JOURNAL OF CLINICAL MICROBIOLOGY, June 1983, p. 1120-11260095-1137/83/061120-07$02.00/0

Vol. 17, No. 6

In Vitro Evaluation of the BACTEC Resin-Containing BloodCulture Bottle

SHARON M. SMITH1 AND ROBERT H. K. ENG2*Microbiology Section, Laboratory Service,1 and Infectious Disease Section, Medical Service,2 VeteransAdministration Medical Center, East Orange, New Jersey 07019, and Department ofMedicine, School of

Medicine, University of Medicine and Dentistry ofNew Jersey, Newark, New Jersey 071032

Received 25 January 1983/Accepted 11 March 1983

The BACTEC resin-containing blood culture bottle (16B; Johnston Labora-tories) was designed to bind antibiotics and to support bacterial growth. However,only modest increases in the recovery rates of pathogens have been found inclinical studies. This in vitro study evaluated the ability of the 16B to recoverorganisms from human serum containing clinically achievable concentrations ofantibiotics. Escherichia coli, Enterobacter cloacae, Klebsiella pneumoniae, Pseu-domonas aeruginosa, Haemophilus influenzae, Staphylococcus aureus, Strepto-coccus faecalis, and a viridans streptococcus were added to serum containingantibiotic(s), and at various time intervals of antibiotic exposure, portions wereremoved and inoculated into both the 16B and the 6B (conventional aerobic)bottles. The studies of the kinetics of killing of the bacterial strains by the variousantibiotics showed a good correlation between those combinations of bacteria-antibiotic(s) which produced slow killing and the combinations of bacteria-antibiotic(s) which were recovered preferentially in the 16B bottles. Low recoveryrates were noted when the antibiotics killed the organisms rapidly. The indicationsfor use of the resin-containing blood culture bottle should be limited to thosesituations in which the patient is receiving antibiotics and the bacteremia issuspected to involve a pathogen which is killed slowly by the administered drug(s)or when the bacteremia is continuous. The failure of the BACTEC 16B bloodculture bottle to recover organisms may in part reflect the bactericidal activity ofthe antibiotics administered.

In the seriously ill patient, antimicrobialagents may be given before the completion of alldiagnostic tests, including the collection ofblood cultures. Often a patient may becomefebrile while on antimicrobial therapy and bloodcultures need to be obtained. In both of theseclinical situations, isolation of a blood-bornepathogen is desirable and should help in theclinical management of the patient. Unfortunate-ly, the blood cultures obtained more often thannot do not yield a viable organism(s) by theconventional methods, possibly due to the pres-ence of antibiotics in the blood. With the in-creasing number of available antibiotics and thewidespread use of broad-spectrum antimicrobialdrugs, a device is needed to remove such agentsfrom the blood sample during or before theincubation of the blood culture bottles. Such adevice should increase the recovery rate of thesepathogens.The first available antibiotic-binding device

was the Antimicrobial Removal Device (ARD;Marion Scientific, Kansas City, Mo.). The ARDcontains cationic and anionic resins, SPS, and

saline (2, 5). After the addition of blood, theARD bottle is rotated for 15 min to enhanceantibiotic removal by the resins. Then the bloodis removed from the resins and inoculated intoconventional blood culture bottles and incubat-ed. The second device is the BACTEC resin-containing blood culture bottle (16B) from John-ston Laboratories, Towson, Md. The 16B bottleis a combination of the aerobic blood culturebottle (tryptic soy broth and other supplements),a nonionic adsorbing resin (13.3% [wt/vol]), anda cationic exchange resin (0.8% [wt/vol]) (1).

Clinical trials with the ARD and the BACTECresin-containing blood culture bottle have beendisappointing in that although increased recov-ery rates have been reported in some studies, amarked increase in the recovery rate of orga-nisms has not been observed (1, 2, 5, 6) (R. P.Gruninger, M. L. Simpson, R. E. Klicker, andM. Y. Khan, Program Abstr. Intersci. Conf.Antimicrob. Agents Chemother. 22nd, MiamiBeach, Fla., abstr. no. 312, 1982; P. E. Por-embski, G. W. White, M. D. Batt, and R. E.Lee, 22nd ICAAC, abstr. no. 411; A. C.

1120

on April 16, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

EVALUATION OF BACTEC 16B BLOOD CULTURE BOTTLE 1121

Sonnewirth and A. Weissfeld, Abstr. Annu.Meet. Am. Soc. Microbiol. 1982, C121, p. 291;and P. Yungbluth, E. Aqui, and H. M. Som-mers, 22nd ICAAC, abstr. no. 410). The studieshave indicated that the recovery rate of gram-negative rods was very low and that the primaryorganisms recovered were the staphylococci.Information given in these reports did not indi-cate the time or dose of antimicrobial agentsadministered or the antibiotic concentrations inthe serum at the time the blood cultures wereobtained.

In our institution, the resin-containing 16Bblood culture bottles have been used only forpatients on antimicrobial therapy and are sub-mitted with the standard BACTEC aerobic andanaerobic blood culture bottles. Our results forthe 16B bottles have been similarly disappoint-ing in that no significant increase in the yield ascompared with the standard aerobic bottle hasbeen observed.The failure to enhance the recovery of orga-

nisms with the ARD and the resin-containing16B bottles presents the need for a systematicevaluation of the antibiotic-binding devices topossibly narrow the indications for their use.This study was designed to evaluate the antibiot-ic-binding ability of the BACTEC 16B bottle, thekinetics of killing of the clinical isolates whichare commonly encountered in clinical practice,and the ability of the 16B to recover theseorganisms after exposure to an antimicrobialagent(s) in human serum in usually achievableblood concentrations.

MATERIALS AND METHODS

Organisms. Organisms used were clinical isolatesfrom blood cultures obtained from the Clinical Micro-biology Laboratory of the Veterans AdministrationMedical Center, East Orange, N.J. Isolates were fro-zen at -70°C and subcultured onto blood or chocolateagar plates (BBL Microbiology Systems, Cockeys-ville, Md.) before testing. Gram-negative rods wereidentified by the API 20E (Analytab Products, Plain-view, N.Y.). Pseudomonas aeruginosa was identifiedby growth at 42°C and pigment production; Staphylo-coccus aureus by production of coagulase; Haemophi-lus influenzae by Haemophilus Quad plate (Carr Scar-borough, Decatur, Ga.); Streptococcusfaecalis by salttolerance, esculin hydrolysis, growth on bile-contain-ing agar, and by the API 20S Streptococcus strip; andthe viridans streptococcus by hemolysis and failure togrow on medium containing bile or 6.5% NaCl orfailure to hydrolyze esculin.

Antibiotics. The following laboratory antibiotic testpowders were obtained from their respective distribu-tors: cefoperazone (Pfizer Laboratories Division, NewYork, N.Y.); cephalothin, tobramycin, and moxalac-tam (Eli Lilly & Co., Indianapolis, Ind.); gentamicin(Schering Corp., Kenilworth, N.J.); cefotaxime(Hoechst-Roussel Pharmaceuticals Inc., Sommerville,N.J.); ampicillin and nafcillin (Bristol Laboratories,

Syracuse, N.Y.); and ticarcillin (Beecham Labora-tories, Bristol, Tenn.). All powders were stored dessi-cated and frozen at -25°C until use.MICs and MBCs. The minimum inhibitory concen-

trations (MICs) were determined by the microdilutiontechnique. Serial twofold dilutions of antibiotics weremade in Mueller-Hinton broth (Difco Laboratories,Detroit, Mich.) in sterile microtiter trays (Flow Labo-ratories, McLean, Va.) by using microdiluters. Aninoculum was added to yield a final concentration of105 organisms per ml from a 4-h log-phase culture. ForH. influenzae, Schaedler broth supplemented with 5%Fildes (BBL) was used. Schaedler broth was also usedin the dilutions of antibiotic and organism for theviridans streptococcus. The microtiter plates weregently mixed and incubated at 35°C. After 16 to 18 h,the MICs were determined (lowest concentration withno visible growth). The minimum bactericidal concen-trations (MBCs) were determined by subculturing 10RI from the well contents onto Mueller-Hinton agar orchocolate agar plates, and the criterion of 99.9% killwas used. S. aureus ATCC 25923, E. coli ATCC25922, and P. aeruginosa ATCC 27853 were includedas control organisms.

Binding experiments. The resin-containing bloodculture bottles (16B) were examined for the ability ofthe resins within the bottles to bind various antibioticsafter inoculation and during incubation. Antibioticswere added to 3 ml of human serum (Flow Labora-tories) to make a final concentration of 25 jig ofampicillin, cephalothin, cefoperazone, cefotaxime,moxalactam, or nafcillin per ml, 5 ,ug of gentamicin ortobramycin per ml, or 120 ,g of ticarcillin per ml. Theserum was heat inactivated at 56°C for 30 min beforeuse. The serum was added immediately to the 16Bresin-containing bottle. The bottles were either rotatedat 35°C in a water bath at 200 cycles per min orincubated at a stationary position. After a 15- or 30-min rotation or after a 15-min stationary incubation, 1ml of Mueller-Hinton broth containing 150 to 300 CFUof E. coli or S. aureus per ml was added to each bottle.The number of CFU per milliliter was equivalent to thetotal number of organisms inoculated per bottle in laterexperiments. E. coli was used for all bottles containingantibiotics except for bottles containing nafcillin towhich S. aureus was added. Bottles which were agitat-ed before the addition of organisms were again incu-bated at 35°C in a water bath shaker (American OpticalCorp., Buffalo, N.Y.) at 200 cycles per min, andbottles which were incubated at a stationary positionbefore the addition of organisms were placed in a 35°Cincubator. After 18 h, all bottles were observed forvisible growth and subcultured onto blood agar platesfor the detection of viable organisms.

Recovery of organisms during simulated blood cul-ture conditions from the resin-containing (16B) bottleversus recovery from the conventional aerobic (6B)bottle. Organisms from a 4-h culture were diluted andadded to serum to yield a final concentration of 50 to100 organisms per ml. Two portions of serum werethen removed and each inoculated into either resin-containing (16B) or conventional aerobic (6B) bloodculture bottles to check for viability and recovery oforganisms not exposed to antibiotics. The antibiotics(25 pg/ml concentration for all antibiotics except ticar-cillin [120 pg/ml] and the aminoglycosides [5 pg/ml])were added to the serum containing the organisms and

VOL. 17, 1983

on April 16, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

TABLE 1. MICs for the organisms studied

MIC (,ug/ml) of:Organism Ampi- Cepha- Cefo- Cefoper- Moxa- Genta- Tobra- Ticar- Naf-

cillin lothin taxime azone lactam micin mycin cillin cillin

E. coli 4 8 S0.10 '0.10 0.15 0.5 0.5 4 NDaK. pneumoniae >128 8 '0.10 0.15 0.15 0.5 ND ND NDE. cloacae >128 >128 '0.10 0.30 0.30 0.5 ND ND NDH. influenzae 0.25 ND 0.03 ND 0.07 ND ND ND NDP. aeruginosa ND ND ND ND ND ND 1 16 NDS. aureus ND ND ND ND ND 0.5 ND ND 0.5S. faecalis 1 ND ND ND ND 2 ND ND NDViridans strep- <0.06 ND ND ND ND 0.25 ND ND ND

tococcus

a ND, Not determined.

mixed. Portions (3 ml) were immediately removed andinoculated into the 16B and the 6B blood culturebottles (time 0). At various time intervals, 3-ml por-tions of this serum containing antibiotics and orga-nisms were used to inoculate both types of bottles. Forserum containing gram-negative rods and the viridansstreptococcus, the intervals were 0, 15, and 30 min, 1,2, and 3 h, and for other gram-positive cocci were 0,0.5, 1, 2, 4, and 6 h. All blood culture bottles inoculat-ed with H. influenzae were supplemented with 5%Fildes. Bottles were gently mixed several times toensure mixing of the serum, antibiotic, bacteria, andthe antibiotic-binding resins. All bottles were incubat-ed in a stationary position at 35°C for a total of 7 daysunless growth was evident earlier. All bottles werechecked daily for visible growth and were subculturedonto blood or chocolate agar plates at 16 to 18 h and 3,and 7 days after inoculation.

Kinetics of killing. The number of viable organismswas determined at various time intervals of exposureto antimicrobial agents by the standard method. Theconcentrations of antibiotics studied included 25 ,ug ofampicillin, cephalothin, cefotaxime, moxalactam,cefoperazone, and nafcillin per ml, 5 ,ug of gentamicinand tobramycin per ml, or 120 ,ug of ticarcillin per ml.The final volume was 5.0 ml of Mueller-Hinton brothcontaining 5 x 105 organisms of a log-phase culture perml. For H. influenzae, 5% Fildes supplement wasadded to the Mueller-Hinton broth. At various timeintervals, portions were removed and quantitativelycultured by dilution and subcultured onto blood orchocolate agar plates.

RESULTSThe MICs of the antibiotics tested and the

respective organisms studied are listed in Table1. The MICs (c0.3 jxg/ml) of the third-genera-tion cephalosporins were low for the organismsstudied. The strain of S. faecalis studied wassusceptible to gentamicin (2 ,ug/ml). All MBCswere within one twofold dilution of the MICswith the exception of S. faecalis and ampicillin.The ability of the resins to bind various antibi-

otics was studied by adding antibiotics to serumat concentrations which are clinically achiev-able. Portions (3 ml) of serum containing a study

antibiotic or antibiotic combination were addedto the resin-containing bottles. Bottles were theneither rotated for 15 or 30 min or left stationaryto allow for antibiotic binding to the resins. E.coli or S. aureus was then added to the bottles.If the organisms grew, then sufficient quantitiesof the antibiotics to which the organisms weresusceptible were assumed to have been bound tothe resins. The results of this experiment areshown in Table 2. Whether the bottles wererotated or held stationary did not affect therecovery of the test strains for ampicillin, cepha-lothin, cefoperazone, gentamicin, tobramycin,or nafcillin. However, in the 16B bottles, suffi-cient amounts of ticarcillin and moxalactamwere still available to kill the strain of E. colitested. Bottles which were left stationary afterthe addition of cefotaxime were able to supportbacterial growth; however, when the bottleswere rotated no growth was detected. Thesesame findings were observed on repeated trials.

Various organisms were added to serum con-taining clinically achievable concentrations of anantibiotic or a combination of antibiotics. Theabilities of both the BACTEC resin-containing16B bottle and the conventional aerobic 6Bbottle to recover the added organisms after thevarious time intervals of antibiotic exposureshown in Table 3. The maximum time interval ofantibiotic exposure studied for the gram-nega-tive rods was 3 h and for the S. aureus and S.faecalis was 6 h. The conventional aerobic bot-tles failed to recover most organisms even whenthe mixture of antibiotic-organism-serum wascultured in these bottles immediately after theaddition of antibiotics (time 0). The test strainsof S. faecalis and the viridans streptococcuswere recovered in both the 16B and the 6B bottleeven after a prolonged exposure to ampicillinalone. For all test strains exposed to cefotaximeand cefoperazone, viable organisms were recov-ered from the 16B bottles up to the maximumantibiotic exposure time studied (3 h); however,

1122 SMITH AND ENG J. CLIN. MICROBIOL.

on April 16, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

EVALUATION OF BACTEC 16B BLOOD CULTURE BOTTLE 1123

TABLE 2. Ability of the resin-containing bloodculture bottle to support growth of E. coli and S.aureus after preinoculation with an antibiotic

Amt Incubation conditionAntibiotic ( Lg/bottle) Organism Stationary' Rotatedb

Ampicillin 75 E. coli +C +Cephalothin 75 E. coli + +Cefotaxime 75 E. coli + _dCefoperazone 75 E. coli + +Moxalactam 75 E. coli _Gentamicin 15 E. coli + +Nafcillin 75 S. aureus + +Tobramycin 15 E. coli + +Ticarcillin 360 E. coli - -

a Bottle kept stationary after the addition of antibi-otic, after inoculation, and during incubation.

b Bottle gently rotated at 200 cycles per min afterthe addition of antibiotic, after inoculation, and duringincubation.

c +, Growth.d -, No growth.

these same organisms were rarely recoveredafter exposure to moxalactam. The time intervalof antibiotic exposure after which the organismswere recovered was less when antibiotic combi-nations to which the organisms were susceptiblewere employed. For the combination of ampicil-lin and gentamicin, the viridans streptococcuswas not recovered by the 16B beyond 15 min ofantibiotic exposure as compared to recoveryafter 3 h of exposure to ampicillin alone. S.faecalis could not be recovered after 1 h ofexposure to the drug combination as comparedto recovery after 6 h of exposure to ampicillinalone. Markedly decreased recovery rates werealso noted for the gram-negative rods after just abrief exposure to antibiotic combinations, andoften only those 16B bottles inoculated at time 0were able to recover viable organisms.

Bacterial growth of antibiotic-damaged orga-nisms in the 16B resin-containing blood culturebottle was detected generally within 18 to 24 h ofinoculation by visual observation, subculture forall gram-negative rods except H. influenzae (re-quired 48 h of incubation for detectable growth),or both. For S. aureus exposed to nafcillinalone, growth was detected 18 h after inocula-tion; however, after 2 h of exposure to nafcillinand gentamicin, 4 days were required for thedetection of growth, whereas with shorter peri-ods of antibiotic exposure, organisms were de-tected within 18 h of inoculation. For S. faecalis,all growth was detected after the initial 18-hincubation in both types of bottles. For theviridans streptococcus, 48 h of incubation wasrequired for detectable growth in the 16B bottleafter exposure to ampicillin and gentamicin. Ingeneral, whenever both the conventional aero-

bic (6B) and the resin-containing (16B) bloodculture bottles yielded viable organisms, growthwas detected in both types of bottles on thesame day.The rates of killing of organisms by individual

or combinations of antibiotics are shown in Fig.1. In general, the rate of killing for the third-generation cephalosporins was relatively slow.Only a 1 to 2 log decrease in CFU was notedafter 3 h of exposure of Enterobacter cloacae,Klebsiella pneumoniae, and H. influenzae tomoxalactam, cefotaxime, and cefoperazone.The combinations of ampicillin and gentamicinand of cephalothin and gentamicin decreased thenumber of viable organisms by 5 logs within 1 hfor all gram-negative rods studied. A 3 logdecrease in 1 h was observed for P. aeruginosaafter exposure to ticarcillin (120 ,ug/ml) andtobramycin (5 ,ug/ml), and no viable organismswere detected after 4 h. After a 3-h exposure ofthe viridans streptococcus and H. influenzae toampicillin at a concentration of 25 ,ug/ml, theviable counts decreased approximately 1 log.After a 6-h exposure of S. faecalis to ampicillin,a decrease of only 1 log was noted. The combi-nations of nafcillin and gentamicin or ampicillinand gentamicin markedly decreased the numberof organisms recovered as compared to nafcillinor ampicillin alone for S. aureus, S. faecalis, andthe viridans streptococcus.

DISCUSSIONIn the critically ill patient with a fever and

suspected bacteremia, various antibiotics orcombinations of antibiotics are often adminis-tered before the collection of appropriate cul-tures. When the identity of the pathogen isunknown, therapy must be broad spectrumenough to include activity for all the organismswhich are suspected. In large or tertiary carefacilities, these organisms may include relativelyresistant enteric gram-negative rods, staphylo-cocci, and P. aeruginosa, which may dictate theuse of certain antimicrobial agents such asmoxalactam or antibiotic combinations such asticarcillin and tobramycin.

Recently, two devices have been made avail-able which contain resins and are reported tobind the antibiotics in the blood samples sent forculture. The ARD and the BACTEC resin-con-taining blood culture bottle (16B) have beendesigned to and should aid in the recovery oforganisms from patients on antimicrobial thera-py. The ARD has been evaluated by severalinstitutions; however, its success as measuredby an increased recovery rate of organisms hasbeen modest (3, 5, 6) (Gruninger et al., 22ndICAAC, abstr. no. 412; Porembski et al., 22ndICAAC, abstr. no. 411; Yungbluth et al., 22ndICAAC, abstr. no. 410). The organisms most

VOL. 17, 1983

on April 16, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

TABLE 3. Bacterial recoveries in the resin-containing (16B) and the conventional aerobic (6B) blood culturebottles after exposure in serum to antibiotics

Organism Antibiotics' Length of Time of last recoverystudy (h) 16B 6B

E. coli Ampicillin + gentamicin 3 0 min 0 minCephalothin + gentamicin 3 0 min 0 minCefoperazone 3 3 h NonebCefotaxime 3 3 h NoneMoxalactam 3 None None

K. pneumoniae Cephalothin + gentamicin 3 0 min 0 minCefoperazone 3 3 h NoneCefotaxime 3 3 h NoneMoxalactam 3 None None

E. cloacae Cephalothin + gentamicin 3 30 min 15 minCefoperazone 3 3 h NoneCefotaxime 3 3 h NoneMoxalactam 3 None None

H. influenzae Ampicillin 3 3 h NoneCefotaxime 3 3 h NoneMoxalactam 3 15 min None

P. aeruginosa Tobramycin + ticarcillin 3 1 h None

S. aureus Nafcillin 6 4 h NoneNafcillin + gentamicin 6 2 h None

S. faecalis Ampicillin 6 6 h 2 hAmpicillin + gentamicin 6 1 h 0 min

Viridans streptococcus Ampicillin 3 3 h 3 hAmpicillin + gentamicin 3 15 min None

a All cell wall antibiotics were used at a serum concentration of 25 ,ug/ml except for ticarcillin which was usedat 120 jLg/ml. Gentamicin and tobramycin were used at a concentration of 5 ,ug/ml.

b None, no viable organisms detected at any time interval sampled, including the sample taken immediatelyafter the addition of antibiotics (0 min).

often recovered by the ARD bottles have beenthe gram-positive cocci, especially the staphylo-cocci. A multicenter study has also been con-ducted evaluating the BACTEC resin-containingblood culture bottle (1). In this study, the addi-tion of this bottle to the standard two-bottleBACTEC system has increased the recoveryrate of pathogens from patients receiving antimi-crobial therapy. Again the increased recoverywas noted mainly for the gram-positive cocci.Because of the difficulty in obtaining clinicaldata, a majority of the studies were not able toindicate the type of antimicrobial therapy in-volved, the blood concentration of the antimi-crobial agents, or the MICs of the organismsrecovered.An adequate detailed clinical evaluation of the

effectiveness of the 16B bottle or the ARD isdifficult because the recovery rate for pathogensis very low for all currently available bloodculture systems. The failure of a device toincrease the recovery of certain organisms may

not be due to the failure of the device but mayreflect the rapid killing of the blood-borne patho-gen by the antimicrobial agents administeredbefore processing of the blood sample by thedevice. In this in vitro study, many of thevariable factors of a clinical study were con-trolled. Commonly isolated pathogens were ex-posed in serum to various commonly used anti-biotics, and the ability of the BACTEC resin-containing blood culture bottle to recover theseorganisms was evaluated and compared to therecovery rate with the conventional aerobicBACTEC blood culture bottle. The resins boundadequate quantities of all antibiotics tested withthe exception of moxalactam and possibly ticar-cillin and allowed for the growth of the bacteria.Partial binding of ticarcillin may explain thegrowth of P. aeruginosa, which has an MIC of16 ,ug/ml, and the no growth of E. coli, whichhas a much lower MIC of 4 jig/ml.

Failure to recover certain organisms afterexposure to antibiotics or combinations of anti-

1124 SMITH AND ENG J. CLIN. MICROBIOL.

on April 16, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

EVALUATION OF BACTEC 16B BLOOD CULTURE BOTTLE 1125

HOURS

6

I 5111.4L 4

20 30-J

2

8

7

-j6

I.U.,20

3

C

- Kx

0.25 0.5 I

2

32HOURS

0.25 0.5 2 3

HOURS

*D

I-0.

0.5 -I2 6

HOURS

IIL

00-j

0.5 2

8

7

6

5~

4

3

2

4HOURS

-F

2 4 6

HOURS

-G7

6

-i

0Li 4

- 30

0

0.5 3HOURS

-H

0.5HOURS

2 3

FIG. 1. Kinetics of antibiotic-induced killing for the clinical isolates studied. Clinical isolates ofE. coli (A), K.pneumoniae (B), E. cloacae (C), P. aeruginosa (D), S. aureus (E), S. faecalis (F), a viridans streptococcus (G),and H. influenzae (H) were used. Symbols: 0-*, control, no antibiotics; x-x, ampicillin; O-O, moxalac-tam; O-O, cefoperazone; *-----, cefotaxime; x---x, cephalothin and gentamicin (also ampicillin andgentamicin for E. coli); 0---0, ampicillin and gentamicin; O---O, ticarcillin and tobramycin; O---O0, nafcillin;* ---*, gentamicin and nafcillin.

VOL. 17, 1983

6

5!4

3

2

IL

co0-J

IL.

'o0-i

-J

a

U1.

2

0-i

a

7

6

5

4

3

2

Ep.

14

-B

0! a

"X,

x

7

8

7

6

4

3

2

0

ill

on April 16, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

1126 SMITH AND ENG

biotics with both the resin-containing and theconventional bottles was consistently observed.When the kinetics of killing of the various agentswere studied, the decrease in the ability torecover organisms after antibiotic exposurefrom the 16B bottles correlated well with therate of decrease in the number of viable orga-nisms in the kill curve. E. coli, K. pneumoniae,and E. cloacae were not recovered after a 30-min exposure to ampicillin and gentamicin orcephalothin and gentamicin. The killing kineticsfor these antibiotics and organisms also indicat-ed that these organisms were rapidly killed. Thecombination of ticarcillin and tobramycin alsoproduced a rapid decrease in the number ofviable P. aeruginosa, and the 16B bottle recov-ery of this organism was not possible after theorganisms were incubated in this antibiotic com-bination for greater than 60 min.For antibiotics which killed organisms more

slowly (cefotaxime, moxalactam, and cefopera-zone), increased recovery of the organisms inthe resin-containing bottle was consistently not-ed. The recovery of the gram-negative rods wasgreatly increased even when susceptible orga-nisms were exposed to relatively high serumlevels for up to 3 h. For gram-positive orga-nisms, S. aureus, S. faecalis, and the viridansstreptococcus, increased recovery was noted forthe antibiotic combinations of ampicillin andgentamicin or nafcillin and gentamicin as com-pared to the conventional bottle. These data areconsistent with results of clinical studies (1, 2, 5,6).Many factors are important in the successful

isolation of a blood-borne pathogen (4). Both theresin-containing 16B and the conventional 6Baerobic blood culture bottles result in the samefinal dilution (approximately 1:10) of the bloodsample and, hence, the contained antibiotics.For organisms which have relatively high MICsof the administered antibiotics or slow killing orfor antibiotics with limited stability in serum andblood, a 1:10 dilution alone by the 6B bottle maybe enough to permit bacterial growth. Althoughthe role of SPS cannot be discounted, equalconcentrations (0.025%) were contained in bothtypes of bottles. The rate of shedding of theorganism into the bloodstream of the host (tran-sient versus continuous) and the timing and rateof infusion of antibiotics may also be determin-

ing factors in the recovery of organisms fromblood cultures. The length of the incubation ofthe blood cultures will also affect the recoveryrate offastidious and slow-growing organisms aswell as possibly antibiotic-exposed organisms.

In conclusion, the data presented suggest thatthe BACTEC 16B resin-containing blood culturebottle has the potential to increase the recoveryrate of antibiotic-exposed microorganisms.However, not all antibiotics are bound or inacti-vated by the contents of this culture bottle (e.g.,moxalactam and ticarcillin). The antibiotic con-centration in the serum at the time of bloodcollection may also influence the success ratefor the recovery of the pathogen. The clinicianand the laboratory personnel must also be awarethat organisms which are killed more slowly aremore likely to be recovered from the resin-containing bottles, whereas those organismswhich are rapidly killed by appropriate therapyare less likely to be recovered. Awareness ofthese factors should render the indications foruse of the 16B BACTEC system more specificand the interpretation of the culture results moreunderstandable.

ACKNOWLEDGMENTS

We thank Rodney Broman of Johnston Laboratories for hisinterest and for supplying the BACTEC blood culture bottlesused in this study.

This study was supported in part by the General MedicalResearch Fund of the East Orange Veterans AdministrationMedical Center.

LITERATURE CITED

1. Applebaum, P. C., D. G. Beckwith, J. R. Dispersio, J. W.Dyke, J. F. Salventi, and L. L. Stone. 1982. Enhanceddetection of bacteremia with a new BACTEC resin bloodculture medium. J. Clin. Microbiol. 17:48-51.

2. Appleman, M. D., R. S. Swinney, and P. N. R. Heseltine.1982. Evaluation of the antibiotic removal device. J. Clin.Microbiol. 15:278-281.

3. Lindsey, N. J., and P. E. Riely. 1981. In vitro antibioticremoval and bacterial recovery from blood with an antibi-otic removal device. J. Clin. Microbiol. 13:503-507.

4. Reller, L. B., P. R. Murray, and J. D. MacLowry. 1982.Cumitech 1A, Blood cultures. Coordinating ed., J. A.Washington II. American Society for Microbiology, Wash-ington, D.C.

5. Wallis, C., J. L. Melnick, R. D. Wende, and P. E. Riely.1980. Rapid isolation of bacteria from septicemic patientsby use of an antimicrobial agent removal device. J. Clin.Microbiol. 11:462-464.

6. Wright, A. J., R. L. Thompson, C. A. McLimans, W. R.Wilson, and J. A. Washington II. 1982. The antimicrobialremoval device a microbiological and clinical evaluation.Am. J. Clin. Pathol. 78:173-177.

J. CLIN. MICROBIOL.

on April 16, 2020 by guest

http://jcm.asm

.org/D

ownloaded from