Antimicrobial activity of ceftriaxone compared with cefotaxime in the presence of serum albumin

Department of Laboratory Medicine and Division of Infectious Diseases, Henderson General Hospital, Hamilton, Ontario Correspondence and reprints: Dr SK Nath, Laboratory Medicine, Henderson General Hospital, Hamilton, Ontario L8V 1C3. Telephone (905) 527-4322 ext 2051, Fax (905) 575-2581 Received for publication January 31, 1994. Accepted June 14, 1994 SK NATH, GA FOS TER, LA MAN DELL, C ROT STEIN. An ti mi cro bial ac tiv ity of cef tri ax one com pared with ce fo taxime in the pres ence of se rum al bu min. Can J In fect Dis 1995;6(1):2127. The ef fect of se rum al bu min on the an ti mi cro bial ac tiv ity of cef tri ax one, ce fo taxime, and a 1:1 ra tio of ce fo taxime and its de sace tyl me tabo lite against nonpseu do mo nal Gramnegative ba cilli was de ter mined. An ti mi cro bial ac tiv ity of drugs was evalu ated by meas ur ing mini mum in hibi tory (MIC) and bac te ri cidal (MBC) con cen tra tions in broth with and with out hu man se rum al bu min. The analy sis of loga rith mi cally trans formed mean MICs and MBCs showed that there was a highly sig nifi cant in ter ac tion be tween drug and se rum al bu min (P<0.0001). The in hibi tory and bac te ri cidal ac tivi ties were great est for ce fo taxime fol lowed by ce fo taxime/de sace tylce fo taxime and cef tri ax one (P<0.01). Timekill ki net ics dem on strated that cef tri ax one was less bac te ri cidal than ce fo taxime in broth with al bu min. On the ba sis of these re sults it was con cluded that the in vi tro an ti mi cro bial ac tiv ity of cef tri ax one com pared with that of ce fo taxime was sig nifi cantly di min ished in the pres ence of se rum al bu min.

T HIRD-GENERATION CEPHA LO SPORINS ARE OF TEN USED IN the ther apy of se ri ous no so comial Gram-negative bac te rial in fec tions.Cef tri ax one and ce fo taxime are simi lar in an ti bacte rial spec trum and re sis tance to beta-lactamases, but dif fer mark edly in se rum pro tein bind ing and elimi na tion half-life (cef tri ax one: 95% pro tein bound, elimi na tion half-life of 8.8 h ver sus ce fo taxime: 35%, 1.2 h) (1,2).It has been sug gested that se rum pro tein bind ing may di min ish thera peu tic ef fi cacy by af fect ing an ti bac te rial ac tiv ity and drug dis tri bu tion (3).Thus, high se rum pro tein bind ing of cef tri ax one may in fact limit its in vi tro ac tiv ity (4) and im pair its thera peu tic ef fi cacy com pared with that of ce fo taxime (5).
In the pres ent in ves ti ga tion, us ing sta tis ti cal meth ods we evalu ated dif fer ences in the in hibi tory and bac te ri cidal ac tivities of cef tri ax one, ce fo taxime, and 1:1 com bi na tion of ce fotaxime and its me tabo lite de sace tyl ce fo taxime against 121 clini cal iso lates of nonpseu do mo nal Gram-negative ba cilli in the pres ence of hu man se rum al bu min.Us ing a time-kill kinetic method we also as sessed the di min ished kill ing of organ isms by cef tri ax one com pared with that of ce fo taxime in broth with al bu min.

MA TE RI ALS AND METH ODS
One hun dred and twenty-one bac tere mic iso lates of nonpseu do mo nal Gram-negative ba cilli (77 Escheri chia coli, 18 Kleb siella pneu mo niae, four Kleb siella oxy toca, five Proteus mirabilis, four Citro bac ter fre un dii, nine En tero bac ter cloa cae and four En tero bac ter ae ro genes) were used.The iso lates were from a stock cul ture col lec tion, and an ti mi crobial sen si tivi ties were de ter mined us ing the Mi croScan system (Tra ve nol Labo ra to ries, New Jer sey) with break point mini mum in hibi tory con cen tra tion (MICs) in the sus cep ti ble cate gory (less than 8 µg/mL) for both cef tri ax one and ce fotaxime.Six other nonpseu do mo nal Gram-negative bac te rial strains (two E coli, two E cloa cae and two K pneu mo niae) isolated from blood and spu tum of pa tients with pneu mo nia in an in ten sive care unit were em ployed in the study of kill ing ki netics in broth only and in broth con tain ing 48% and 95% equivalent lev els of hu man se rum al bu min.
Cation-adjusted Mueller-Hinton Broth (CAMHB) (Difco Can ada [Lot No G3DXRS]) and nu tri ent agar (Bec ton Dick inson BBL Can ada [Lot No E4DYYC]) were em ployed.For each me dium, a sin gle batch was used for all in vi tro test ing.
The an ti bi otic stan dard pow ders em ployed in the study and their sup pli ers were: cef tri ax one (Hoffmann-La Roche Ltd [Lot No 12592]); ce fo taxime (Hoechst-Roussel Inc [Lot No 112009]); and de sace tyl ce fo taxime (Hoechst-Roussel Inc [Lot No 8A0203B]).
MICs for the or gan isms were de ter mined by the broth micro di lu tion method (6).An ti bi otic stock so lu tions were prepared ac cord ing to the manu fac tur er's in struc tions.Cef tri ax one, ce fo taxime and the 1:1 com bi na tion of ce fotaxime with its de sace tyl me tabo lite were tested in broth with and with out se rum al bu min us ing se rial dou bling di lu tions of each an ti bi otic (0.015 to 8 mg/L).Broth with se rum al bu min con tained 19 g/L (for time-kill ki net ics only) and 38 g/L al bumin, equiva lent to 48% and 95% (weight/vol ume) al bu min in nor mal hu man se rum.The se rum al bu min was ob tained from blood col lected from healthy vol un teer do nors by the Blood Serv ices of the Ca na dian Red Cross So ci ety (manu fac tured and dis trib uted as 25% so lu tion, USP 100 mL by Miles Canada Inc).A sin gle batch of se rum al bu min was used for all in vi tro test ing.All strains grew in the se rum al bu min and were in cluded in this study.The drug con cen tra tion that fi nally inhib ited growth with out any amount of visi ble growth at higher con cen tra tions was re corded as the MIC af ter in cu ba tion of the pan els for 18 h at 35°C.Qual ity con trol tests were con ducted by in clud ing E coli ATCC 25922 in each batch of tests.MICs for cef tri ax one and ce fo taxime were al ways found to be within the ac cept able ranges.
For mini mum bac te ri cidal con cen tra tion (MBC) de ter mi nations, du pli cate 0.01 mL sam ples were re moved from each well with out growth and dis pensed by spread ing across the sur face of nu tri ent agar plates.Re sults were re corded af ter in cu ba tion of plates for 18 h at 35°C.The ini tial in ocula were de ter mined and ranged from 3 to 4x10 5 colony-forming units (cfu)/mL.The MBC was de fined as the low est con cen tra tion pro duc ing at least 99.9% kill ing of the origi nal in ocu lum, us ing the ap pro pri ate re jec tion val ues de scribed by Pear son et al (7).
Bac te ri cidal ac tiv ity of cef tri ax one and ce fo taxime against six nonpseu do mo nal Gram-negative bac te ria was also measured us ing the time-kill ki netic method (8).The an ti bi otic concen tra tions that were used rep re sented the MBC for each strain tested.In ocu lated broths con tain ing ap proxi mately 3.5x10 5 cfu/mL and antibi otic were in cu bated, and at 0, 1, 2, 4, 6, 18 and 24 h in ter vals, a 0.01 mL ali quot was re moved from each of the sam ple tubes and se ri ally di luted (1:10).A 0.1 mL of di lu ent was plated on nu tri ent agar.Vi able counts were de ter mined af ter in cu ba tion of the plates for 18 h at 35°C.The bac te rial sur vival meas ured as log 10 cfu was plotted against time (hours).A time-kill curve was es tab lished for each strain.The nega tive ef fect of se rum al bu min on bac te rial kill ing was as sessed by a 2log 10 or greater in crease in cfu (100-fold de crease in kill ing) at 6 h in stead of 24 h for the conven ience of plot ting and ex trapo la tion.
For data analy ses, MIC and MBC re sults less than 0.015 and greater than 8 mg/L were con verted to 0.007 and 16 among the means of the natu ral loga rith mic trans formed meas ure ments (ie, geo met ric means).Two-way analy sis of vari ance (ANOVA) was used to de ter mine whether there was an in ter ac tion be tween an ti bi otic and se rum al bu min con centra tion.One-way ANOVA was used to de ter mine whether there were dif fer ences among an ti bi ot ics at each of the se rum al bumin con cen tra tions.When dif fer ences oc curred, the Scheffé post-hoc test in di cated which mean(s) dif fered sig nifi cantly from the oth ers.

RE SULTS
The sus cep ti bil ity data of the bac te rial strains in broth and broth with se rum al bu min are shown in Ta ble 1.A marked effect of se rum al bu min on the MIC 90 and MBC 90 was noted.The cef tri ax one MIC 90 and MBC 90 in creased eight fold in the presence of se rum al bu min.The MIC 90 of ce fo taxime and ce fotaxime/de sace tyl ce fo taxime were vir tu ally un af fected by the ad di tion of se rum al bu min.There were mi nor dif fer ences in the MBC 90 of ce fo taxime and ce fo taxime/de sace tyl ce fo taxime in broth com pared with broth with se rum al bu min.

Fig ure 1) Fre quency dis tri bu tion of mini mum in hibi tory con cen tra tions (MICs) and mini mum bac te ri cidal con cen tra tions (MBCs) (mg/L). All or ganisms were tested against cef tri ax one (MIC, MBC; a, b), ce fo taxime (MIC, MBC; c, d), and ce fo taxime/de sace tyl ce fo taxime (MIC, MBC; e, f). Solid bars: in vi tro test ing in broth with out se rum al bu min; hatched bars: in vi tro test ing in broth with 95% equiva lent se rum al bu min
The fre quency (number) dis tri bu tion for the or gan isms inhib ited (MIC) or killed (MBC) at the an ti bi otic con cen tra tions are shown in Fig ure 1.For ce fo taxime and ce fo taxime/de sace tylce fo taxime, the ma jor ity of the MIC re sults were clus tered at the three drug con cen tra tions (Fig ure 1c,e).For cef tri ax one MICs (Fig ure 1a), the iso lates were dis trib uted over two clusters (0.03 to 0.06 mg/L in ab sence of al bu min and 0.25 to 0.5 mg/L in pres ence of al bu min).Simi lar dif fer ences are observed for the MBC re sults of ce fo taxime and ce fo taxime/desace tyl ce fo taxime ver sus cef tri ax one (Fig ure 1d,f ver sus 1b).Since the dis tri bu tions of MIC and MBC raw data were highly skewed, sta tis ti cal analy ses were per formed on the geo met ric means.Two-way ANOVA of MICs and MBCs showed that there was a highly sig nifi cant in ter ac tion be tween drug and se rum al bu min (P<0.0001).The ex tent of this ef fect was evalu ated by one-way ANOVA of MICs tested with the Scheffé post-hoc com pari son.There were sig nifi cant dif fer ences in geo met ric mean MICs and MBCs among the three drugs in broth only (P=0.05) and broth with 95% equiva lent se rum al bu min (P=0.0001)(Ta ble 2).The one-way ANOVA of MICs with the Scheffé post-hoc com pari son de fini tively showed that ce fotaxime and ce fo taxime/ de saceytl ce fo taxime were sig nificantly more ac tive than cef tri ax one when tested with al bu min (P<0.01).Ce fo taxime was found to be sig nifi cantly more inhibi tory than a 1:1 com bi na tion of ce fo taxime and its de sacetyl me tabo lite in broth with out and with al bu min (P<0.01).Even when se rum al bu min was added, ce fo taxime had the most po tent bac te ri cidal ac tiv ity (P<0.01).
The ef fects of se rum al bu min on the bac te ri cidal ac tiv ity of cef tri ax one and ce fo taxime were also as sessed by the timekill ki netic analy sis.Any re duc tion in bac te ri cidal ac tiv ity was as sessed by de ter mi na tion of the sur vival rate (log 10 cfu/mL ver sus time [hours]) of bac te ria when ex posed to a drug in broth con tain ing 48% and 95% equiva lent con cen tra tions of se rum al bu min.Six strains with ele vated MBCs of cef tri ax one were tested: two E coli strains (Fig ure 2a,b) with cef tri ax one MBCs of 0.03 and 0.125 mg/L that in creased to 0.5 and 2.0 mg/L, re spec tively, by ad di tion of se rum al bu min; two E cloacae strains (Fig ure 2c,d) with cef tri ax one MBCs of 0.5 and 0.25 mg/L that rose to 2.0 mg/L in broth with se rum al bu min; and two K pneu mo niae strains (Fig ure 2e,f) with cef tri ax one MBCs of 0.06 and 0.125 mg/L that in creased to 0.5 and 4.0 mg/L in se rum al bu min.Only one K pneu mo niae strain tested (Fig ure 2f) had an ele va tion in ce fo taxime MBC (0.03 to 1.0 mg/L).Serum al bu min at 48% or 95% equiva lent se rum lev els did not overtly af fect the bac te ri cidal ac tiv ity of ce fo taxime based on 2log 10 or greater in crease in sur vival (cfu/mL) of bac te ria at 6 h or 24 h.In con trast, a marked nega tive ef fect on the bac te ricidal ac tiv ity of cef tri ax one was dem on strated for all six strains with the 95% equiva lent se rum al bu min con cen tra tion and for four strains at the 48% equiva lent con cen tra tion.There was no dif fer ence in the ex tent of kill ing at 6 h or 24 h, when the bac te ri cidal ac tiv ity of the an ti bi ot ics was com pared in the pres ence of 48% or 95% equiva lent se rum al bu min.

DIS CUS SION
The re la tion ships be tween se rum pro tein bind ing and the dis tri bu tion of beta-lactam an ti bi ot ics has been ex ten sively stud ied and re viewed over the past two dec ades (9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19).The dif fer ences in se rum pro tein bind ing be tween cef tri ax one and ce fo taxime have been es tab lished (1,2).Meth od ol ogy has been de vel oped to com pare the phar ma coki net ics of an ti bi otics that dif fer in pro tein bind ing (20).Pro tein bind ing may be of im por tance in an ti bac te rial che mo ther apy be cause the bound an ti bi otic is es sen tially un avail able for an ti bac te rial ac tiv ity (14,21).This ef fect should be re flected by an in crease in MIC and MBC un der the in flu ence of se rum pro tein.The pres ent study was not in tended to quan tify the to tal and free con centra tions of cef tri ax one and ce fo taxime in se rum (1,2,19), but rather to evalu ate the change in in vi tro in hibi tory and, more im por tant, bac te ri cidal ac tiv ity of cef tri ax one, ce fo taxime and cefto taxime/de sace tyl ce fo taxime (in 1:1 com bi na tion) against 121 nonpseu do mo nal Gram-negative bac te ria in the presence of hu man se rum pro tein.Since cef tri ax one is pre dominantly bound to al bu min (22), the pres ent in ves ti ga tion em ployed 48% and 95% equiva lent lev els (19 and 38 gm/L) of hu man se rum al bu min, in stead of di lut ing with 25% or 50% (vol ume/vol ume) of hu man se rum as de scribed by Jones and Barry (4).
We found that MIC 90 (1 mg/L) and MBC 90 (2 mg/L) for ceftri ax one in the pres ence of 95% equiva lent se rum al bu min in creased eight fold and were at least a two fold di lu tion lower than those of ce fo taxime (0.25 and 1 mg/L) and ce fotaxime/de sace tyl ce fo taxime (0.5 and 1 mg/L), re spec tively.Be cause the MIC 90 , a fre quently used per cen tile to evalu ate in vi tro ac tiv ity, is prone to er ror, even in sym met ri cal data dis tribu tions (23,24), and with non sym met ri cal or skewed dis tri butions one should be cau tious of us ing any per cen tile as an es ti ma tor of in vi tro ac tiv ity.We em ployed the natu ral loga rithmic trans formed meas ure ments (ie, geo met ric means) of MICs and MBCs to achieve a more ac cu rate meas ure of cen tral ten dency (23,24).The geo met ric mean MIC and MBC were the low est for cef tri ax one in di cat ing it had the best an ti mi cro bial ac tiv ity in broth only.How ever, based on the geo met ric means, ac tiv ity of ce fo taxime was su pe rior in broth with serum al bu min.
Highly sig nifi cant ele va tions in the geo met ric mean MICs and MBCs of cef tri ax one were ob served by two-way ANOVA when in vi tro test ing was switched from 0% (broth only) to 95% equiva lent se rum al bu min.There were also marked differ ences be tween ce fo taxime and ce fo taxime/de sace tyl ce fotaxime.Ce fo taxime proved more ac tive than the 1:1 com bi na tion of ce fo taxime and its me tabo lite, al though an ear lier re port docu mented the ad van tages of an ti mi cro bial agents that have ac tive me tabo lites (25).Man dell and Af nan (26) also re ported syn er gis tic ac tiv ity of de sace tyl ce fo taxime and its parent com pound in the pres ence of neu tro phils.The one-way ANOVA of geo met ric mean MICs and MBCs tested by Scheffé post-hoc showed that the in hibi tory and bac te r i cidal ac tivi ties were great est for ce fo taxime fol lowed by ce fotaxime/de sace tyl ce fo taxime and cef tri ax one.Be cause the time-kill ki netic method is con sid ered to be more sen si tive than sim ple MIC-MBC meth od ol ogy for evalu ating an ti mi cro bial ac tiv ity, it is pos si ble to ob serve wide dif ferences in the kill ing ki net ics among strains with simi lar MICs (8).
CAN J INFECT DIS VOL 6 NO 1 JANU ARY/FEB RU ARY 1995 Since there was no dif fer ence in the ex tent of kill ing at 6 or 24 h, the ex trapo lated re sults from the con ven ience plot of 6 h time-kill tested at MBC lev els of drugs were con sid ered to be pre dic tors of the bac te ri cidal ac tiv ity.All six strains tested resisted kill ing by cef tri ax one in broth with 95% se rum al bu min.A marked nega tive ef fect of se rum al bu min on the kill ing by cef tri ax one was also dem on strated for four of six strains at a re duced, 48% equiva lent se rum al bu min level in the test ing broth.The kill ing rate was also slowed for two of six strains exposed to ce fo taxime in broth with se rum al bu min.
The pres ent study dem on strated a sig nifi cant re duc tion in an ti mi cro bial ac tiv ity of cef tri ax one in the pres ence of hu man se rum al bu min (P=0.0001).In vivo ani mal model stud ies such as that of Ta wara et al (27) have shown that the thera peu tic ef fi cacy of cephem an ti bi ot ics in clud ing cef tri ax one in K pneu mo niae-in duced ex peri men tal pneu mo nia cor re lated with to tal drug level in se rum.How ever, in a rat model of K pneu mo niae pneu mo nia de scribed by Bakker-Woudenberg et al (28) pro tein bind ing of these an ti mi cro bial agents in se rum ad versely af fected both drug pene tra tion into lungs and therapeu tic ef fi cacy.
The clini cal sig nifi cance of the nega tive ef fect of cef tri axone bind ing to se rum al bu min re mains con tro ver sial.While the phar ma coki netic and thera peu tic ef fi cacy data from the four clini cal stud ies of cef tri ax one and ce fo taxime (5,(29)(30)(31) are equivo cal in show ing a dif fer ence be tween the two agents, Man dell et al (31) and Smith et al (32) re ported that the thera peu tic ef fi cacy of cef tri ax one (2 g every 24 h) was com pa ra ble with that of ce fo taxime (2 g every 6 h) in treat ing se ri ous no so comial bac te rial in fec tions.
The re sults of our study sug gest that there was a nega tive in flu ence of se rum al bu min on the in vi tro an ti mi cro bial ac tiv ity of cef tri ax one.The clini cal im por tance of this has not been deter mined.

TA BLE 2 Sta tis ti cal analy sis of raw data of mini mum in hibi tory con cen tra tions (MICs) and mini mum bac te ri cidal con - cen tra tions (MBCs) of cef tri ax one (CAX), ce fo taxime (CFT) and ce fo taxime/de sace tyl ce fo taxime (1:1; CFT/ DACT) in broth only (0%) and broth with 95% equiva lent hu man se rum al bu min
Based on the one-way analy sis of vari ance CAN J INFECT DIS VOL 6 NO 1 JANU ARY/FEB RU ARY 1995 *