Initial detection of bacteremia by subculture of unvented tryptic soy broth blood culture bottles

DIAGNMICROBIOLINFECTDIS 1984;2:107-111

107

Initial Detection of Bacteremia by Subculture of Unvented Tryptic Soy Broth Blood Culture Bottles Nancy K. Henry and John A. Washington II

Routine subculture of macroscopically negative blood cultures is a traditional blood culture procedure. The need to perform routine early (6-17 hr) and late (48 hr) subculture of unrented blood culture bottles when u simultaneous subculture of the vented bottle is performed has been questioned. Blood cultures in paired vented and unvented tryptic soy broth (TSB) bottles from 4574 patients were examined retrospectively. Subculture of unvented TSB bottles provided initial detection of 412 (5.0%) isolates from 277 (6.1%) patients and was comparable to that of vented TSB bottles for Pseudomonas and all other microorganisms, except for the Enterobacteriaceae (p < 0.001; vented TSB), Candida (p < 0.001; vented TSB), and Haemophilus influenzae (p < 0.01; unvented TSB). Of the H. int]uenzae isolates, 40% were detected initially by subculture of the unrented TSB bottles; early subculture recovered 67% of these isolates. The value of subculture of unrented TSB bottles is minimized when subculture of the vented TSB bottle is routinely performed; however, routine subculture of the unrented bottle is recommended whenever TSB is used for detection of bacteremia in patients in whom H. in~/uenzae infection is possible. Routine aerobic subculture of macroscopically negative blood culture bottles is a traditional blood culture procedure. Subculture of bottles containing broth alone is recommended for vented or unvented bottles (Relier et al., 1982). Pfaller et al. (1983} reported that aerobic subculture of unvented thiol broth-containing bottles was rarely useful, but cautioned that results may differ with the use of other culture media or subculture procedures and with fastidious bacteria such as Haemophilus. We report our experience with subculture of unvented blood culture bottles containing tryptic soy broth (TSB}. MATERIALS AND METHODS During the study period (1976-1961), blood for culture was collected by trained phlebotomists from patients with suspected bacteremia or fungemia. Throughout the study 20-30 ml of blood was collected from adult patients and 10-ml aliquots were inoculated into two or three blood culture bottles. Blood samples (2-10 ml} were collected from pediatric patients and equal volumes inoculated into two blood culture bottles. The study was restricted to those patients with blood cultures consisting of two bottles containing 100 ml of TSB under vacuum with 0.025% sodium polyanetholesulfonate (SPS) and 5-10% CO2 (Difco Laboratories, Detroit, MI). Upon arrival in the laboratory one of the bottles was vented transiently to release the vacuum; the other remained unvented. From the Mayo Clinic and Mayo Foundation, Rochester, Minnesota. Address reprint requests to: Nancy K. Henry, Ph.D., Section of Clinical Microbiology, Mayo Clinic, Rochester, MN 55905. Received June 27, 1983; revised and accepted December 5, 1983.

Science Publishing Co., Inc. 52 Vanderbilt Avenue,New York, NY 10017

© 1984 Elsevier

0732-g893/84/$03.00

N.K. Henry and J.A. Washington II

108

Bottles were incubated at 35°C in room air and examined macroscopically 6-17 hr after collection/incubation, later the same day, and daily for 7 days and on day 14 before discarding. Subcultures were performed on all macroscopically negative blood culture bottles after 6-17 hr and after 48 hr of incubation. Subcultures were performed by withdrawing an aliquot of broth and inoculating a quadrant of a chocolate blood agar plate, which was then incubated for 48 hr in 5-10% CO2. Anaerobic incubation of subcultures was not performed. Positive cultures were analyzed according to the bottle(s) and means of microorganism detection, that is, macroscopic, early subculture, or late subculture. Patients with positive cultures were analyzed according to the bottle and means of detection of their first or only positive blood culture. The number of microorganisms and the number of positive cases in which the first means of detection was by subculture of unvented bottles were examined to determine the importance of subculture from these bottles. Statistical analyses were performed using the X2 test (Dixon and Massey, 1969). RESULTS A total of 8321 microorganisms was recovered from 4574 patients. No distinction was made between presumed pathogens and contaminants in this study. There were 412 (5.0%) isolates from 277 (6.1%) patients that were recovered by subcultures from unvented TSB blood culture bottles; 542 (6.5%; p ~ 0.001) isolates were recovered from 325 (7.1%; not significant) patients by subculture from vented TSB bottles. Detection of H. inf]uenzae was strongly influenced by subculture of u n r e n t e d TSB blood culture bottles. Subculture provided initial detection of 62 of 135 (46%) isolates, which represented 39 of 81 (48%) patients with H. influenzae bacteremia (Tables 1 and 2). Subcultures from u n r e n t e d TSB bottles detected significantly more isolates of H. influenzae than did those from vented TSB bottles (p ~ 0.01), as shown in

TABLE 1. Microorganisms Detected Initially in Early and Late Subcultures of Unvented and Vented TSB Bottles Number (%) of isolates recovered Unvented bottle only Microorganism Gram-negative bacteria Enterobacteriaceae b Haemophilus influenzae Pseudomonas Other Gram-positive bacteria Staphylococcus aureus

All means ° 2544 135 542 101

Early 92

(3.6)

42 (31.1) 11 (2.0) 2 (2.0)

Late 9

Vented bottle only Early

Late

(0.4)

152 (6.0)

15

(0.6)

20 (14.9) 49 (9.0) 5 (5.0]

13 (9.6) 30 (5.5) 7 (7.0)

23 (17.0) 50 (9.2) 17 (17.0)

983

9

(0.9)

16

(1.6)

19 (1.9)

13

(1.3)

1049 1127 709

22 45 5

(2.1) (4.0) (0.7)

73 8 3

(7.01 (0.7) (0.4)

16 (1.5) 42 (3.7) 9 (1.3)

101 8 13

(9.6) (0.7) (1.8)

269

1

(0.4)

4 (1.5)

10

(3.7)

Staphylococcus epidermidis Streptococci Other Candida

°Includes only those genera of bacteria and yeast detected macroscopicallyor by subculture from one both blood culture bottles. bE. coli, 55%; Klebsiella, 20%; Serratia and Enterobocter, 15%; other, 10%.

or

16 (2.9) 5 (0.8)

13 (1.6) 2 (0.3) 3 (0.5)

54 (6.8)

7 (1.7)

21 (8.1) 2 (3.7]

5 (0.4] 3 (3.7)

Late

13

2

3

2 1

(2.3)

(0.3)

(0.7)

(0.8) (1.9}

12 (0.9) 15 (18.5)

Early 1 (0.1) 8 (9.9)

Late

3 (0.5)

9 (1.1)

2 (0.5)

11 (4.3) 2 (3.7)

Earlierc

i3 (2.3) 9 (1.4]

11 (1.4)

6 (1.5)

10 (3.9) 3 (5.7)

39 (3.1) 3 (3.7)

Early

Only

2 (2.1)

4 (0.7) 10 (1.5)

79 (9.9)

7 (1.7)

11 (4.3) 5 (9.4)

4 (0.3) 7 (8.6)

Late

1 (1.0)

7 (1.3)

2 (0.3)

4 (0.9)

4 (1.6)

38 (3.1) 5 (6.2)

Early

3 (0.2) 7 (8.6)

Late

2 (2.1)

3 (0.5) 1 (0.2)

7 (0.9)

3 (0.7)

11 (4.3) 4 (7.5)

Earlier

Subculture of vented TSB

"Includes only those genera for which a positive blood culture was detected initially by unvented or vented TSB subculture. bAll positive blood cultures from a patient detected initially by subculture alone. CPat/ent's first positive blood culture detected sooner by this subculture.

96

551 659

Streptococci Other

Candida

795

2 (0.5)

4 (1.6) I (1.9}

258 53

411

42 (3.4} 13 (16)

Early

1239 81

Total °

Staphylococcus epidermidis

Qureus

Staphylococcus

Gram-positive bacteria

Other

Haemophilus influenzoe Pseudomonas

Gram-negative bacteria Enterobacteriaceae

Microorganism

Only b

Subculture of unvented TSB

Number {%) of patients

TABLE 2. Patients with Positive Blood Cultures Detected Initially by Early and Late Subculture

o

o

c~

C ,< t~

N.K. Henry and J.A. Washington II

110

Table 1. The early unvented TSB subculture detected twice (42) the number of

Haemophilus isolates detected by the late subculture (20). Among the 81 patients with H. influenzae bacteremia (Table 2), subcultures from unvented TSB blood culture bottles were responsible for detecting significantly more positive blood cultures than were those from vented TSB bottles (p < 0.05). The early subculture was the only subculture positive in 13 patients with H. influenzae bacteremia and was the first means of detection in an additional 15 patients. Late subculture was the only subculture positive in three patients and was the first means of detection in an additional eight patients with H. influenzae bacteremia. With other gram-negative bacteria (Table 1), subculture of unvented TSB bottles was not as likely to provide initial detection of Enterobacteriaceae as was that of the vented TSB bottle (p < 0.001). Positive blood cultures from patients with Enterobacteriaceae bacteremia (Table 2) were more often detected initially by subculture of vented TSB bottles (p < 0.05). There were no significant differences in the detection of Pseudomonas and other miscellaneous gram-negative or of gram-positive microorganisms. Candida was detected more often by subcultures from vented TSB bottles (p < 0.001; Table 1). The number of patients with candidemia detected by subculture was too small for statistical analysis. DISCUSSION

The role of the unvented blood culture bottle has been reexamined (Ilstrup and Washington, 1983; Tenney et el., 1982; L.B. Relier, S. Mirrett, and W.-L.L. Wang, Abstr Annu Meet Am Soc Microbiol, 1983, C88, p 326). Ilstrup and Washington (1983) recommend that an unrented TSB blood culture bottle be used with a vented blood culture bottle, as 27% of Bacteroidaceae, 44% of peptostreptococci, and 39% of Haemophilus would have been missed without the use of the unvented bottle. Anaerobic incubation of routine subcultures was determined to be of no value by Murray and Sontag (1978) and by Paisley et el. (1978), however, the question remains whether one should perform subcultures routinely from the unvented bottle. Pfaller et el. (1983) found that subculture of unvented thiol blood culture bottles was of limited value when subculture of vented TSB blood culture bottles was also performed. They cautioned, however, that they had few isolates of Pseudomonas or Haemophilus in their study. Our study examined the role of subcultures from unvented TSB bottles and included isolates of Haemophilus and Pseudomonas, which constituted 1.6 and 6.5% of the total isolates, respectively. The detection of positive blood cultures by subculture of unrented TSB bottles was comparable to that of vented TSB bottles for Pseudomonas and all other microorganisms except the Enterobacteriaceae, Candida, and Haemophilus. Bacteremia caused by members of the Enterobacteriaceae was detected earlier by subculture of vented TSB bottles, whereas that caused by H. influenzae was detected earlier by subculture from unvented TSB bottles. Our study showed that 46% of H. influenzae isolates were detected initially by subculture of the unvented TSB bottle and that the early subculture recovered 67% of these isolates. Although we concur with Pfaller et el. (1983) that routine subculture of unrented blood culture bottles is of limited value when routine subculture is performed from the vented blood culture bottle, we do recommend the former procedure whenever TSB is used for the detection of bacteremia in patients in whom H. influenzae infection is possible. Subcultures of TSB blood culture bottles provided initial detection of 30-40% of isolates, which reemphasizes the value of subcultures of blood culture bottles containing broth alone (Pfaller et al., 1982; Henry et el., 1984). Whether the yield from

Unvented Tryptic Soy Broth Subcultures

111

subculture onto a chocolate blood agar plate is comparable to that from the agar slant of a biphasic blood culture bottle remains uncertain; however, we have shown that a vented bottle containing TSB (Roche) with the SeptiChek slide attachment was superior to the routinely subcultured unvented TSB (Difco} bottle for the recovery of gram-negative aerobic and facultatively anaerobic bacilli (p < 0.01}, particularly for P. aeruginosa (p < 0.05) {Henry et al., 1984}. Thus far, our data show that the Roche bottle is comparable to the unvented TSB bottle for the recovery of Haemophffus, but the numbers are stilltoo small--Difco TSB, 9 of 9; Roche TSB, 7 of 9 {unpublished data}--for statisticalanalysis. Until a biphasic system is demonstrated to be equivalent in performance to a chocolate blood agar subculture plate incubated in 5-10% CO2 for the recovery of H. influenzae, the latterprocedure should be used. REFERENCES Dixon W], Massey F] ]r (1969) Introduction to Statistical Analysis. 3rd ed. New York: McGraw-Hill. Henry NK, Grewell CM, McLimans CA, Washington ]A II (1984) Comparison of the Roche SeptiChek blood culture bottle with a brain heart infusion biphasic medium bottle and with a tryptic soy broth bottle. ] Clin Microbiol (in press). Ilstrup DM, Washington ]A (1983) Effects of atmosphere of incubation on recovery of bacteria and yeasts from blood cultures in tryptic soy broth. Diagn Microbiol Infect Dis 1:215. Murray PR, Sontag ]E (1978) Evaluation of routine subcultures of macroscopically negative blood cultures for detection of anaerobes. ] Clin Microbiol 8:427. Paisley ]W, Rosenblatt ]E, Hall M, Washington ]AII (1978) Evaluation of a routine anaerobic subculture of blood cultures for detection of anaerobic bacteremia. ] C/in Microbiol 8:764. Pfaller MA, Sibley TK, Wesffall LM, Hoppe-Bauer ]E, Keating MA, Murray PR (1982) Clinical laboratory comparison of a slide blood culture system with a conventional broth system. ! C/in Microbiol 16:525. Pfaller MA, WesffaU LM, Murray PR (1983} Value of routine aerobic subculturing of unvented blood culture bottles. ] C/in Microbiol 17:601. Railer LB, Murray PR, MacLowry ]D (1982) Cumitech 1A, Blood Cultures II. Coordinating ed., ]A Washington II. Washington, DC: American Society for Microbiology. Tenney ]H, Reller LB, Mirrett S, Weinstein MP, Wang W-LL (1982) Controlled evaluation of the effect of atmosphere of incubation on detection of bacteremia and fnngemia in supplemented peptone broth. ] C/in Microbiol 16:437.