Annals of

Ann Hematol (1992) 64:28-34

Hematology 9 Springer-Verlag 1992

Original article The relevance of cultures of catheter-drawn blood and heparin-lock fluid to diagnose infection in hematologic patients H. E L. Guiot 1,2, A.V. Helmig-Schurter 1,2, and J.M. van 't Noordende a Department of Infectious Diseases 2 j.A. Cohen Institute for Radiopathology and Radiation Protection 3 Department of Electronmicroscopy, University Medical Center, P.O. Box 9600, 2300 RC Leiden, The Netherlands Received August 19, 1991/Accepted November 15, 1991

Summary. The results of bacteriologic cultures of blood and heparin-lock fluid, both drawn from the central venous catheters of 54 consecutive oncohematologic patients, have been used to determine their value for the diagnosis of systemic and catheter-associated infection. In 30 patients with clinical signs of infection (bacteremia or septicemia), 74 of 1000 (7.407o) heparin-lock fluid cultures, 114 of 542 (2107o) catheter-drawn blood cultures, and 36 of 134 (26070) venipuncture blood cultures became positive, whereas in 24 patients without clinical signs of infection the respective values were 5 of 700 (0.7%), one of 220 (0.4%), and none of ten cultures. Comparison of the results of cultures sampled on the same day reveals that the positive and negative predictive values for catheter-drawn blood cultures, with the venipuncture blood cultures taken as the standard for bacteremia, are 8207o and 95% respectively. The results of heparin-lock fluid are indicative for clinically relevant colonization of the catheter. Three or more positive heparin-lock fluid cultures, sampled on subsequent days, were correlated with the occurrence of bacteremia or septicemia with a positive predictive value of 100070. The conclusions are supported by the results of scanning electron microscopy. Key words: Intravenous catheters - Blood cultures Diagnosis - Bacteremia

bacteriologic cultures, because during and after chemotherapy or bone marrow transplantation patients often are febrile. Central venous catheters offer a simple way to obtain blood and to avoid the discomfort of painful punctures, hematomas, and the frustration of trying to obtain vascular access in patients with fibrotic or poorly visible blood vessels. Although in many hematologic centers it is customary to sample blood via the central venous catheter, the results of these blood cultures are thought to be clinically irrelevant. The objection is that the bacteria isolated from catheter-drawn cultures are not necessarily the same as those cultured from blood obtained by venipuncture. So the results might reflect only colonization of the catheter and not bacteremia responsible for clinical symptoms. The aim of our study was to determine the value of catheter-drawn blood cultures for diagnosing bacteremia in selectively decontaminated oncohematologic patients, and the value of heparin-lock fluid cultures for diagnosing catheter-associated infection at an early stage. Heparin locks (see " M e t h o d s " ) were used to prevent intraluminal clotting inside the cannula. To support the conclusions of the study, scanning electron microscopy of portions of the intravenous catheter of three patients was performed, and the results are presented as three case reports.

Introduction

Patients and methods

One of the parts of hospitalization that oncohematologic patients remember most is the recurrence of venipuncture to obtain blood for various purposes [11]. Especially in these patients blood samples are frequently needed for

Address for correspondence: H. EL. Guiot, Department of

Haematology/BMT, Bldg. 1, C 6 - 106, University Medical Center, P.O. Box 9600, 2300 RC Leiden, The Netherlands

All patients admitted consecutively to the Department of Hematology and Bone Marrow Transplantation during the period 1988-1990 to undergo various cytotoxic treatments [10, 15] for remission induction or conditioning prior to bone marrow transplantation were included in the study. All patients became granulocytopenic (< 100/mm 3) for at least 2 weeks following their cytotoxic treatment. Infection prevention consisted of selective decontamination (oral regimen: neomycin, polymyxin B, pipemidic acid, amphotericin B) in combination with local antiseptic treatment of the oral cavity, protective isolation, and well-cooked food [4, 5]. In addition, most patients were given penicillin G prophylaxis for streptococcal infection [6].

29

Management of intravenous catheters Central venous access was realized through percutaneously installed, subcutaneously tunneled subclavian-atrium catheters (pyrogenfree ethylene oxide-sterilized silicone, 1.2-2.0 mm, code 2181.20 Vigon BP7-95440; Ecouen, France) [12]. The catheter was sutured to the skin near the entry site. The external infusion system was renewed daily. After removal of this system, instead of a drip infusion to prevent intraluminal clotting, the intravenous cannula was filled with heparin (Thromboliquine, Organon; Boxtel, The Netherlands) and locked with a Luer-lock cap. The catheter remained locked for a median of 2 h (range 30 min - 12 h, depending on whether it was needed again for infusion; 90% remained locked for 1 - 3 h).

Sampling of cultures Prior to the connection of a new infusion system, the heparin-lock fluid was drawn from the intravenous catheter by puncture of the iodine-sterilized rubber membrane of the Luer-lock cap. This sample (approximately 1 ml) is cultured daily and called heparin-lock fluid. When the patient had fever ( > 38~ axillary), blood cultures were taken via the intravenous line and percutaneously by intravenous puncture. To exclude contamination of the catheter-drawn blood culture with infusion fluid, the first 2 ml aspirated from the cannula were discarded. Then approximately 12 ml of blood was drawn via the iodine-sterilized rubber membrane of the Luer-lock cap of the stopcock. Venipuncture blood (12 ml) was sampled by percutaneous puncture after disinfection of the skin (1070 iodine in 70~ alcohol).

solid agar culture medium (3 ml blood in 2 ml liquoid, Roche; Basel, Switzerland, mixed with 10 ml melted - 5 6 ~ BHI agar, Oxoid, and cooled down to room temperature). All cultures were incubated overnight (37~ for maximally 5 days and subcultured using routine bacteriologic techniques.

Typing of bacterial strains. To demonstrate differences or conformity between bacterial strains isolated from the cultures obtained during episodes of infection, macroscopic and microscopic morphology, antimicrobial sensitivity patterns, and biochemical characteristics (e.g., API-20 E, API-20 NE, API-20 STREP; France) were studied. Bacteremia/septicemia. A patient with positive venipuncture blood cultures but without clinical signs of major infection is considered to have bacteremia. A patient with positive venipuncture blood cultures and signs of major infection, i.e., fever ( > 38~ axillary) with at least chills and malaise, is considered to have septicemia.

Scanning electron microscopy. Portions of the proximal tip and the distal subcutaneous tunnel were bisected and submerged in 1.5% glutaraldehyde in cacodylate buffer (0.14 M, pH 7.4) for 90 min at room temperature. Next, they were placed in 70% alcohol and stored at - 2 ~ Finally, the catheter portions were dehydrated in 100070 ethanol Critical-point drying was performed with CO2. The samples were metallized with gold and examined with a Philips 525 scanning electron microscope.

Results

Culture techniques Heparin-lock fluid. The samples were added to brain-heart infusion broth (9 ml bated overnight (blood agar no. used to identify

BHI - Oxoid Ltd.; Basingstoke, England), incu(37~ and subcultured on sheep blood agar 2, Oxoid). Routine bacteriologic techniques were the micro-organisms.

Catheter-drawn blood and venipuncture blood. Approximately 12 ml of blood was distributed over aerobic culture medium (5 ml blood in 70 ml Trypticase Soy Broth; Becton Dickinson, Cockeysville, MD 21030, USA), anaerobic culture medium (3 ml blood in 30 ml cysteine HC1 BHI broth, Oxoid; London, England), and

Table 1. Flow chart of results of cultures of heparin-lock fluid, catheter-drawn blood and venipuncture blood in 54 consecutive oncohematologic patients with a central venous catheter

Over a 2-year period, 59 consecutive patients were studied, representing 59 episodes of central venous access by indwelling catheters. With five patients, it was not clear whether they developed clinically relevant infection. At the time the blood cultures became positive their clinical conditions were not compatible with infection, their granulocytes recovered, and their indwelling catheters were removed. For that reason they are excluded from the study. During 24 of the 54 remaining episodes of central venous access the patients involved were not suspected of

No. of patients

54

No. of episodes No. of venipuncture blood cultures

24 10

30 134

Bacteriologically documented bacteremia/septicemia

no

yes

720 (30)

1023 (34)

No. of days with central venous access (per patient)

No. of cultures positive (+), negative ( - ) -heparin-lock fluid -catheter-drawn blood -venipuncture blood Positive cultures (%) -heparin-lock fluid -catheter-drawn blood -venipuncture blood

+ 5 1 0

695 228 10

0.7 0.4 0

a Two or three blood cultures per day are counted separately

+ 74 114 a 36

m

926 428 98 7.4 21 27

30

systemic bacterial or fungal infection, either because they had no signs of infection or because fever and malaise arose from other causes. During these 24 episodes (720 days), 939 cultures were performed, 0.6~ of which were positive, i.e., five of 700 heparin-lock fluid cultures, one of 229 catheter-drawn blood cultures, and none of ten venipuncture blood cultures (Table 1). The six positive cultures yielded coagulase-negative staphylococci; this occurred only once during the episodes involved, and the patients were not given antibiotic therapy. In 30 patients bacteriologically documented bacteriemia or septicemia occurred; i.e., positive venipuncture blood cultures were associated with clinical signs of bacterial or fungal infection. During the corresponding 30 episodes (1023 days, including days without infection), 1676 cultures were performed, 13.5~ of which were positive, i.e., 74 of 1000 (7.4~ heparin-lock fluid cultures, 114 of 542 (21070) catheter-drawn blood cultures, and 36 of 134 (27O7o) venipuncture blood cultures (Table 1). To illustrate the chronologic distribution of positive and negative cultures, the results obtained in four patients are given as an example (Fig. 1). The 24 patients without documented infection are represented by patient A. During 19 days of catheterization 17 daily heparinlock fluid samples were cultured, none of which were positive. On day 6 fever started - however, without other clinical signs of infection. During 5 consecutive days seven catheter-drawn blood cultures and three venipuncture blood cultures were obtained, all of which were negative. Normalization of the body temperature occurred without antimicrobial therapy. In the nine patients represented by patient B, the heparin-lock fluid cultures became positive more or less simultaneously with the occurrence of clinical signs of infection. Both catheter-drawn and venipuncture blood cultures yielded the same species as those cultured from the heparin-lock fluid, although not all cultures sampled on the same day were positive. The corresponding infections were caused by various spe-

patient A hep.lock cath.drawn bid venipunct.bld

0000000000-0-000000 - 00000 000

patient B hepJock cath.drawn bid venipunc.bld

O00000000XO00000000000 X X X X 0

patient C hep.lock cath.drawn bid venipunc.bld

0000000

X X X X X X 00 X X X X X X --

patient D hep.lock cath.drawn bid venipunc,bld DAYS

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 X X 0 0 X 1

5

10

15

20

Fig. 1. Chronologic distribution of positive and negative heparinlock fluid (hep.lock) cultures, catheter-drawn blood (cath,drawn bld) and venipuncture (venipunc. bld) cultures. Circle: negative; cross." positive; line: no cultures taken

cies: i.e., Pseudomonas aeruginosa (n = 1), Escherichia coli (n = 1), Salmonella spp (n = 1), Staphylococcus aureus (n = 2), Capnocytophaga ochracea (n = 1), Candida albicans (n = 2), mixed S. epidermidis and S. mitis (n = 1). The two Candida infections were conspicuous because of the persistence of positive cultures despite antifungal therapy, i.e., 4 + 1 heparin-lock fluid cultures, 7 + 5 catheter-drawn blood, and 1 + 3 venipuncture blood cultures positive during 5 and 9 days, respectively. Both patients died due to disseminated Candida infection. In the 11 patients represented by patient C, the hepatin-lock fluid cultures became positive prior to the clinical symptoms of infection. When fever started a few days later and blood cultures were performed, these cultures yielded the same species and types as those cultured from the heparin-lock fluid; all patients had at least one positive venipuncture blood culture. The corresponding infections were all caused by coagulase-negative staphylococci. In Fig. 2 the cumulative results of the cultures are shown by an excision from the whole episode of central venous access, i.e., 10 days before and 10 days after the first positive blood culture. Half of the patients had two or more heparin-lock fluid cultures yielding staphylococci prior to the occurrence of clinical signs of infection. One patient (p. 12) was an exception, with two catheterdrawn and two venipuncture blood cultures yielding Staphylococcus epidermidis but no positive heparin-lock fluid cultures. From the same figure (Fig. 2) it can be " r e a d " that six catheters were removed due to persisting positive cultures. In the nine patients represented by patient D (Fig. 1), bacteremia was documented by venipuncture blood cultures, i.e., at least one culture positive. The catheterdrawn blood cultures yielded the same species, but the majority of heparin-lock fluid cultures were negative. All these infections were caused by streptococci; S. mitis (n = 7), S. sanguis (n = 1), and S. morbilorum (n = 1). The cumulative results are shown in Fig. 3. Counting the number of cultures by pairs obtained on the same day, both positive or both negative and pairs with opposite results, reveals a discrepancy of only 4o70, i.e., 25 venipuncture blood cultures were positive and 47 negative versus 28 catheter-drawn cultures positive and 44 negative. Both types of cultures give similar results (/7 = 0.36). Calculation of the corresponding accuracy indices of cultures sampled on the same day (Table 2) reveals that the positive and negative predictive values for catheterdrawn blood cultures, with the venipuncture blood cultures as the standard for bacteremia, are 82~ and 95% respectively. The results of heparin-lock fluid cultures with both catheter-drawn and venipuncture blood cultures as standard, give positive and negative predictive values of 77~ and 94~ respectively. These predictive values are biased, because heparin-lock fluid was cultured on a daily basis, whereas blood cultures were performed only when the patient had fever. Actually, the predictive values are used incorrectly because the heparinlock fluid cultures were not aimed as alternative blood cultures. The purpose of the heparin-lock fluid cultures

31 1 4 !~

10

Li 12 10

i 6

8-

6

4

42 20 ~

0

U m

b

B

10

m b

e

r

n u

e

r

8-

765 4-

6

3-

4

22 0

1 i

i

~Ji ] i i~

i

i

p

4

i[-

~' . . . . . . . . . . .

0

C

12

5

8

4

e 84

3

4-

2

o

i

i''a

-10

... ill i

"J

~

r

-5

i

"4

i

i

i

i

F

i

~

i

i

i

i

E

i

~

i

i

I

6

lO

2

i

i

r

i

L

i

C

1 i

0

r

r

5

q

,

i

~

"l ~

:

10

number of days

0 -10

-5

0

5

10

number of days

Fig. 2 A - C . The cumulative results of cultures obtained in 12 patients with bacteremia or septicemia with coagulase-negative staphylococci. A, Heparin-lock fluid cultures; B, catheter-drawn blood cultures; C, venipuncture blood cultures; black bars, all cultures; hatched bars, positive cultures; day O, first day with positive blood cultures. In the majority of patients the episode of catheterization exceeded the number of days indicated in the figure

Fig. 3 A - C . The cumulative results of cultures obtained in 12 patients with bacteremia or septicemia with viridans streptococci. A, Heparin-lock fluid cultures; B, catheter-drawn blood cultures; C, venipuncture blood cultures; black bars, all cultures; hatched bars, positive cultures; day O, first day with positive blood cultures. In the majority of patients the episode of catheterization exceeded the number of days indicated in the figure

Table 2. Cross-tabulation of the results of venipuncture blood cul-

to predict catheter-associated infection is illustrated by the finding that all patients with two or more positive heparin-lock fluid cultures on consecutive days sooner or later had positive venipuncture blood cultures, with the same type o f bacteria. Finally, scanning electron microscopy (SEM) o f portions o f the catheters o f three patients has been performed. Regarding results o f cultures, these patients fulfilled the characteristics o f patients A, C, and D, as shown in Fig. 1.

tures, catheter-drawn blood cultures, and heparin-lock fluid cultures sampled on the same day

Catheter-drawn blood Positivea Negative

Heparin-lock fluid Positive a Negative

Number of cultures Venipuncture blood cultures Positive a Negative 23 5 2 42 Venipuncture and/or catheter-drawn blood cultures Positivea Negative 37 11 33 546

a The same bacterial species (types) were cultured from the corresponding cultures

Case report L In a patient without clinical signs o f infection during 26 days o f catheterization, all 26 heparin-lock fluid cultures and two catheter-drawn b l o o d cultures were negative. The intraluminal surface o f the tip is covered with a thin layer o f fibrous material (Fig. 4 A ) containing

32

Fig. 5. A, The lumen of the tip of a bisected central venous silicone catheter (30 days in situ, Staphylococcus epidermidis bacteremia. Inside the lumen is a thick layer of slime and fibers, x25 B, Detail of A. Large numbers of spherical structures, embedded and associated with fibers appearing like staphylococci and fibrin. x 2500

spherical structures with the appearance o f staphylococci (Fig. 4B). The outer surface o f the tip is clean except for some a m o r p h o u s particles (Fig.4C); the intraluminal and outer surfaces o f the subcutaneous tunnel portion are clean too (not shown).

Fig. 4. A, The lumen of the tip of a bisected central venous catheter (26 days in situ, no infections) is covered with a thin layer of fibrous material, x 39 B, Detail of A. Spherical structures appearing morphologically like staphylococci are closely associated with fibers, probably fibrin. • 8800 C, Outer surface of the tip shown in A and B. Except for some amorphous particles, the tip is clean and relatively smooth. •

Case report IL In another patient, five catheter-drawn blood cultures and two heparin-lock fluid cultures yielded Staphylococcus epidermidis. Therapy with vancomycin resulted in negative cultures and disappearance o f fever. Following removal o f the cannula 14 days after the last positive culture, SEM revealed that the intraluminal surface o f the tip was covered by a huge flake o f material (Fig. 5A) c o m p o s e d o f condensed glycocalyx [8, 13], a fibrous material containing large numbers o f erythrocytes and coccoid bacteria, intimately associated with or embedded in the matrix (Fig. 5 B). At the distal part o f the

33

Fig. 6. The lumen of the tip of a bisected central venous silicone catheter (50 days in situ, Streptococcus mitis septicemia and Lactobacillus easei rhamnosis bacteremia) is covered with a very thin, delicate layer of fibrous material. •

catheter, the subcutaneous tunnel portion, the fibrin and erythrocytes have disappeared. However, coccoid bacteria 9 embedded in slime are still present. The outer surface of the catheter is clean (not shown). Case report IIL In the third patient Streptococcus mitis septicemia occurred, followed by bacteremia with Lactobacillus casei (nine blood cultures). All heparin-lock fluid cultures were negative. The catheter, suspected to be the source of bacteremia, appears to be clean; only at the intraluminal surface of the tip is there a delicate layer of fibrous material (Fig. 6), containing some erythrocytes but no morphological structures resembling bacteria.

cohort with subsequent cultures. These single positive cultures probably reflected either clinically irrelevant transient bacteremia or clinically irrelevant colonization of the intravenous catheter [13]. The results suggest that if blood aspirated from the catheter yields bacteria, then these bacteria are also present in the circulation. The explanation might be that the bacteria which colonize the catheter are firmly adherent [3, 13] and relatively slow growing [13]. Those small numbers that are released from the biomass are washed away by infusion fluids, causing obscure transient bacteremia. The more the catheter is washed, the lower the chance is that catheter-drawn blood cultures contain bacteria originating from the intraluminal wall. Only after excessive proliferation, as shown in the SEM graph of Fig. 5 B, may the number of released bacteria, or fragments of biomass, be enough to ensure positive catheter-drawn blood cultures. However, by that time venipuncture blood cultures will also become positive, due to infusion fluid shedding released bacteria into the circulation in numbers enough to produce positive cultures. The results of the heparin-lock fluid cultures indicate that they may reveal clinically relevant catheter colonization. In 11 patients such colonization occurred prior to clinical signs of bacteremia. In a number of other patients, heparin-lock fluid cultures were negative although blood cultures were positive. These results may be taken as an indication that the catheter is not colonized like the SEM graph of Fig. 6, or that obscure colonization is not (yet) relevant, i.e., colonization like in the SEM graph of Fig. 4 B. In conclusion, in our oncohematologic patients catheter-drawn blood cultures may be used to diagnose bacteremia, and heparin-lock fluid cultures may be used for central venous catheter surveillance.

References Discussion The starting point of the present study was the assumption that micro-organisms cultured from blood drawn from the central venous catheter may reflect colonization of the catheter only and not necessarily bacteremia, which would mean that catheter-drawn cultures are unreliable for diagnosing bacteremia. In a number of articles [1, 2, 7] it was shown that micro-organisms cultured from catheter aspirates either are not present in the peripheral blood or belong to other species. On the other hand, some authors described a good correlation between venipuncture blood cultures and cultures from central venous catheter aspirates [9, 14]. Our results are in agreement with those of the latter authors. In the 30 patients with proven bacteremia the correlation between catheter-drawn blood and venipuncture blood culture was high, and in the 24 patients without bacteriologically documented bacteremia only 0.4% of the catheter-drawn blood cultures were positive. Moreover, those few that were positive were not positive in

1. Applefeld JJ, Caruthers TE, Reno DJ, Civetta JM (1978) Assessment of the sterility of long-term cardiac catheterization using the thermodilution Swan-Ganz catheter. Chest 74:377-380 2. Bryant JK, Strand CL (1987) Reliability of blood cultures collected from intravascular catheter versus venipuncture. Am J Clin Pathol 88:113-116 3. Evans RC, Holmes CJ (1987) Effect of vancomycin hydrochloride on Staphylococcus epidermidis biofilm associated with silicone elastomer. Antimicrob Agents Chemother 31:889-894 4. Guiot HFL, van den Brock PJ, van der Meet JWM, van Furth R (1983) Selective antimicrobial modulation of the intestinal flora of patients with acute nonlymphocytic leukemia: a doubleblind, placebo-controlled study. J Infect Dis 147:615-623 5. Guiot HFL, Helmig-Schurter AV, van der Meet JWM, van Fnrth R (1986) Selective antimicrobial modulation of the intestinal microbial flora for infection prevention in patients with hematologic malignancies. Evaluation of clinical efficacy and the value of surveillance cultures. Scand J Infect Dis 18: 153160 6. Guiot HFL, Peters WG, van den Brock PJ, van der Meet JWM, Kramps JA, Willemze R, van Furth (1990) Respiratory failure elicited by streptococcal septicaemia in patients treated with cytosine arabinoside, and its prevention by penicillin. Infection: 131-137

34 7. Hudson-Civetta JA, Civetta JM, Martinez OV, Hoffman TA (1987) Risk and detection of pulmonary artery catheter-related infection in septic surgical patients. Crit Care Med 15:29-34 8. Marrie TJ, Costerton JW (1984) Scanning and transmission electron microscopy of in situ bacterial colonization of intravenous and intra-arterial catheters. J Clin Microbiol 19:687-693 9. Moyer MA, Edwards LD, Farley L (1983) Comparative culture methods on 101 intravenous catheters. Routine, semi-quantitative, and blood cultures. Arch Intern Med 143:66-69 10. Peters WG, Willemze R, Colly LP, Guiot HFL (1987) Side effects of intermediate- and high-dose cytosine arabinoside in the treatment of refractory or relapsed acute leukaemia and non-Hodgkin's lymphoma. Neth J Med 30:64-67 11. Raaf JH (1985) Results from use of 826 vascular access devices in cancer patients. Cancer 55:1312-1321

12. Rosen M, Latto IP, Shang NgW (1981) Handbook of percutaneous central venous catheterization. Saunders, Philadelphia 13. Tenney JH, Moody MR, Newman KA, Schimpff SC, Wade JC, Costerton JW, Reed WP (1986) Adherent microorganisms on lumenal surfaces of long-term intravenous catheters. Arch Intern Med 146:1949 14. Tonnesen A, Peuler M, Lockwood WR (1976) Cultures of blood drawn by catheters vs venipuncture. Blood Cultures 235: 1877 15. Willemze R, Peters G, van Hennik MB, Fibbe WE, Kootte AMM, van Berkel M, Lie R, Rodenburg C J, Veltkamp JJ (1985) Intermediate and high-dose Ara-C and m-AMSA (or daunorubicin) as remission and consolidation treatment for patients with relapsed acute leukaemia and lymphoblastic non-Hodgkin-lymphoma. Scand J Haematol 34:83-87

The relevance of cultures of catheter-drawn blood and heparin-lock fluid to diagnose infection in hematologic patients.

The results of bacteriologic cultures of blood and heparin-lock fluid, both drawn from the central venous catheters of 54 consecutive oncohematologic ...
2MB Sizes 0 Downloads 0 Views