JOURNAL OF CLINICAL MICROBIOLOGY, Feb. 1988, p. 250-256 0095-1137/88/020250-07$02.00/0 Cpyright 1988, American Sciety fr Micrbilgy Vl. 26, N. 2 Effects f Medium and Inculum Variatins n Screening fr High-Level Aminglycside Resistance in Entercccus faecalis DANIEL F. SAHMl* AND CARMEN TORRES2 Clinical Micrbilgy Labratries and Department f Pathlgy, University f Chicag Medical Center, Chicag, Illinis 60637-1463,' and Hspital Ramn y Cajal, Servici de Micrbilgia, Madrid, Spain2 Received 27 July 1987/Accepted 28 Octber 1987 Entercccus faecalis islates that are refractry t aminglycside-penicillin synergy can be detected by their ability t grw in the presence f high cncentratins f aminglycside (2,000,g/ml). In past studies investigatrs have used a variety f media and inculum sies t perfrm high-level aminglycside resistance screens, but little is knwn abut hw these variatins affect test accuracy. We screened 63 E. faecalis strains n different media by using varius inculum sies and crrelated the results with synergy test results btained by time-kill studies. Screens were dne with dextrse-phsphate agar, brain heart infusin agar, Trypticase sy agar with 5% sheep bld, Mueller-Hintn agar with 5% sheep bld, dextrse-phsphate brth, and Mueller-Hintn brth. Agar screens were inculated with 102, 104, and 106 CFU; and brth screens cntained a final inculum f 105 CFU/ml. The E. faecalis islates were tested fr high-level resistance t streptmycin, kanamycin, amikacin, gentamicin, and tbramycin. Of the 63 islates tested, 21 did nt shw high-level resistance t any f the aminglycsides tested, and 42 demnstrated high-level resistance t ne r mre drugs. The sensitivity f mst screens was -90%. Regardless f the inculum sie r medium used, false-resistance results were seldm encuntered. Screen specificity, which was used as the indicatr f false susceptibility, was markedly influenced by bth the inculum sie and the drug being tested. Specificity was lw whenever a 102-CFU inculum was used, when amikacin was tested with any inculum, and when tbramycin was tested in brth media. Data fr kanamycin culd be used t predict amikacin-penicillin synergy, and the highly accurate gentamicin screen bviated the need fr the testing f tbramycin. We recmmend a 106-CFU inculum fr agar screens and a 105-CFU/ml inculum fr brth screens. The type f medium used did nt substantially influence screen accuracy. Amng the aminglycsides, nly streptmycin, gentamicin, and, ccasinally, kanamycin need t be used t screen E. faecalis islates fr aminglycside-penicillin synergy. The synergistic antientercccal activity that aminglycside and penicillin cmbinatins exhibit has been knwn fr sme time (7, 11, 14). Use f these antibitics is the mst effective treatment fr serius entercccal infectins, and they are recmmended as the treatment f chice (8, 12, 17-19). Enterccci can acquire high-level resistance t aminglycsides (i.e., resistant t >2,000,ug/ml), hwever, and becme refractry t penicillin-aminglycside synergy (1, 2, 6, 9, 15, 16, 20, 22). High-level resistance has been reprted fr every aminglycside that might be cnsidered fr cmbinatin therapy; these include streptmycin, kanamycin, amikacin, tbramycin, gentamicin, and netilmicin (1, 2, 6, 9, 12, 15, 16, 20, 22). The emergence f these resistant strains can significantly limit the therapeutic chices fr serius entercccal infectins. Because high-level aminglycside resistance is mst frequently mediated by cnstitutively prduced aminglycside-mdifying enymes that are encded n transferable plasmids, the incidence f enterccci that are refractry t cmbinatin therapy is expected t increase (3, 10, 13, 15, 25). Indeed, reprts by Zervs et al. (24) have indicated that as much as 55% f their entercccal islates demnstrated high-level gentamicin resistance. The prbability that synergy-resistant enterccci will emerge in many health care facilities, cupled with the serius cnsequences that this resistance might have fr the therapeutic management f patients, indicates that perhaps the mst relevant suscepti- * Crrespnding authr. 250 bility testing methds fr enterccci are thse that predict the susceptibility f the rganism t synergy. The tw mst cmmn methds fr determining synergy, the checkerbard technique and time-kill tests, are ften t cumbersme, time-cnsuming, and labr intensive fr rutine use in many clinical labratries. Because entercccal resistance t synergy is mst frequently mediated by resistance t high levels f aminglycsides, hwever, a simple screening test riginally described by Mellering et al. (16) can be used t predict the susceptibility f Entercccus faecalis islates t synergy. By this methd, an agar medium is supplemented with a particular aminglycside t a final cncentratin f 2,000,ug/ml, and E. faecalis islates are inculated nt the agar surface. Strains that grw are exhibiting high-level aminglycside resistance and, therefre, wuld likely be refractry t drug synergy. On the ther hand, strains that fail t grw are prbably susceptible t aminglycside-penicillin synergy. Althugh several investigatrs have used tests based n the principle f this high-level aminglycside resistance screen, the type f media and sie f the inculum emplyed have varied cnsiderably. The different media used have included Trypticase sy yeast brth (BBL Micrbilgy Systems, Cckeysville, Md.) (22), catin-supplemented Mueller-Hintn brth (MBH) (9), Trypticase sy brth (8), dextrse-phsphate agar (DPA) (2, 16), 5% defibrinated hrse bld agar (1), brain heart infusin agar (BHI) (15, 20), and nutrient agar (18). Similarly, the inculum sies have varied frm 104 t 106 CFU/ml (8, 9, 15, 20, 22), and sme investigatrs (2, 16, 18) have inculated test agars simply by
VOL. 26, 1988 streaking a single entercccal clny ver the surface withut quantitating the number f bacteria that were present in the inculum. The effect that these test variatins have n the accuracy f the screen fr predicting entercccal resistance t synergy has nt been thrughly examined. The present study was undertaken t determine the effect that variatin in the type f medium and inculum sie wuld have n the ability f the screen t detect high-level aminglycside resistance accurately and, therefre, t predict susceptibility t aminglycside-penicillin synergy. The infrmatin gained frm this investigatin culd be useful fr establishing the mst reliable cnditins fr perfrming the screen. MATERIALS AND METHODS Organisms. Fr this study we used 63 strains f E. faecalis. Fifty-six f the islates were btained frm patient specimens submitted t the University f Chicag Clinical Micrbilgy Labratries, and the ther seven strains were graciusly prvided by Steven R. Crider (Naval Hspital, Camp Pendletn, Calif.). We cnfirmed the identify f each E. faecalis islate n the basis f clny mrphlgy, Gram stain, and recmmended bichemical characteristics (5). Antibitics and media. The antibitics used fr this investigatin included gentamicin (Schering Crp., Blmfield, N.J.), tbramycin (Eli Lilly & C., Indianaplis, Ind.), streptmycin and penicillin (Sigma Chemical C., St. Luis, M.), and kanamycin and amikacin (Bristl Labratries, Syracuse, N.Y.). Mueller-Hintn agar (MHA) and MHB were btained frm BBL Micrbilgy Systems. All ther media, which included Trypticase sy agar (TSA), DPA, BHI, and dextrse-phsphate brth (DPB), were prvided by GIBCO Labratries, Madisn, Wis. MHA and TSA were the nly media that were supplemented with 5% sheep bld. High-level aminglycside resistance screen. Fr the highlevel aminglycside resistance screen, each aminglycside was incrprated int each agar (MHA, TSA, DPA, BHI) and brth (DPB, MHB) medium t a final cncentratin f 2,000,ug f drug per ml. Incula fr the agar screens were prepared frm vernight (18 t 24 h) E. faecalis cultures that had been grwn in DPB at 35 C. By using a Steers replicatr, prtins f each vernight culture, which cntained apprximately 109 CFU/ml, were inculated nt the varius agars t give a final inculum f 106 CFU. Subsequently, the remainder f each culture was diluted with sterile 0.85% saline t match the turbidity f a 0.5 McFarland standard (ca. 108 CFU/ml). Frm this suspensin apprpriate dilutins were made, and a Steers replicatr was used t inculate the agar media with the 102 and 104 CFU inculum f each E. faecalis strain. Media that lacked an aminglycside supplement were used as psitive grwth cntrls. All media were incubated fr 18 t 24 h in an ambient atmsphere at 35 C. Fllwing incubatin the agar screens were examined fr the presence f bacterial grwth; if bacterial grwth was evident, the relative amunt (i.e., light t heavy) was nted. Fr the brth (i.e., DPB and MHB) screening methd, the inculum was prepared in saline frm cultures grwn vernight in DPB, as was dne fr the agar screen incula. Hwever, the final inculum was 105 CFU in 1.0 ml f brth. Media that were nt supplemented with an aminglycside were used as psitive grwth cntrls. Brths were incubated in an ambient atmsphere at 35 C and, fllwing 18 t 24 h f incubatin, were examined visually fr the presence f bacterial grwth. SYNERGY SCREENS FOR E. FAECALIS 251 Time-kill studies. T perfrm time-kill studies with each E. faecalis islate that was screened fr high-level aminglycside resistance, we used a methd based n a prcedure described previusly by Mellering et al. (17). After the rganisms were grwn in DPB vernight at 35 C, the turbidity f the brth cultures was adjusted t equal a 0.5 McFarland standard. A 0.1-ml prtin f the vernight culture was added t 0.9 ml f fresh DPB t give a final rganism cncentratin f 107 CFU/ml. Prir t inculatin, each tube f fresh DPB was supplemented with the apprpriate aminglycside, either alne r in cmbinatin with penicillin. A DPB tube that did nt cntain antibitic was used as a psitive grwth cntrl. The cncentratin f antibitic that was used clsely apprximated the maximum achievable serum level f each drug in serum and was less than the MIC fr the E. faecalis strains tested. The drug cncentratins used were as fllws: streptmycin, 25,ug/ml; amikacin, 20,ug/ml; kanamycin, 20,ug/ml; gentamicin, 5,ug/ml; tbramycin, 5,ug/ml; penicillin G, 10 U/ml. Inculated brths were incubated in an ambient atmsphere at 35 C. At 0-, 4-, and 24-h intervals after inculatin, a 0.01-ml prtin was remved frm each f the DPB tubes and diluted and 0.1 ml f diluent was plated n TSA. By using the viable cunts determined at each time interval, a 24-h time-kill curve was established fr each E. faecalis strain. Susceptibility t aminglycside-penicillin synergy was defined as a.100-fld increase in killing by the drug cmbinatin ver the killing accmplished by the mst active f the tw drugs when tested separately. Resistance t synergy was a <100-fld increase in killing (18). Data analysis. Results f the high-level resistance screens were crrelated with results f the time-kill synergy studies. Based n the crrelatins, screen results were classified as true psitive, true negative, false psitive, r false negative. True psitive results were thse in which bth the screen test and the time-kill test shwed an islate t be synergy susceptible. True negative results were thse in which an islate was synergy resistant by bth tests. Results in which an islate was synergy susceptible by the screen test but resistant by the time-kill test were cnsidered t be false psitive, and thse that were synergy resistant by the screen test but susceptible by the time-kill methd were cnsidered t be false negative. The specificities and sensitivities f each f the varius high-level aminglycside resistance screens were calculated by using the fllwing frmulas: sensitivity (%) = [number f true psitives/(number f true psitives + number f false negatives)] x 100; specificity (%) = [number f true negatives/(number f true negatives + number f false psitives)] x 100. By these frmulas, sensitivity indicated the incidence f false resistance bserved with the varius screens, and specificity indicated the incidence f false susceptibility. RESULTS Examples f time-kill curves generated fr tw E. faecalis islates are given in Fig. 1. Results btained with a strain that was susceptible t the synergistic activity f every aminglycside-penicillin cmbinatin tested are shwn in Fig. 1A, and results btained with a strain that was resistant t every antibitic cmbinatin tested are shwn in Fig. 1B. Such time-kill plts were generated fr each E. faecalis islate and were used as the standard by which islates were characteried as being susceptible r resistant t the synergy f varius aminglycside-penicillin cmbinatins.
252 SAHM AND TORRES J. CLIN. MICROBIOL. 0 A a-- 0 -J B -j c,) - :D CD -J ILL c1 C) TIME (HOURS) -J C) TIME (HOURS) FIG. 1. Effect f penicillin and each aminglycside, tested alne and in cmbinatin, n the grwth f a E. faecalis strain that was susceptible t synergy between penicillin and each aminglycside tested (A) and a E. faecalis strain that was resistant t synergy between penicillin and each aminglycside tested (B). The fllwing antibitics were tested: penicillin (PC), 10 U/ml; streptmycin (ST), 25,ug/ml; amikacin (AK), 20,ug/ml; kanamycin (KN), 20,ug/ml; gentamicin (GM), 5,ug/ml; tbramycin (TB), 5 kg/ml; cntrl (C), n antibitics. All time-kill studies were cnducted by using DPB. Of the 63 E. faecalis islates selected fr this study, 21 were susceptible t every drug cmbinatin tested, and 42 strains exhibited resistance t ne r mre aminglycsidepenicillin cmbinatins (Table 1). Amng these 42 islates, tw were resistant nly t streptmycin-penicillin synergy, whereas the ther 40 were resistant t mre than ne aminglycside-penicillin cmbinatin. All 40 f these multiply-resistant islates shwed kanamycin-penicillin and amikacin-penicillin resistance, 33 shwed streptmycin-penicillin resistance, and 22 shwed bth gentamicin-penicillin and tbramycin-penicillin resistance. The effects that varius agar media and inculum sies had n the specificities f the high-level resistance screens are given in Table 2. Regardless f the aminglycside tested r the agar medium emplyed, use f the 102 inculum frequently resulted in specificities that were well belw 90% and several that were less than 80%. Extremely lw specificities were als bserved with all f the amikacin screens. N matter which agar medium r inculum sie was used fr these screens, the specificity never exceeded 23%. When kanamycin was substituted fr amikacin t predict amikacinpenicillin synergy, hwever, the specificity increased sub- TABLE 1. Aminglycside resistance patterns f E. faecalis islates N.tf Aminglycside resistance pattern f: Streptmycin...... 2 Streptmycin, kanamycin, amikacin... 18 Streptmycin, kanamycin, amikacin, gentamicin, tbramycin... 15 Kanamycin, amikacin, gentamicin, tbramycin... 7 Susceptibility t all...... 21 strains a As determined by time-kill studies (Fig. 1). TABLE 2. Specificity f high-level aminglycside resistance screens tested with varius agar media and inculum sies Specificity (%) n the fllwing Aminglycsideb (CFU) agar media: DPA BHI TSA MHA Streptmycin 102 100 93 79 82 104 100 100 100 97 106 100 100 100 97 Gentamicin 102 90 81 76 76 104 95 95 90 95 106 95 95 95 95 Tbramycin 102 82 72 81 50 104 91 91 91 88 106 95 95 91 91 Kanamycin 102 90 69 78 79 104 100 97 97 100 Amikacin 102 17 10 21 17 104 17 15 10 17 106 22 23 20 21 Kanamycin-amikacinc 102 91 68 83 81 104 100 96 97 100 a The frmula used t calculate specificity (incidence f false susceptibility) is given in the text. b Each aminglycside screen was tested at 2,000 p.g/mi. c Fr kanamycin-amikacin, kanamycin was used in place f amikacin t predict amikacin-penicillin synergy.
VOL. 26, 1988 stantially. This increase was especially evident with the use f 104- and 10'-CFU incula, which resulted in specificities f.96%. Similarly, the specificities f the ther aminglycside screens were frequently.95% when either 104- r 106-CFU incula were used. In general, when all f the aminglycside screens are cnsidered, the specificities btained with the 106-CFU inculum tended t be slightly higher than the specificities btained with 104-CFU inculum. The sensitivities btained with varius aminglycside screens are given in Table 3. Regardless f the medium and inculum sie used, streptmycin, gentamicin, tbramycin, and amikacin screens ali had high sensitivities, with mst being 100%. In general, the sensitivity f the kanamycin screens was als high. Only when a 106-CFU inculum was tested n BHI did kanamycin sensitivity fall t a relatively lw 80%. Fr the screens in which kanamycin was used in place f amikacin t predict amikacin-penicillin synergy, the sensitivities were -95% with 102- and 104-CFU incula, but were lwer with the 106-CFU inculum. Mst ntable was the 77% sensitivity btained with the use f BHI. When the E. faecalis islates were screened fr high-level resistance, sme f the tests resulted in very light grwth at the site f inculatin. These instances f light grwth, which were characteried by the presence f '-10 individual clnies r a fine film f grwth, were categried as questinable screen results. Instead f including these results in the data fr calculating the specificities and sensitivities given in Tables 2 and 3, respectively, these questinable results were tabulated accrding t bth the agar medium used and the aminglycside tested and were crrelated with resuits f the time-kill synergy studies (Table 4). Because f the lw specificities assciated with all f the amikacin Sensitivity f high-level aminglycside resistance TABLE 3. screens tested with varius agar media and inculum sies Aminglycsidel Inculum ide"(cfu) Sensitivity (%) n the fllwing agar media: DPA BHI TSA MHA Streptmycin 102 100 96 100 100 104 100 96 100 100 106 96 96 100 96 Gentamicin 102 100 100 100 100 Tbramycin 102 100 100 100 100 Kanamycin 102 100 100 100 100 106 94 80 95 95 Amikacin 102 100 100 100 100 Kanamycin-amikacinc 102 100 100 100 100 104 100 96 100 95 106 94 77 90 90 a The frmula used t calculate sensitivity (incidence f false resistance) is given in the text. b Each aminglycside screen was tested at 2,000,ug/ml. C Fr kanamycin-amikacin, kanamycin was used in place f amikacin t predict amikacin-penicillin synergy. SYNERGY SCREENS FOR E. FAECALIS 253 TABLE 4. Number f questinable screen results btained with the use f varius aminglycsides and agar media Ttals by Questinable results n the Aminglycside fllwing agar mediaa: amin- DPA BHI TSA MHA glycside Streptmycin 0/0 0/0 3/4 0/0 3/4 Gentamicin 0/1 0/1 2/3 2/2 4/7 Tbramycin 1/1 1/1 1/1 6/6 9/9 Kanamycin 1/6 1/3 5/8 3/6 10/23 Ttals by medium 2/8 2/5 11/16 11/14 26/43b a Questinable screen results were defined as '10 clnies per spt inculum r very light hae f grwth. Results btained with amikacin and with the 102-CFU inculum were excluded. Results are expressed as the number f questinable results btained with synergy-resistant strains/ttal number f questinable results. bthe ttal number f questinable results was 1.7% f all results. screens and with the screens inculated with 102 CFU (Table 2), the questinable results btained with these tests were nt included in Table 4. Of the 2,520 remaining screen results, nly 43 (1.7%) were questinable. Of these 43 questinable results, 26 ccurred with strains that were characteried as being synergy resistant t particular aminglycside-penicillin cmbinatins, and the remaining 17 ccurred with synergy-susceptible strains. The kanamycin screens mst frequently resulted in light grwth, and amng the agar media, questinable results ccurred mst ften with TSA and MHA. The sensitivities and specificities f the brth screens fr high-level aminglycside resistance are given in Table 5. Because agar screen results shwed that incula f 104 t 106 CFU generally resulted in high specificities and sensitivities (Tables 2 and 3, respectively), and because 105 CFU/ml is the final inculum recmmended fr brth dilutin susceptibility testing (21), the brth screens were tested with a single inculum sie f 105 CFU/ml. With the exceptin f the streptmycin screen results btained in DPB, the sensitivities f these tests were 100%. Hwever, the specificities f the screens varied cnsiderably with the aminglycside being tested. Whereas streptmycin, gentamicin, kanamycin, and kanamycin substituted fr amikacin synergy screens demnstrated specificities abve 90%, the specificities f the amikacin and tbramycin screens were cnsiderably lwer, regardless f the brth that was used. The lw specificities f the amikacin screens perfrmed in brth were cnsistent TABLE 5. Sensitivities and specificitiesa f high-level aminglycside resistance screens tested in brth mediab Results n the fllwing brth media: Aminglycside DPB MHB Sensitivity Specificity Sensitivity Specificity M% (%) (%) M% Streptmycin 96 94 100 91 Gentamicin 100 95 100 95 Kanamycin 100 95 100 97 Tbramycin 100 82 100 43 Amikacin 100 10 100 7 Kanamycin-amikacinc 100 97 100 100 a Frmulas used t calculate sensitivity (incidence f false resistance) and specificity (incidence f false susceptibility) are given in the text. b Final inculum fr all tests, 105 CFU/ml. C Fr kanamycin-amikacin, kanamycin was used in place f amikacin t predict amikacin-penicillin synergy.
254 SAHM AND TORRES with the lw specificities nted with varius agar media (Table 2). In cntrast, the specificities f the brth screens fr tbramycin were much lwer than the specificities btained with agar media (Table 2). The 43% specificity f the tbramycin screen in MHB resulted frm the fact that 13 truly resistant E. faecalis islates failed t grw in the brth screen. These islates were retested by using higher incula (106 and 107 CFU/ml). With the 106-CFU/mnl inculum, nne f the 13 islates grew, but all strains grew when the inculum was i07 CFU/ml. Hwever, with the 107-CFU/ml inculum, several strains that were susceptible t tbramycin-penicillin synergy als grew. A similar attempt t increase the specificity f the tbramycin screen in DPB by increasing the inculum sie als resulted in a number f false-resistant results. DISCUSSION As indicated by the high sensitivities given in Table 3, very few instances f false resistance ccurred when varius agar media and inculum sies were used t test E. faecalis islates fr synergy resistance. Exceptins were nted when a 106-CFU inculum was used in kanamycin screens fr determining bth kanamycin-penicillin synergy and fr amikacin-penicillin synergy (kanamycin substituted fr amikacin), particularly when BHI was emplyed. Because the same inculum sie did nt result in lw sensitivities fr the ther aminglycside screens, regardless f the agar medium that was used, the explanatin fr the false resistance bserved with kanamycin screens n BHI is unclear. In any çase, ur results indicate that if BHI is used, the sensitivities f the kanamycin screens may be significantly influenced by the inculum sie. In cntrast, sensitivities f the ther aminglycside synergy screens are nt ntably affected by the agar medium r inculum sie. In cntrast t the sensitivities f the screens, the specificit.es were substantially influenced by the inculum sie (Table 2). Use f 102 CFU frequently resulted in falsesusceptible test results (i.e., lw specificity). Such results indicate that ineffective aminglycside-penicillin cmbinatins culd be used fr the treatment f serius E. faecalis infectins. Therefre, the 102-CFU inculum sie must be cnsidered inadequate fr synergy screening. Althugh gd specificities were achieved with bth 104- and 106-CFU incula, slightly fewer false-susceptible results were nted with the larger inculum. Regardless f the inculum sie r the agar medium used, the amikacin screens shwed extremely lw specificities (Table 2). The high incidence f false susceptibility that we bserved agrees with results reprted by Basker et al. (1) and Calderwd et al. (2). Results f their studies shwed that sme E. faecalis strains failed t grw in the presence f 2,000 ptg f amikacin per ml, but when tested in time-kill studies, the strains were refractry t amikacin-penicillin synergy. Why these strains are inhibited by 2,000,ug f amikacin per ml but are nt synergistically killed by the cmbinatin f penicillin and serum-achievable cncentratins f amikacin is nt knwn. In any case, as demnstrated previusly (1, 2) and in ur study, kanamycin can be used in place f amikacin t accurately predict amikacin-penicillin synergy (Tables 2, 3, and 5). Tw majr factrs are respnsible. First, E. faecalis resistance t kanamycin and amikacin is mediated by either ne f tw aminglycside-mdifying enymes, 3'-phsphtransferase-III r the bifunctinal enyme 2"-phsphtransferase-6'-acetyltransferase (3, 10, 13). Therefre, because any kanamycin-resistant strain is als amikacin-resistant, the sensitivity f the kanamycin J. CLIN. MICROBIOL. screen fr amikacin-penicillin synergy shuld be cmparable t the sensitivity bserved fr kanamycin-penicillin synergy, as shwn by the results given in Table 3. Secnd, in cntrast t the 2,000-,ug amikacin screen, mst strains that are susceptible t 2,000,ug f kanamycin per ml are synergistically killed by the kanamycin-penicillin cmbinatin, as well as by the amikacin-penicillin cmbinatin. Thus, with the kanamycin screen fr amikacin-penicillin synergy, few instances f false susceptibility wuld be expected and the specificities shuld be high (Table 2). Except fr the amikacin screens and thse perfrmed with the 102-CFU inculum, the specificities and sensitivities f mst agar screens were high (Tables 2 and 3). Of 2,520 tests that were perfrmed, in nly 43 (1.7%) was questinable grwth shwn (Table 4). Thus, results f mst high-level screens perfrmed n agar media demnstrate either synergy resistance r susceptibility. Hwever, when questinable grwth des ccur, neither the resistance nr the susceptibility f the islate can be presumed. Organisms that give such results shuld be tested by the time-kill methd. Mst questinable results ccurred with TSA and MHA (Table 4), but they represented nly 3% f the screen results btained with these tw media. Similarly, mst questinable results ccurred when kanamycin was tested, but the 23 questinable results cnsisted f nly 5% f the results btained with this particular aminglycside. These data indicate that TSA and MHA and kanamycin are acceptable fr reliable synergy screening. Except fr amikacin and tbramycin, the sensitivities and specificities f the aminglycside resistance screens perfrmed in brth media were high (Table 5). The same mechanism that was respnsible fr the lw specificities in the agar screen was prbably respnsible fr the lw specificities bserved fr the brth amikacin tests. As was true fr the agar screens, kanamycin can be used in the brth methd t predict amikacin-penicillin synergy. The reasn fr the high number f false-susceptible results bserved with the tbramycin screens is unknwn. An increase in the inculum t i07 CFU/ml resulted in grwth f the truly resistant strains that failed t grw when a 105-CFU/ml inculum was used, but use f the higher inculum increased the incidence f false resistance seen amng synergy-susceptible strains. The prblems encuntered with the tbramycin brth screens culd be eliminated by using gentamicin t screen fr tbramycin-penicillin synergy. Because the bifunctinal aminglycside-mdifying enyme (2"-phsphtransferase- 6 -acetyltransferase) that mediates high-level resistance t gentamicin is the enyme that als mediates tbramycin resistance (3), islates that are resistant t gentamicinpenicillin synergy are als resistant t tbramycin-penicillin synergy. Therefre, gentamicin screens, which shwed high specificities and sensitivities in bth agar and brth media (Tables 2, 3, and 5), culd be used t predict nt nly gentamicin-penicillin synergy but als tbramycin-penicillin synergy. Of imprtance is the fact that because the brth screens were perfrmed in 1-ml vlumes, the results btained might be different frm thse that may ccur with the use f a brth micrdilutin system. Studies t investigate the accuracy f micrdilutin prcedures fr detecting synergy resistance are needed. The findings f this study indicate that if labratries are ging t rutinely screen clinically significant E. faecalis islates fr resistance t aminglycside-penicillin synergy, they shuld fllw certain guidelines and take specific precautins. With respect t inculum sie, use f an insufficient inculum (i.e., 102 CFU) can have an adverse effect n
VOL. 26, 1988 the specificity f the agar screens. Therefre, we d nt recmmend the use f unquantifiable inculatin prcedures such as the streaking f a single clny ver the agar surface f an antibitic-cntaining plate. Because the fewest instances f false-susceptible results were btained with 106 CFU, this inculum sie is recmmended fr agar screens. With the 106 CFU inculum, the instances f false-resistant screen results may be increased when kanamycin is tested n BHI. Hwever, use f an agar medium ther than BHI (DPA, TSA, MHA) wuld alleviate this prblem. Fr synergy screens perfrmed in brth media, we recmmend a final inculum sie f 105 CFU/ml, the same as that which is used fr ther brth dilutin susceptibility testing prcedures (21). Because the specificities and sensitivities f the synergy screens were nt ntably affected by the type f agar r brth medium used, except when a 106-CFU inculuni was used t test kanamycin n BHI, in general, the medium that is available may be used. In many clinical micrbilgy labratries, this wuld be TSA, MHA, r MHB. The aminglycsides selected fr screening depend n the drugs that are being cnsidered fr therapy and the antibitics that give the mst reliable screen results. The aminglycsides that might be cnsidered fr cmbinatin therapy with a cell wall-active agent are streptmycin, kanamycin, amikacin, gentamicin, tbramycin, and netilmicin. Amng these drugs, nly streptmycin, gentamicin, and ccasinally, kanamycin shuld be cnsidered fr synergy screening. Regardless f whether E. faecalis resistance t streptmycin is mediated by ribsmal mutatin (4, 25) r enymatic inactivatin (13), nly streptmycin can be used t screen fr streptmycin-penicillin synergy. Because the same bifunctinal enyme (i.e., 2"-phsphtransferase-6'-acetyltransferase) that mediates high-level gentamicin resistance als mediates tbramycin and netilmicin resistance (3), a gentamicin screen alne currently is sufficient t reliably predict synergy between penicillin and gentamicin, tbramycin, r netilmicin. In additin, this enyme als mediates resistance t kanamycin and amikacin s that gentamicinresistant E. faecalis islates are als resistant t these tw aminglycsides (3). Sme E. faecalis islates are susceptible t gentamicin (L.e., lack 2"-phsphtransferase-6'-acetyltransferase), hwever, but are resistant t kanamycin and amikacin (Table 1). These islates prduce 3'-phsphtransferase-III, an enyme that inactivates kanamycin and amikacin but nt gentamicin (13). Therefre, althugh all gentamicin-resistant islates are als resistant t kanamycin and amikacin, nt all gentamicin-susceptible islates are susceptible t kanamycin and amikacin. The susceptibilities f these gentamicin-susceptible strains t kanamycin and am'ikacin can be determined nly by a screen fr high-level kanamycin resistance. Finally, we must emphasie that the results f this study pertain nly t E. faecalis islates and nt t the less cmmnly encuntered Entercccus faecium. E. faecium' is resistant t mst aminglycside-penicillin cmbinatins, and is ften susceptible nly t gentamicin-penicillin synergy. High-level resistance screens perfrmed with these rganisms frequently d nt crrelate with results f synergy studies (23). Further studies are needed t establish reliable screening prcedures that may be used t test E. faecium susceptibility t aminglycside-penicillin synergy. LITERATURE CITED 1. Basker, M. J., B. Sicmbe, and R. Sutherland. 1977. Aminglycside-resistant enterccci. J. Clin. Pathl. 30:375-380. SYNERGY SCREENS FOR E. FAECALIS 255 2. Calderwd, S. A., C. Wennersten, R. C. Mellerihg Jr., L. J. Kun, and D. J. Krgstad. 1977. Resistance t six aminglycside amincyclitl antibitics amng enterccci: prevalence, evlutin, and relatinship t synergism with penicillin. Antimicrb. Agents Chemther. 12:401-405. 3. Curvalin, P., C. Carlier, and E. Cllat. 1980. Plasmid-mediated resistance t amincyclitl antibitics in grup D streptccci. J. Bacteril. 143:541-551. 4. Elipuls, G. M., B. F. Farber, B. E. Murray, C. Wennersten, and R. C. Mellering, Jr. 1984. Ribsmal resistance f clinical entercccal islates t streptmycin. Antimicrb. Agents Chemther. 25:398-399. 5. Facklam, R. R., and R. B. Carey. 1985. Streptccci and aerccci, p. 154-175. In E. H. Lennette, A. Balws, W. J. Hausler, Jr., and H. J. Shadmy (ed.), Manual f clinical micrbilgy, 4th ed. American Sciety fr Micrbilgy, Washingtn, D.C. 6. Hrdniceanu, T., L. Bugueleret, N. El-Slh, G. Bieth, and F. Delbs. 1979. High-Level, plasmid-brne resistance t gentamicin in Streptcccus faecalis subsp. ymgenes. Antimicrb. Agents Chemther. 16:686-689. 7. Hunter, T. H. 1947. Use f stréptmycin in the treatment f bacterial endcarditis. Am. J. Med. 2:436-442. 8. lannini, P. B., J. Ehret, and T. C. Eickff. 1976. Effects f ampicillin-amikacin and ampicillin-rifampin n enterccci. Antimicrb. Agents Chemther. 9:448-451. 9. Ikeda, D. P., A. L. Barry, and S. G. Andersn. 1984. Emergence f Streptcccus faecalis islates with high-level resistance t multiple amincyclitl aminglycsides. Diagn. Micrbil. Infect. Dis. 2:171-177. 10. Jacb, A. E., and S. J. Hbbs. 1974. Cnjugal transfer f plasmid-brne multiple antibitic resistance in Streptcccus faecalis var. ymgenes. J. Bacteril. 117:360-372. 11. Jawet, E., J. B. Gunnisn, and V. R. Cleman. 1950. The cmbined actin f penicillin and streptmycin r chlraniyycetin n enterccci in vitr. Science 111:254-256. 12. Kreniski, O. M., C. Wennersten, R. C. Mellering, Jr., and M. A. Sande. 1978. Penicillin-netilmicin synergism against Streptcccus faecalis. Antimicrb. Agents Chemther. 13: 430-434. r 13. Krgstad, D. J., T. R. Krfhagen, R. C. Mellering, Jr., C. Wennèrsten, and M. N. Swart. 1978. Aminglycside-inactivating enymes in clinical islates f Streptcccus faecalis. J. Clin. Invest. 62:480-486. 14. Mandell, G. L., D. Kaye, M. E. Levisn, and E. W. Hk. 1970. Entercccal endcarditis. Arch. Intern. Med. 125:258-264. 15. Mederski-SamraJi, B. D., and B. E. Murray. 1983. High-level resistance t gentamicin in clinical islates f enterccci. J. Infect. Dis. 147:751-757. 16. Mellering, R. C., Jr., C. Wennersten, T. Medrek; and A. N. Weinberg. 1971. Prevalence f high-level resistance t aminglycsides in clinical islates f enterccci, p. 335-340. Antimicrb. Agents Chemther. 1970. 17. Mellering, R. C., Jr., C. Wennersten, and A. N. Weinberg. 1971. Studies n antibitic synergism against enterccci. I. Bacterilgical studies. J. Lab. Clin. Med. 77:821428. 18. Mellering, R. C., Jr., C. Wennersten, and A. N. Weinberg. 1971. Synergy f penicillin and gentamicin against enterccci. J. Infect. Dis. 124(Suppl.):S207-S209. 19. Mellering, R. C., Jr., C. Wennersten, and A. J. Weinstein. 1973. Penicillin-tbramycin synergism against enterccci: a cmparisn with penicillin and gentamicin. Antimicrb. Agents Chemther. 3:526-529. 20. Murray, B. E., J. Tsa, and J. Panida. 1983. Enterccci frm Bangkk, Thailand, with high-level resistance t currently available aminglycsides. Antimicrb. Agents Chemther. 23:799-802. 21. Natinal Cmmittee fr Cliiical Labratry Standards. 1985. Methds fr dilutin antimicrbial susceptibility tests fr bacteria that grw aerbically; apprved standard. Publicatin M7-A. Natinal Cmmittee fr Clinical Labratry Standards, Villanva, Pa. 22. Standifrd, H. D., J. B. demaine, and W. M. M. Kirby. 1970.
256 SAHM AND TORRES Antibitic synergism f enterccci. Arch. Intern. Med. 126: 255-259. 23. Wennersten, C. B., and R. C. Mellering, Jr. 1980. Mechanism f resistance t penicillin-aminglycside synergism in Streptcccus faecium, p. 710-712. Current chemtherapy and infectius disease. Prceedings f the 11th Internatinal Cngress f Chemtherapy and the 19th Interscience Cnference n Antimicrbial Agents and Chemtherapy. American Sciety fr J. CLIN. MICROBIOL. Micrbilgy, Washingtn, D.C. 24. Zervs, M. J., C. A. Kauffman, P. M. Therasse, A. G. Bergman, T. S. Mikeseli, and D. R. Schaberg. 1987. Nscmial infectin by gentamicin-resistant Streptcccus faecalis. Ann. Intern. Med. 106:687-691. 25. Zimmerman, R. A., R. C. Mellering, Jr., and A. N. Weinberg. 1971. Mechanism f resistance t antibitic synergism in enterccci. J. Bacteril. 105:873-879.