Detection of Yeast Septicemia by Biphasic and Radiometric

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, Apr. 1981, p /81/ $0.00/0 Vol. 13, No. 4 Detecton Yeast Septcema by Bphasc and Radometrc Methods ELENA PREVOST* AND EDWARD BANNISTER Dvson Clncal Mcrobology, Department Laboratory Medcne, Medcal Unversty South Carolna, Charleston, South Carolna 9403 Receved 14 August 1980/Accepted 1 October 1980 From January 19 to Aprl 1980 our mcrobology laboratory used a commercal bphasc bran heart nfuson vented culture method for fungal blood cultures and a commercal radometrc (BACTEC 460, Johnson Laboratores, Cockeysvlle, Md.) method for bacteral blood cultures. A total 668 bphasc fungal blood cultures were processed, whch 30 grew yeasts from 19 patents. There were 38,34 BACTEC blood cultures processed for bactera, whch 184 grew yeasts from 85 patents. The overall detecton tme for all yeasts averaged 8.3 days for the bphasc method and.4 days for the radometrc method. The BACTEC aerobc bottle detected over sx tmes as many yeasts as dd the anaerobc bottle. Candda albcans was the most frequently solated yeast n both methods, beng detected n an average. days n the bphasc method and 1.6 days n the aerobc BACTEC bottle. It s concluded from ths study that the radometrc method s far superor to the bphasc method, because () t has a shorter detecton tme, () t can be used smultaneously wth bacteral methods, savng blood and money, and () t requres no specal or separate meda or nstructons for yeasts, thus allevatng confuson n the blood collecton process. In the past 10 to 15 years, yeast septcema has been prmary concern n the management hosptalzed and compromsed patents (1, 5,, 13, 14, 19, 1, ). Blood cultures postve for yeasts, however, have not always concded wth actual nfecton. Koznn et al. (14) and Bodey (1) both found that blood cultures were postve n less than 50% dssemnated Candda nfectons. Conversely, postve blood cultures may not always reflect dssemnated nfectons, but rather may reflect transent canddema whch resolves upon removal a contamnated catheter (, 8). Further, t s known that persstent and ultmately fatal yeast nfectons are ten catheter nduced (19, ). A prospectve study by Rose reported that, 69 patents wth systemc Candda nfectons, 55 had venous catheter-assocated nfecton (19). In addton, a study by Currey and Que (5) reported that as many as 16% patents recevng hyperalmentaton developed canddema. Technques for solatng fung from blood have been a source controversy for several years (, 3, 9-11, 13, 1, 18). Komorowsk and Farmer (13) approached the problem tryng to mprove blood culture technques for detectng yeasts by usng a lyss-fltraton technque, whch was found to greatly enhance the recovery yeasts from blood. Ths method s currently recommended by Haley and Calaway (10) n a Centers for Dsease Control mycology manual. However, the method s cumbersome and rsks contamnaton, and most laboratores fnd t unsutable to ther stuaton. In 194, Gantz et al. (9) determned that commercal vacuumed blood culture bottles used n routne bacterologcal methods nhbted the growth Canda spp., but that ventng these bottles wth sterle cotton plugged needles shortened ncubaton tme and enhanced the recovery yeasts from blood specmens. In 195, ths observaton was confrmed by Braunsten and Tomasulo () n smulated blood cultures. Roberts and Washngton (18) descrbed a method usng 100-ml bran heart nfuson (BHI) Castaneda-type vented bottles whch yelded a sgnfcantly hgher re- 655 covery rate. A dsadvantage to ths method s that 100-ml BHI bphasc bottles are not commercally avalable, although a 50-ml sze s manufactured by BBL Mcrobology Systems (Cockeysvlle, Md.) and GIBCO Dagnostcs (Madson, Ws.). Caplan and Merz (3) found that postve yeast cultures were ncreased by 13% when two these 50-ml commercal bottles were used nstead a sngle bottle. Captolo et al. (C. Captolo, T. E. Kehn, and J. B. Mayo, Abstr. Annu. Meet. Am. Soc. Mcrobol, 1980, C68, p. 319) reported that commercal BHI bphasc

2 656 PREVOST AND BANNISTER bottles were superor to conventonal BHI aerobc vented blood culture bottles n the detecton yeasts. Recently, Hopfer et al. (1) reported an automated radometrc technque (BACTEC, Johnston Laboratores, Cockeysvle, Md.) for yeasts that was beng used for bacteral blood cultures. They found that all postve yeast cultures were detected wthn days. Yeasts n ther study were detected n an average.9 days overall. The BACTEC system s based on the nstrument's abty to detect mnor changes n radolabeled `4C levels when an organsm utlzes labeled substrates and produces a radolabeled boproduct. Bacteral growth can be detected by the BACTEC nstrument n as lttle as 4 h after collecton the blood culture (6, 0). Our laboratory has been usng a commercal bphasc blood culture method (two vented bottles per culture) for fungal cultures and the BACTEC 460 radometrc method for bacteral blood cultures for over 3 years. We have made a study these two systems wth respect to yeast detecton durng a 40-month perod (1 January 19 to 30 Aprl 1980). MATERIALS AND METHODS All bacteral blood cultures were collected n the usual manner by placng 3 to 5 ml blood each n the anaerobc and aerobc BACTEC bottles. After arrval n the laboratory the aerobc bottles were read by the BACTEC 460 nstrument and flushed wth fresh aerobc gases twce on the day collecton (day 0), twce on day 1, and once each on days, 3, 4, 5, and ; after each readng, the bottles were placed n a 3 C shakng ncubator. Anaerobc bottles were read once a day for days begnnng on day 1, flushed wth fresh anaerobc gases, and ncubated n a nonshakng 3 C ncubator. All aerobc and anaerobc bottles were routnely subcultured on days 1 and 5 and Gram staned on day 1. All postve machne readngs (-35 for aerobc bottles and -0 for anaerobc bottles) were Gram staned and subcultured both aerobcally and anaerobcally. AUl negatve cultures were reported as fnal on day. Yeasts or other fung detected and solated were taken to the mycology laboratory for complete dentfcaton. Genus and speces yeasts were determned by morphology on cornmeal-tween agar, by API-0C, and by other ancllary tests (gas producton, urea, ntrate reducton). Fungal blood cultures were collected n the same manner by placng 3 to 5 ml venous blood (or arteral blood f fungal endocardts was suspected) n each two 50-ml BHI bphasc bottles (BBL). Upon recept n the laboratory a sterle cotton-plugged needle was permanently nserted nto each bottle to release the vacuum and allow ventng. The bottles were ncubated at 5 to C and read daly for 30 days. Each day the agar surface was nspected for vsual growth and, f growth was present, worked up accordngly. bottles were gently shaken to J. CLIN. MICROBIOL. subculture the blood-broth mxture over the agar surface, and the bottles were rencubated. cultures were reported as fnal on day 30. RESULTS From 1 January 19 to 30 Aprl 1980 the bacterology laboratory processed 38,34 BAC- TEC blood cultures (more than 6,000 bottles, snce most cultures conssted an aerobc and an anaerobc bottle), whch 184 bottles grew yeasts from 85 patents: 91 Candda albcans, 38, C. parapsloss, 3 C. kruse, 19 Cryptococcus neormans, 8 C. (Torulopss) glabrata, and 3 Rhodotorula rubra. The aerobc BACTEC bottle detected 158 (86%) yeasts and the anaerobc bottle detected 6 (14%) yeasts (Table 1). In 16 pared postve BACTEC cultures (consstng one aerobc bottle and one anaerobc bottle) there were four nstances (.4%) when only the anaerobc bottle yelded the solate, nstances (13.6%) when both bottdes yelded the solate, and 136 (84%) nstances when only the aerobc bottle yelded the solate. Of the 184 yeasts detected by BACTEC, 151 (8%) were,, or C. parapsloss. The aerobc bottle detected these yeasts n a mean tme 1.66 days (range, 0 to 6 days); the anaerobc bottle requred.38 days (range, 0 to days). The mycology laboratory processed 668 fungal blood culture bottles (usually two bottles per culture) n the same tme span. Of these, 30 grew yeasts from 19 patents: 15, C. tropcals, 1 C. kruse, 4 C. parapsloss, 6 Cryptococcus neormans, and C. (Torulopss) glabrata (Table 1). Seventy percent these yeasts were,, or C. parapsloss and were detected n a mean tme.8 days (range, to 0 days). Overall detecton tme for all yeasts (Table 1) n the BACTEC aerobc bottle averaged.3 days, and that n the anaerobc bottle averaged.4 days. For the bphasc bottles overall detecton tme for all yeasts averaged 8.3 days. The longest detecton tme for BACTEC was for C. kruse, requrng days; for the bphasc bottle the longest tme was for and C. (Torulopss) glabrata, both whch requred 10 days. For C. kruse the detecton tme was smlar n both systems, requrng 6 days wth the bphasc bottle and days wth the aerobc BACTEC bottle. There were 53 blood cultures taken by both methods on the same day from 13 patents on 15 dfferent occasons (Table ). There were three nstances (patents, 5, and ) when the bphasc bottles were postve and the BACTEC bottles were negatve, fve nstances (patents 1, 3, 8 [nd specmen], 9 [nd spec-

3 VOL. 13, 1981 DETECTION OF YEAST SEPTICEMIA 65 TABLE 1. Results blood cultures taken by radometrc (BACTEC) and bphasc methods, 1 January 19 to 30 Aprl 1980 Bphasc BACTEC Yeast solate Aerobc Anaerobc No. Avg no. cultures days to detect No. Avg no. No Avg no. cultures to detect days cultures to detect days Candda albcans Candda tropcals Candda parapsloss Candda kruse Cryptococcus neormans oa Candda (Torulopss) glabrata Rhodotorula rubra Total a Ths culture came from a patent wth fatal dssemnated cryptococcoss shortly before death. The culture was drawn n early mornng, and the aerobc bottle was postve on the second readng on day 0 (day collecton). The anaerobc bottle (routnely read frst on day 1) s read on day 0 f the aerobc bottle s postve. men], and 11) when both methods were postve, and seven nstances (patents 4, 6, 8 [lst specmen], 9 [lst specmen], 10, 1, and 13) when only the BACTEC detected yeasts. Patents 4, 5, 8, 9, 11, and 13 were already known to have yeast septcema from one or more prevous postve BACTEC blood cultures (wth the same yeast). Of the 15 occasons when both methods were used, 1 (80%) were postve by BACTEC and 9 (60%) were postve by the bphasc method. Of the 4 cultures taken by BACTEC, 1 (0%) were postve n an average.6 days (range, 0 to days); the 9 cultures taken by the bphasc method, 13 (45%) were postve n an average 8.5 days (range, to 0 days). The BACTEC 460 nstrument faled on 1 occasons to detect yeast growth durng the - day ncubaton perod (Table 3). These yeasts were detected by blnd subculture on days 1 (10 solates) and 5 ( solates). DISCUSSION Successful clncal management and treatment canddoss and other fungal nfectons depend to a great extent on the stage the nfecton. The earler the physcan knows about a postve blood culture, the sooner he can plan a course acton (0, ). A bacteral blood culture that grows yeast may be one the frst sgnfcant fndngs an early or smolderng yeast septcema, transent fungema, or an unsuspected dssemnated nfecton. Unfortunately, one the man obstacles n the detecton yeast septcema s that fungema s usually only consdered after bacteral cultures or treatment have yelded no defmtve dagnoss or clncal mprovement. Indeed, ths s compounded when fungal blood cultures requre specal permsson or dfferent meda from bacteral cultures, as has been the recent trend (11, 1, 18). It s, no doubt, greatly desrable to physcans to be able to culture for both bactera and yeasts wth the same meda and rest assured that each organsm wll be detected wth much the same relablty. The extra blood and expense utlzed n two separate systems s evdent wth patent 11 (Table ), from whom was drawn three bacteral blood cultures (sx bottles) and fve fungal blood cultures, amountng to some 55 ml blood n 1 day (over $150). Smulated blood cultures have been used to demonstrate advantages one system over another (3, 11, 1). However, premature conclusons may be made regardng a system's clncal applcaton- based on results from smulated blood cultures (3). In realty, smulated blood cultures show lttle relatonshp to actual patent blood contanng a plethora antbotcs, antcancer drugs, or a host other varables that obvously affect the growth and growth rate mcroorgansms. As an example, one study (1) reported that ther homemade BHI bphasc bottles detected yeast n 1 to 4 days n smulated blood cultures and n.6 to 3. days n nne patent blood cultures. When Roberts and Washngton (18) evaluated ths system wth 3 postve patent blood cultures, the mean detecton tme rose to 5.3 days (range, to 14 days). We fnd that our use ths same system from a commercal source further ncreased the detecton tme to 8.3 days (range, to 0 days), makng ths partcular commercal system nadequate for the rapd detecton yeast septcema n our nsttuton. In our study many nstances yeast septcema (ncludng three nstances from two patents n Table ) had been

4 658 PREVOST AND BANNISTER J. CLIN. MICROBIOL. TABLE. Blood cultures from 13 patents drawn by both radometrc (BACTEC) and bphasc methods on the same day' BACTEC Bphasc Patent No. No. days Isolate days to Isolate to detect detect 1 C. kruse C. kruse 6 C. kruse 3 4C 5C 6 8C (lst specmen) 8 (nd specmen) 9C (lst specmen) 9 (nd specmen) 10 "lc 1 d 1 o C. parapslossb 13c 1 a Each BACTEC culture represents a set consstng an aerobc and an anaerobc bottle; bphasc cultures represent a sngle bottle. Patents lsted twce represent cultures drawn on dfferent dates. b Ths solate was thought to represent skn contamnaton (see text). ' Indcates that one or more prevous BACTEC cultures were postve wth same yeast. d Ths solate was not detected by the BACTEC 460 nstrument but was solated from the aerobc bottle from the day 1 blnd subculture detected by BACTEC wthn 48 h, the physcan had been notfed and had taken a specfc clncal course acton, and n some cases the patent had been dscharged before the bphasc bottle became postve. Of the three nstances when the bphasc botte was the only postve yeast culture on the day when both methods were used (Table ), one the patents (patent 5) had already been dagnosed by BACTEC from earler postve cul-

5 VOL. 13,1981 TABLE 3. Yeasts detected by subculture methods but not by the BACTEC 460 nstrument durng the -day ncubaton perod Day Patent No. No. ~~~~~~detecton Isolate by subculture (source)a A 1 1 (A) B 1 1 (A) C 1 C. (Torulopss) glabrata 5 (A) D 1 (A) E 4 1 (A) F 1 1 (A) G 4 C. (Torulopss) glabrata 5 (A) H 1 1 (AN) I 1 C. (Torulopss) glabrata 5 (A) J 1 5 (AN) Indcates only bottle found postve on subculture. A, Aerobc BACTEC bottle; AN, anaerobc BACTEC bottle. tures. The C. parapsloss found only n the bphasc bottle n patent was thought by hs physcan to represent skn contamnaton and not to be clncally sgnfcant; the patent expred from another mycotc nfecton (rhnocerebral zygomycoss) before further cultures could be drawn. In any case, hs sera dd not show antbodes to C. parapsloss antgen when specfcally tested for ths organsm at the Centers for Dsease Control. Patent, then, represented the only occason when the bphasc method was the only known postve source (two four bottles) and the sngle BACTEC culture was negatve. Al seven nstances when the BACTEC bottle was the only postve culture (Table ) were consdered sgnfcant fndngs by the attendng physcans. Durng our study blood cultures by these two methods t became clear that most the postve BACTEC yeast cultures were a serendptous fndng; vz., the physcan, suspectng a bacterema, had ordered bacteral blood cultures at the frst sgn the septcema and found that the etologcal agent was a yeast rather than a bacterum. Ths allowed a quck change n management plans, such as replacng or removng contamnated catheters or startng antfungal therapy early n the nfecton. On the other hand, t was found that many the postve bphasc fungal blood cultures were n severely compromsed patents who already had numerous postve yeast cultures from several other stes such as urne, wound, sputum, and throat. In other cases, the fungal bphasc blood cultures were ordered by the physcans only after beng notfed that the BACTEC bottles were postve for yeasts. DETECTION OF YEAST SEPTICEMIA 659 Routne blnd subcultures are not specfcally recommended by the manufacturer the BAC- TEC nstrument. However, others (4, 0) have prevously noted the need for blnd subcultures to ncrease the recovery certan bactera. Our study confrms a report by Rosner (0) that subcultures are also necessary to ncrease the recovery some yeast strans, snce although the nstrument dd not detect ther presence n 1 bottles wthn the days, they were readly solated by conventonal methods from the same bottles. Thus, the culture medum tself seems adequate, but there does not appear to be enough metabolc actvty n some yeast strans to produce a suffcent detectable amount labeled boproduct. A hgh recovery rate and rapd detecton tme are desrable n any culture system (0, ). Overall, the BACTEC system was found to detect yeasts an average more than three tmes faster than the bphasc fungal bottles. Others (10, 15, 16) have noted that the BACTEC detected yeasts more rapdly than conventonal bacteral methods. Although t s possble that the commercal source may have had some nhbtory effect on the growth yeasts n the bphasc bottles, we beleve that contnuous ncubator shakng and frequent flushng wth fresh aerobc gases n tbe aerobc bottle accounts for a faster growth rate than does shakng the bphasc bottle once a day. Perhaps contnuous shakng vented bphasc or vented conventonal bottles would also shorten the detecton tme for yeasts n those methods. We suspect that blood cultures have hstorcally been a poor ndcator yeast septcema manly because the method utzed, e.g., vacuumed or other conventonal bacteral bottles (1,, 1, ). In concluson, we beleve that the aerobc bottle n the BACTEC radometrc method s superor to other currently used methods to detect yeasts. Snce the aerobc bottle detected sx tmes as many yeasts as dd the anaerobc bottle, t s possble that n specal stuatons such as suspected Candda endocardts, the use two aerobc BACTEC bottles may ncrease the chances solatng yeast as the etologcal agent. The advantages usng the BACTEC method for detectng yeast n blood cultures are as follows: () rapd detecton tme s acheved, n most cases less than 48 h after collecton; () radometrc detecton systems are far more senstve to small amounts growth than s the human eye; () no specal handlng, ventng, or specal fungal request s requred; (v) t costs no more to the patent (n blood or money) or to the laboratory to do both bacteral and fungal cultures n the same medum smultaneously;

6 660 PREVOST AND BANNISTER and (v) the physcan can culture for both organsms wth the same order whether or not he has consdered yeast septcema as a possblty, thus ncreasng the frequency dagnosng yeast nfectons earler. ACKNOWLEDGMENTS We apprecate the advce and crtcsm John P. Manos n the preparaton the manuscrpt. We are also grateful to Patrca Gddens for searchng and retrevng blood culture records and to Jane Dngle for clercal assstance. J. CLIN. MICROBIOL. LITERATURE CITED 1. Bodey, G. P Fungal nfectons complcatng acute leukema. J. Chronc Ds. 19: Braunsten, H., and M. Tomasulo A quanttatve study the growth Candda albcans n vented and unvented blood culture bottles. Am. J. Cln. Pathol. 66: Caplan, L. M., and W. G. Merz Evaluaton two commercally prepared bphasc meda for recovery fung from blood. J. Cln. Mcrobol. 8: Caslow, M., P. D. Ellner, and T. E. Kehn Comparson the BACTEC system wth blnd subculture for the detecton bacterema. Appl. Mcrobol. 8: Curry, C. R., and P. G. Que Fungàl septcema n patents recevng parenteral hyperalmentaton. N. Engl. J. Med. 85: DeBlanc, H. J., Jr., F. DeLand, and H. N. Wagner, Jr Automated radometrc detecton bactera n,96 blood cultures. Apple. Mcrobol. : Ells, C. A., and M. L. Spvack The sgnfcance canddema. Ann. Intern. Med. 6: Ganes, J. D., and J. S. Remmngton. 19. Dssemnated canddass n surgcal patents. Surgery : Gantz, N. M., J. L. Swan, A. A. Mederos, and T. F. O'Bren Blood culture bottles nhbtng growth Candda and fosterng growth Bacterodes. Lancet : Haley, L., and C. S. Callaway Laboratory methods n medcal mycology, 4th ed., p CDC Publ Centers for Dsease Control, Atlanta. 11. Hopfer, R. L., K. Mlls, and D. Groschel Improved culture medum for the radometrc detecton yeasts. J. Cln. Mcrobol. 9: Hopfer, R. L., A. Orengo, S. Chestnut, and M. Wenglar Radometre detecton yeasts n blood cultures cancer patents. J. Cln. Mcrobol. 1: Komorowsk, R. A., and S. Farmer Rapd detecton canddema. Am. J. Cln. Pathol. 59: Koznn, P. J., R. S. Galen, et al The precptn test n systemc canddass. J. Am. Med. Assoc. 35: Randall, E. L Long-term evaluaton a system for radometrc detecton bacterema, p In D. Schlessnger (ed.), Mcrobology-195. Amercan Socety for Mcrobology, Washngton, D.C. 16. Renner, E. D., L. A. Gathrdge, and J. A. Washngton II Evaluaton radometrc system for detectng bacterema. Appl. Mcrobol. 6: Roberts, G. D., C. Horstmeer, M. Hall, and J. A. Washngton II Recovery yeast from vented blood culture bottles. J. Cln. Mcrobol. : Roberts, G. D., and J. A. Washngton I Detecton fung n blood cultures. J. Cln. Mcrobol. 1: Rose, H. D Venous catheter-assocated canddema. Am. J. Med. Sc. 5: Rosner, R Comparson macroscopc, mcroscopc, and radometrc examnatons clncal blood cultures n hypertonc meda. Appl. Mcrobol. 8: Sande, M. A., C. R. Bowman, and R. A. Calderone. 19. Expermental Candda albcans endocardts: characterzaton the dsease and response to therapy. Infect. Immun. 1: Seelg, M. S., C. P. Speth, P. J. Koznn, E. F. Ton, and C. L. Taschdjan Candda endocardts after cardac surgery. J. Thorac. Cardovasc. Surg. 65: Washngton, J. A., and P. K. W. Yu Radometrc method for detecton bacterema. Appl. Mcrobol. :