134

Notes

6. Beck-Sague CM, Jarvis WR, Brook JH, Culver DH, Potts A, Gay E, Shotts BW, Hill B, Anderson RL, Weinstein MP: Epidemic bacteremia due to Acinetobacter baumannii in five intensive care units. American Journal of Epidemiology 1990, 132: 723-733. 7. Seifert H, Baginski R, Schulze A, Pulverer G: Antimicrobial susceptibility of Acinetobacterspecies. Antimicrobial Agents and Chemotherapy 1993, 37: 750753. 8. Weaver RE, Actis LA: Identification of Acinetobacter species. Journal of Clinical Microbiology 1994, 32: 1832. 9. Bouvet PJ, Grimont PA: Identification and biotyping of clinical isolates of Acinetobacter. Annales de l'lnstitut Pasteur Microbiologie 1987, 138: 569-578. 10. Hartstein AI, Morthland VH, Rourke JW, Freeman J, Garber S, Sykes R, Rashad AL: Plasmid DNA fingerprinting of Acinetobacter calcoaceticus subspecies anitratus from intubated and mechanically ventilated patients. Infection Control and Hospital Epidemiology 1990, 11: 531-538. 11. Seifert H, Schulze A, Baginski R, Putverer G: Comparison of four different methods in the epidemiological typing of Acinetobacter baumannii. Journal of Clinical Microbiology 1994, 32: 1816-1819. 12. AI-Khoja MS, Darrell JH: The skin as the source of Acinetobacter and Moraxella species occurring in blood cultures. Journal of Clinical Pathology 1979, 32: 497-499. 13. Kline MW: Review of recurrent bacterial meningitis. Pediatric Infectious Disease Journal 1989, 8: 630634. 14. Wen DY, Bottini AG, Hall WA, Haines SJ: Infections in neurologic surgery. The intraventricular use of antibiotics. Neurosurgery Clinics of North America 1992, 3: 343-354. 15. Rodriguez K, Dickinson GM, Greenman RL: Successful treatment of gram-negative bacillary meningitis with imipenem/cilastin. Southern Medical Journal 1985, 78: 731-732. 16. S,~nchez JF, Sanz-Hospital J, Guerrero A, MartinezBeltr~n J, Quereda C: Curaci6n con imipenem-cilastino de una meningitis por Acinetobacter catcoaceticus. Enfermedades lnfecciosas y Microbiotogia Clinica 1991, 9: 512-513.

Eur. J. Clin. Microbiol. Infect. Dis.

Role of Catheter Colonization and Infrequent Hematogenous Seeding in Catheter-Related Infections E. Anaissie 1, G. Samonis 2, D. K o n t o y i a n n i s 1, J. C o s t e r t o n 3, U. Sabharwal 4, G. B o d e y 1, I. R a a d 1.

Adult cancer patients were prospectively studied to determine the relationship between ultrastructural and microbiologic catheter colonization and clinical catheter-related infections. Participants included

38 patients

whose

central

venous

catheters were removed because of suspected catheter infection (16 patients) or other noninfectious causes (22 controls). The presence of clinical infection was determined. Catheters were examined by microbiologic methods (sonication and roll-plate culture) and by scanning and transmission electron microscopy. Ultrastructural microbial colonization and biofilm formation were universal and occurred as early as one day after catheter insertion. The extent of biofilm formation was unrelated to the clinical status of patient or the catheter microbiological findings. Secondary seeding of catheters was rarely seen. Catheter microbial biofilm formation occurs early, is universal and does not necessarily represent an infectious condition.

Central venous catheters (CVC) are being used with increasing frequency for the administration

of therapeutic agents (1). Infectious complications have been associated with the use of such devices (2), however, the pathogenesis, diagnosis and management of these infections are topics of controversy (3, 4). The bacterial biofilm on the CVC's surfaces has been viewed as a focus of perl Section of Infectious Diseases, Department of Medical Specialties, University of Texas, M.D. Anderson Cancer Center, 1515 Holcombe Boulevard (Box 47), Houston, Texas 77030, USA. 2University of Crete, Division of Medicine, Heraklion 71110, Crete, Greece. 3Montana State University, Center for Biofilm Engineering, Bozeman, Montana 59717, USA. 4Department of Biological Sciences, University of Calgary, Calgary, Alberta, Canada.

Vol. 14, 1995

sistent infection refractory to antimicrobial agents and host defenses, leading sometimes to relapse of the sepsis (5, 6). This prospective study was undertaken to study whether a relationship exists between vascular catheter biofilm, hematogenous seeding of catheter from noncatheter site and clinical catheter-related infection. Patients and Methods. This study was conducted at The University of Texas M.D. Anderson Cancer Center from August 1986 to January 1987. Nontunnelled CVCs, which were removed from 16 cancer patients because of suspected catheter-related infection (CRI), were prospectively examined. A control group of 22 patients whose catheterization had been discontinued because of the patient's death, treatment completion or CVC dysfunction was also examined. Patients in both groups were consecutively identified by the nurses of the infusion therapy team. Catheter-related infection was suspected if fever (> 101°F) or bacteremia occurred in a patient with a CVC and persisted 48 hours after appropriate antimicrobial therapy was initiated. Catheters were removed at the bedside and processed in the laboratory within 4 hours of removal. After skin preparation with a povidoneiodine solution, the CVC was exposed below the insertion site through a skin incision. The catheter was then removed aseptically and placed in a sterile container. At the time of removal, a nurse from the infusion therapy team examined the CVC site for inflammation. Each CVC was then cut into seven segments from the tip to the hub, with each segment 2 cm long. Two segments were taken from the hub area, three from the subcutaneous area and two from the intravascular distal area (tip). One segment from the tip, subcutaneous area and hub were examined by electron microscopy. Another segment from the tip, two segments from the subcutaneous area and one from the hub were independently cultured quantitatively by the semiquantitative method of Maki et al. (7). The same segments were then cut open longitudinally, then scraped with a scalpel, and the catheter segment and scalpel tip were sonicated for 8 min according to published methods (8). The entire internal and external surfaces of each specimen were examined by scanning and transmission electron microscopy at the Department of Biological Sciences, University of Calgary, according to published methods (8). The extent to which the biofilm covered the biomaterial surface was calculated as the percentage of surface covered by biofilm. Subsequently, these

Notes

135

surfaces were examined at higher magnifications and at regular intervals to ascertain that the amorphous biofilm layer actually contained microbes. The microbial cells and other features of the biofilm surfaces were photographed. The following definitions were used: bloodstream infection- signs and symptoms of infection (fever or chills) and at least one positive blood culture, except for coagulase-negative staphylococci for which at least two blood cultures drawn from separate sites were required; significant catheter colonization - the isolation of 15 or more colonyforming units (cfu) by the roll-plate technique or more than 100 cfu by the scrape-sonication technique from any catheter-segment; catheter-related bloodstream infection - significant colonization of a catheter segment and concurrent bloodstream infection by the same organism (same species and antibiogram) recovered from bloodstream and catheter, and no evidence of another source of infection; primary bloodstream infection - a bloodstream infection with a negative catheter culture and no apparent source of infection; and, secondary bloodstream infection bloodstream infection with a negative catheter culture resulting from a source other than the catheter (such as pneumonia, urinary tract or gastrointestinal tract infection). Differences between frequencies of categorical variables were determined with the chi-square or Fisher's exact test. Continuous variables with normal distribution (such as age) were compared with the Student's t test. Continuous variables not normally distributed (such as the distribution of biofilm on catheters) were compared with the Wilcoxon rank-sum test. A p value < 0.05 was considered significant. Results and Discussion. Although patients with suspected infection were significantly older (mean age 57.5 + 11.3 years) than control patients (mean age 48.3 + 15.7 years) (p = 0.04), their underlying immunosuppression risk factors (leukemia, solid tumors, neutropenia) were comparable. No significant differences were found in the characteristics of the CVCs removed from patients with suspected CRI and control group patients regarding the number of lumen, type of material used, location of insertion and the duration of catheterization. Biofilm was observed universally. Table 1 shows the distribution of biofilm as examined by scanning electron microscopy. The extent of biofilm did not differ between the two groups and was evenly distributed between the inner and outer

136

Notes

Eur. J. Clin. Microbiol. Infect. Dis.

Table 1: Distribution of biofilms on catheters as determined by scanning electron microscopy Mean distribution of biofilm (%)" Patients with suspected CRI Catheter surface

Controlb

Tip

Subcutaneous

Hub

Tip

Subcutaneous

Hub

$1

$2

$3

$4

$1

$2

$3

$4

Inner

9.1

13.1

14.9

3,8

6.4

11,3

17.0

6.4

Outer

8.5

6.3

10.8

3,3

7.4

11,9

14.3

6.0

a M e a n distribution of biofilm was quantitated as a percentage of surface area covered by the biofilm layer. bpatients with no infection. CRI = catheter-related infection.

Table2: Visible organisms on catheters of patients with secondary bloodstream infections Type of secondary bloodstream infections

No. of episodes

Visible organisms on catheters by electron microscopy

Gram-negative bacillemia

6

cocci

Fungemia (yeasts)

1

cocci

Polymicrobial (staphylococci and yeast)

3

cocci on 2 catheters none on I catheter

Polymicrobial (staphylococci and gram-positive bacilli)

2

cocci

surface of the CVC. Whether the CVC was associated with infection or not, the subcutaneous segment had significantly more biofilm than either the hub (Wilcoxon rank-sum test, p < 0.005) or the tip (p < 0.05). Of the 40 CVCs examined, 27 (13 in cases and 14 in controls) showed visible bacteria (mainly coccoid cells) on the internal surface, compared with visible bacteria detected on the external surface of only 16 catheters (6 in cases and 10 in controls) (p = 0.0t). Usually, spherical shapes of bacterial cocci (probably coagulase-negative staphylococci) were buried in the biofilm to varying degrees, regardless of the quantitative culture findings or the patient's clinical status. The colonization started to occur, however, very early in the course of catheterization. Biofilm was found on the CVC of all six patients whose catheters had been in place for less than three days. In one of these patients, the CVC had been in place for less than 24 h. Quantitative catheter cultures revealed that 6 (15 %) CVCs were significantly colonized. Four

(10 %) were associated with catheter-related bloodstream infections caused by Enterobacter

cloacae, Candida albicans, Staphylococcus epidermidis, or Acinetobacter calcoaceticus. During the course of catheterization, 18 episodes of bloodstream infections occurred in 16 patients. Examination of these CVCs by both microbiologic and electron microscopy methods revealed the absence of microbial seeding in most of these patients (Table 2). When classified according to the clinical and microbiologic sources of the infection, the 18 bloodstream infections consisted of: four catheter-related infections, two primary infections and 12 secondary infections. All of the catheter-related and primary bloodstream infections had almost total concordance between the organism isolated from the blood and that detected on the catheter surface by electron microscop): This concordance is based on morphology, and it is possible that the cocci visualized on the catheter are not the same as those cultured from the blood. However, in 67 % of the secondary bloodstream infections, the visible organisms detected on the catheter surface by electron microscopy and the findings of earlier bacteremia or fungemia cultures were totally inconsistent (Table 2). Based on electron microscopy observation, our results indicated a universal colonization of indwelling CVC, mainly by cocci, presumably coagulase-negative staphylococci. This colonization occurred as early as 24 hours after CVC insertion, and it seemed to be somewhat independent of the patient's clinical condition. Our results indicated, therefore, that catheter colonization does not necessarily result in bloodstream infection. We were not able to establish any significant correlation between the clinical features of catheterized patients and the electron microscopy re-

Vol. 14, 1995

sults. The morphologic (coccoid) forms seen by scanning electron microscopy, the cell-wall structure (gram-positive) seen by transmission electron microscopy and the taxonomic identifications

(Staphylococcus epidermidis, Staphylococcus aureus) of recovered organisms all indicate that the skin flora play a maior role in colonization of the catheter, as suggested (9). In addition, the higher density of biofilm in the subcutaneous segments, compared to that in the tip (Table 2), points to the skin and hub, rather than hematogenous seeding from other sites, as the sources of microbial colonization. Although important, the hub is not the major contributor to catheter colonization, in view of the extensive colonization of the external surfaces and the higher density of biofilm in the subcutaneous segments compared with the hub (Table 2). Data from other investigators agree with these findings. Tenney et al. (10) using transmission electron microscopy in Hickman catheters removed from patients with cancer, found that 87 % of these catheters were colonized by gram-positive cocci in the glycocalyx adherent to the surface of the catheter lumen. This almost universal colonization of the Hickman catheters by cocci seemed to correlate poorly with either organisms recovered from catheter culture or the presence and microbiologic type of bloodstream infection. Finally, our data suggested that hematogenous seeding of vascular catheters following secondary gram-negative or fungal bloodstream infection seems to be uncommon. None of the ten CVCs exposed to bloodstream pathogens from noncatheter sites showed any evidence of clinical or microbiologic catheter infections. The microorganism detected on the CVC surface by scanning electron microscopy and transmission electron microscopy failed to correlate with 67 % of the preceding secondary bloodstream infections (mostly gram-negative bacilli and Candida infection) (Table 2). However, a large prospective randomized study would be necessary to determine whether catheter removal is beneficial in the setting of a secondary bloodstream infection. References 1. Bottino J, McCredie KB, Groschel DHM, Lawson M: Long-term intravenous therapy with peripherally inserted silicone elastomer central venous catheters in patients with malignant diseases. Cancer 1979, 43: 1937-1943. 2. Hamptom AA, Sheretz RJ: Vascular-access infections in hospitalized patients. Surgical Clinics of North America t988, 68: 57-71.

Notes

137

3. Coltignon PJ, Munro R: Laboratory diagnosis of intravascular catheter associated sepsis. European Journal of Clinical Microbiology and Infectious Diseases 1989, 8: 807-814. 4. Decker MD, Edwards KM: Central venous catheter infections. Pediatric Clinics of North America 1988, 35: 579-612. 5. Costerton JW, Cheng KJ, Geesey GG, Ladd TI, Nickel JC, Dasgupta M, Marrie TJ: Bacterial biofilms in nature and disease. Annual Review of Microbiology 1987, 41: 435--464. 6. Dickinson G, Birno A: Infections associated with indwelling devices: concepts of pathogenesis: infections associated with intravascular devices. Antimicrobial Agents and Chemotherapy 1989, 33: 597--601. 7. Maki DG, Weise CE, Sarafin HW: A semiquantitative culture method for identifying intravenous-catheterrelated infection. New England Journal of Medicine 1977, 296: 1305-1309. 8. Marrie TJ, Costerton JW: A scanning and transmission electron microscopic study of the surfaces of the intrauterine contraceptive devices. Journal of Obstetrics and Gynecology 1983, 146: 384-394. 9. Vas SL, Low DE, Oreopoulos DG: Peritonitis in peritoneal dialysis. In: Nolph KD, Martinus N (ed): Peritoneal dialysis. The Hague 1981, p. 334. 10. Tenney JH, Moody MR, Newman KA, Schimpff SC, Wade JC, Costerton JW, Reed WP: Adherent microorganisms on lumenal surfaces of long-term intravenous catheters. Archives of Internal Medicine 1986, 146: 1949-1954.

Diversity of Nosocomial

Xanthomonas maltophilia (Stenotrophomonas maltophilia) as Determined by Ribotyping R Gerner-Smidt 1'4., B. Bruun 2, M. Arpi 3,

J. Schmidt I

Seventy-seven clinical isolates of Xanthomonas

maltophilia

( Stenotrophomonas

maltophilia)

were consecutively collected from the Rigs1Department of Clinical Microbiology,StatensSeruminstitut, Artillerivej 5, DK-2300 Copenhagen S, Denmark. 2Department of Clinical Microbiology, Rigshospitalet, Afsnit 9301,Juliane Maries Vej 22, DK-2100 Copenhagen ~, Denmark. 3Department of Clinical Microbiology, Frederiksberg Hospital, Nordre Fasanvej 57, DK-2000, Frederiksberg, Denmark. 4Department of Clinical Microbiology,Bispebjerg Hospital, BispebjergBakke 23, DK-2400Copenhagen NV, Denmark.