- Email: [email protected]
Use of aminoglycosides in immunocompromised patients
Use of Aminoglycosides in lmmunocompromised Patients
LOWELL S. YOUNG, M.D. Los Angeles,
California
From the Department of infectious Diseases, Univereity of California School of Medicine, Center for Health Sciences, Los Angeles, California. Requests for reprints should be addressed to Dr. Lowell S. Young, Department of Medicine, UCLA Medical School, Center for Health Sciences, 10883 La Conte, Los Angeles, California 90024.
For much of the last decade, combination therapy with aminoglycosides has been accepted as the therapeutic approach of choice in immunocompromised hosts. improved clinical results have also correlated with the presence of synergistic interactions between the aminoglycoside and beta-lactam components of a regimen. Differences between the aminoglycosides and beta-la&am agents remain a subject of controversy. Studies at the University of California, Los Angeles, Medical Center suggest that amikacin interacts more frequently in a synergistic manner with beta-lactams than do alternative aminoglycosides. Amikacin has been used experimentally and (following llcensure) without reservation at the University of Caiifornia, Los Angeles, Medical Center since 1973. Almost 199 blood isolates of both Pseudomonas aeruginosa and Klebsiella pneumoniae collected during the last 12 years have been retested, and no evidence of increased aminoglycoside resistance was found. A relatively new development is Interest in empiric therapeutic regimens that employ two beta-lactam agents. In a large, recently completed study, less satisfactory results were observed in P. aeruginosa infections treated wlth the “double beta-lactam” than in those treated with the regimen containing amikacin; furthermore, nephrotoxicity and eighth nerve damage occurred no more commonly in the group receiving amikacin than in recipients of the double beta-lactam regimen. Aminoglycosides have been used to treat infections in immunocompromised patients for many years. In the 1950% prior to the availability of gentamicin, agents such as streptomycin and kanamycin were used frequently as initial therapy in these patients. In immunocompromised patients, the most important organism in terms of mortality, although not necessarily the most prevalent, has always been Pseudomonas aeruginosa. In fact, among university, county, cancer, and even community hospitals, P. aeruginosa is associated with the highestcase-fatality ratio of all causes of gram-negative rod bacteremia [l]. Several years ago, Bodey and colleagues [2] observed that survival from the onset of Pseudomonas sepsis in cancer patients was very poor. Often a significant number of these patients died within one or two days after the first positive blood culture result. For these reasons, therapy in this high-risk group of patients must be empiric: treatment must be initiated before the infecting organisms have been identified and their antimicrobial susceptibilities determined. In order to deal with gram-negative pathogens that might be present, many physicians employ aminoglycosides in the initial therapeutic regimen. However, there are many contro-
July
15, 1995
The
American
Journal
oi Medicine
Volume
79 (suppl
1A)
21
AMINOGLYCOSIDE
TABLE
I
SYMPOSIUM-YOUNG
Controversies
Aminoglycosldes
about
Aminogiycosides
811) ineffective In neutropenic
patients.
ISSUe
Adequacy of dose and achievement of therapeutic dose levels in vivo Better results are achieved by continuous infusion. ISSUeS
Supporting evidence scanty More toxicity in some studies Larger doses actually given Will comblnatlon use llmlt emergence of resistance to betaIactamsi Evidence Shown in some studies and may be related to duration of therapy Amlnoglycosldes will be replaoed by single agents, such as new cephalosporlns. Issue Suggested by some studies but issue unresolved in neutropenic and other high-risk patients
versies concerning the use of aminogiycosides, not only in immunocompromised patients but in other patients as well. Some of these controversies are highlighted in Table I. MAJOR
CONCERNS
ABOUT
AMINOGLYCOSIDES
Many have claimed that aminoglycosides are ineffective in neutropenic patients. However, it should be remembered that the ineffectiveness of a drug or class of drugs in vivo can be substantiated only by compelling and convincing evidence that adequate doses of the drug have been given and that therapeutic levels have actually been achieved. Many studies that claim aminogiycosides are ineffective in neutropenic patients have used subtherapeutic doses and have failed to demonstrate adequate blood levels [3]. Another controversy relating to aminogiycoside therapy involves the concept that better therapeutic results can be achieved by continuous infusion. Although this may seem plausible, a review of the literature suggests that, in fact, the beta-lactam antibiotics rather than aminoglycosides may give better therapeutic results when given in continuous infusion [2,3]. Craig and his colleagues [4,5] have studied the post-antibiotic effects of a variety of drugs. They found that lethal antibiotic activity of ail aminogiycosides is so rapid that, in many cases, there is a post-antibiotic effect of one to four hours against clinically important gram-negative organisms. Craig’s work, among others [6], suggests that when therapeutic levels of aminogiycosides fall below the minimal inhibitory concentration, organisms such as P. aeruginosa will start to grow again unless there is a significant post-antibiotic effect. This regrowth phenomenon, however, is much more striking with the beta-lactam antibiotics when studied in vitro. Therefore, it would seem that the application of the continuous infusion technique has been improperly directed and 22
July 15,1985
The American Journal of Medlclne
should, in fact, be applied more to the beta-iactam agents than to the aminoglycosides. In several studies, researchers have attempted to compare the efficacies of continuous and intermittent infusion of aminogiycosides. in one study by Bodey and associates [A on the continuous infusion of amikacin in neutropenic patients, results showed that more than half of the patients in whom amikacin levels were maintained at greater than 20 pg/mi subsequently showed some form of ototoxicity or nephrotoxicity. These results suggest that continuous exposure of vital eighth nerve and kidney structures to high levels of aminoglycosides increases risk of toxicity. A second issue of the continuous infusion controversy involves the amount of drug actually delivered. in an attempt to achieve sustained high levels of aminoglycosides [3], the amount of drug actually delivered is considerably greater than the amount the patient would have received by a conventional intermittent-dosing regimen. Thus, in those situations in which better clinical results seem to be obtained with continuous infusion, the improvement is likely due to the greater total daily dosage of drug delivered. IS EMERGENCE OF RESISTANCE COMBINATION THERAPY?
LESSENED
BY
Much attention has been paid to the growing concern that prolonged therapy with beta-iactam agents may cause the emergence of antimicrobial resistance. The concept has been advanced that combining a beta-lactam plus an aminogiycoside may limit the development of resistance to the beta-lactam. One of the major questions in current studies concerning any limitation of beta-iactam resistance by aminogiycosides is the duration of therapy. In many of these studies, the combination therapy is administered for only 10 to 14 days, so that significant differences in the rate of emergence of resistance would be very difficult to appreciate. The issue of emergence of resistance is certainly difficult to evaluate in most of the studies comparing immunocompromised and non-immunocompromised patients. One of the best-designed multicenter studies that addresses this issue was conducted in Canada, by Gribble and colleagues [8]. in this study, piperaciiiin was used alone and in combination with an aminogiycoside, usually gentamicin or tobramycin. When piperaciiiin was used as a single agent, it was associated with a significantly higher incidence of resistance (Table ii).
We have assessed therapeutic results in a retrospective study of P. aeruginosa bacteremia [9]. Combinations of the antimicrobial agents that exert synergistic antimicrobial activity against the actual infecting strains were associated with a significantly better therapeutic response. For example, 10 of 12 patients in whom synergistic interaction between gentamicin and carbeniciliin could be demonstrated showed clinical response. Cf the six patients whose isolates were not synergistically affected, none
Volume 79 (suppl 1A)
AMINOGLYCOSIDE
showed resoonse. Similar findinas have been reported bv Klastersky and Zinner [lo]. An especially interesting study was reported by Lerner and Reyes [Ill, in which they evaluated non-immunocompromised hosts with P. aeruginosa endocarditis. In their experience, the only patients with medical cures were those in whom the infecting Pseudomonas organism was synergistically inhibited and killed by the combination of an aminoglycoside and a -_ broad-spectrum penicillin. In a prospective study of empiric therapy of neutropenic patients conducted at the University of California, Los Angeles, Medical Center [12], the protocol called for either amikacin plus carbenicillin or gentamicin plus carbenicillin. Almost concurrently, Love et al [13] at the University of Maryland employed a similar protocol using ticarcillin as their antipseudomonal penicillin. If the results of these two studies are pooled, it becomes evident that the clinical response rate is over 80 percent when the pathogen isolated from blood is susceptible to both antibiotics used in gram-negative bacteremia in neutropenic patients. Although the issue of synergism remains unsettled, additive effects do exist and are likely to be beneficial. Experience at the University of California, Los Angeles, Medical Center [14] in analyzing individual isolate response versus clinical therapeutic response has shown that there is a statistically significant difference between inhibition by only one drug and inhibition by two drugs that exhibit a synergistic effect. Our results indicate that both in vitro synergism and additive effects are associated with improved therapeutic effects. NEW AGENTS: SINGLE AND COMBINED The majority of newly introduced antibiotics for systemic use in the United States are cephalosporins and related agents. Because of claims that these newer agents are more potent and have a broader spectrum than previously available antibiotics, many studies of single-agent therapy with these cephalosporins have been undertaken. Nevertheless, monotherapy with these newer agents in the treatment of some gram-negative bacillary infections caused by such organisms as Pseudomonas, Enterobacter, Serratia, and even Citrobacter has, in the experience of many investigators [15], been associated with a rapid emergence of resistance. In our study [16], the emergence of resistance in P. aeruginosa to single-agent therapy with moxalactam reached 35 percent; a somewhat lower rate of emergence of resistance was found in another study with cefoperazone (unpublished data). Some of our patients did show clinical improvement with moxalactam therapy despite emergence of resistance. However, improvement may have been due to a variety of factors unrelated to antimicrobial treatment per se. For example, a patient with a leg ulcer may have been treated with both antibiotics and surgical debridement, so that clinical response could not be attributed solely to antibiotic monotherapy. July
15, 1996
TABLE II
Emergence ol Resistance* Using Plperkillin Alone and in Combination with Aminoglycosldes Rsglmen
Piperacillin Piperacillin
SYMPOSIUM-YOUNG
Penent
alone + aminoglycoside
Number
42 17
26 24
“Emergence of resistance accounted for five of nine episodes ment failure. Reproduced with permission from [a].
of treat-
To support claims that single-agent therapy with newer cephalosporins-because of their greater potency and broad spectrum-will supplant the aminoglycosides, there are some seemingly well-executed clinical studies [17,18] in which treatment with a third-generation cephalosporin has been compared with combination therapy. In a recent study by Smith et al [la], which was not a strict comparison of monotherapy versus the same agent plus an aminoglycoside, the effects of monotherapy with cefotaxime were compared with those of combination therapy that included an antistaphylococcal agent plus an aminoglycoside, i.e., cefotaxime versus nafcillin plus tobramytin. The overall response rates were 90 percent and 60 percent, respectively. However, upon careful review of these data and the methodology employed in collecting them, some disproportionalities become evident. In this double-blind, randomized, prospective study, there were relatively few gram-negative rod infections, particularly of the bloodstream [la]. The entire study included a total of nine Pseudomonas infections, seven Enterobacter infections, and 10 Klebsiella infections. Further analysis of these 26 infections shows that 20 were treated with nafcillin plus tobramycin, whereas only six received cefotaxime alone. This difference is statistically significant. Additionally, this study enrolled no neutropenic patients. Studies such as this one, however excellent the clinical design, have not convincingly established that third-generation cephalosporins like cefotaxime can replace combination therapy with aminoglycosides for serious gram-negative infections in neutropenic patients. The excellent results of recent studies of cefoperazone in gram-negative bacteremia would, by any measure, be considered impressive. Response rates for bacteremia range from 84 to 100 percent in summaries of case reports collected during the course of some clinical trials [19] conducted before licensure in the United States. It should be borne in mind, however, that evaluability criteria can affect the interpretation of summary data. For example, in a multicenter study conducted by Cohen et al [20], a case was not considered evaluable if a patient received cefoperazone therapy for fewer than five days. Thus, by eliminating those patients who received short-duration therapy, they achieved a response rate of 100 percent in the treatment of gram-negative bacteremia. In contrast, The American
Journal
of Medlclne
Volume
79 (suppl
1A)
23
AMINOGLYCOSIDE
TABLE
III
SYMPOSIUM-YOUNG
Significant interactions between Lactam Antibiotics Demonstrated and in Animal Models
Betain Vitro
DOUBLE
BETA-LACTAM
THERAPY
Bolivar and associates [21] studied cefoperazone in the treatment of bacteremia occurring in neutropenic patients and attained a clinical response rate of 50 percent for Enterobacter infections, 33 percent for Klebsiella, and 50 percent for PSeUdDmOnaSinfections. The major apparent difference between the results of the multicenter cefoperazone study summarized by Cohen et al [20] and the report of Bolivar and colleagues [21] is that the patients in the latter study were neutropenic.
The most recent innovation in empiric antimicrobial therapy in febrile neutropenic patients is “double beta-lactam” treatment. This therapy entails the use of a broad-spectrum penicillin plus a cephalosporin, rather than a betalactam agent with an aminoglycoside. In a study conducted at the University of California, Los Angeles, Medical Center [22], we compared moxalactam plus piperacillin with moxalactam plus amikacin, the latter being considered “conventional therapy.” In patients with Pseudomonas infections, only one of five receiving the double beta-lactam regimen showed a favorable response, as opposed to seven of nine treated with moxalactam plus amikacin. This difference misses attaining statistical significance by only one one-hundredth of a point (p = 0.06). Also in this study, two patients with Pseudomonas bacteremia who had initially had susceptible isolates of P. aeruginosa experienced a relapse of bacteremia associated with multiply beta-lactam-resistant isolates. This is highly suggestive of the phenomenon of inducible beta-lactamase resistance that has been well studied by Sanders and Sanders [15]. Mechanisms for this phenomenon are summarized in Table III. An important point regarding the treatment of immunocompromised patients with double beta-lactam therapy concerns the incidence and severity of neutropenia in these patients. In our study, there was significantly more severe neutropenia among patients receiving the double beta-lactam regimen. It is well established that beta-lactam agents cause neutropenia, which is almost always reversible. Therefore, we must be concerned that use of two beta-lactam agents may actually prolong neutropenia and augment risk of superinfection.
EORTC DATA
NEPHROTOXICITY
Data from the European Organization for the Research and Treatment of Cancer (EORTC) have been discussed at length elsewhere in these symposium proceedings. However, the relevance of their results to this discussion pertains only to one infection-P. aeruginosa occurring in Trial I and Trial III. The last study showed clearly that combination therapy using amikacin plus a broad-spectrum antipseudomonal penicillin, such as azlocillin or ticarcillin, is statistically significantly better than combination therapy with a cephalosporin-even one of the most popular of the new agents, cefotaxime.
In another study at the University of California, Los Angeles, Medical Center, we compared the combination of amikacin plus carbenicillin with the combination of netilmitin plus carbenicillin in more than 230 patients in order to examine the incidence and nature of nephrotoxicity (L.S. Young and W.L. Hewitt, unpublished data). We chose netilmicin because it had been hoped, on the basis of animal studies, that it would be a less nephrotoxic agent than the other aminoglycosides. Our results did not support this contention, however. Our definition of nephrotoxicity required a rise in serum creatinine level of at least 60 per-
Synergism
Antagonism
Mechanism Inactivation of beta-lactamase by inhibitor; enzyme-labile beta-lactam reaches lethal target in cell
Induction of beta-lactam hydrolyzing enzymes
Attack on different lethal targets (penicillin-binding proteins) in cell
Induction of beta-lactamasemediated barrier that blocks access to lethal targets in cell
Examples of Organisms Involved Staphylococcus aureus, gram-negative bacteria producing types II, Ill, IV, or V beta-lactamases
Enterobacteriaceae possessing inducible type I beta-lactamases (e.g., Enterobacter, Serratia, Ciirobacter)
Enterobacteriaceae
Enterobacteriaceae Pseudomonas inducible type lactamases
Adapted
with permission
TABLE
IV
from
(151.
Synergistic Inhibition and/or Killing of 30 Gentamlcin-Reslstant Pseudomonas Amikacin or Tobramycln Plus Any of Six Beta-Lactam Antlmicroblals Ticanillin
Amikacin Tobramycin Total synergism ‘30 strains
24
and possessing I beta-
Thienamycin
Cafaulodin
Moxalacfam
15
0
7 22
2 2
6 5 11
16 13 31
per group.
July 15, 1995
The American Joumel of Medicine
Volume 79 (suppl 1A)
aeruginosa’ Csfoparaxana 22 20 42
by Cefofaxlma 25 20 45
AMINOGLYCOSIDE
TABLE V
In Vitro Synergism against AminoglycosidcSusceptible
Number of strains tested Amikacin+ Ticarcillin Moxalactam Cefoperazone Cefotaxime Tobramycin+ Ticarcillin Moxalactam Cefoperazone Cefotaxime ‘All blood isolates were chosen because ‘K-E-S group: Klebsiella-Enterobacter-Serraia
SYMPOSIUM-YOUNG
Gram-Negative Rods*
pseudomonas aerusinosa
Eschsrichia coli
K-E-S’ group
Acinetobacter
Providencia
13
9
15
3
9
12 13 12 13 8 13 11 13
5 8 7 of piperacillin group
resistance.
cent to a concentration of 1.3 mg/dl or more. Of the 21 patients who met this criterion of nephrotoxicity, we found only two in whom the rise in serum creatinine was attributable only to the aminoglycoside with which they had been treated. In the great majority of cases, nephrotoxicity occurred in patients who were hypotensive, receiving amphotericin B, cisplatin, diuretics, radiologic contrast media, or other potentially responsible agents. Therefore, much of what has been attributed to aminoglycoside toxicity could be due to important cofactors. RELEVANCE OF SYNERGISTIC ACTIONS OF ANTIMICROBIAL COMBINATIONS Numerous synergism studies have been conducted both in vitro and in vivo. Recent interest has been focused on newer agents, such as imipenem and cefsulodin. It is interesting to note that despite the potency of imipenem, for example, it is quite difficult to demonstrate synergistic interaction with aminoglycosides. One explanation for this may be that the killing of P. aeruginosa by imipenem is so complete that there are no surviving subpopulations for the concomitant aminoglycoside to kill. In the study represented in Table IV, of 30 gentamicin-resistant strains of P. aeruginosa, more synergistic effects were observed with the less potent antipseudomonal compounds than with imipenem. Although there seems to be little synergism with imipenem, there is far greater synergism evidenced with cefoperazone or cefotaxime. This phenomenon may, in fact, be related to the incomplete killing with the less potent agents (cefoperazone, cefotaxime). It seems clear that, with these selected strains of P. aeruginosa (selected for gentamicin resistance), there is somewhat more synergism with amikacin plus the beta-lactams than there is with the tobramycin/beta-lactam combination. Correspondingly, in studies of aminoglycoside-susceptible gram-negative rods and organisms selected for piperacillin resistance, amikacin was occasionally more active than tobramycin (Table V). Juty 15,1985
One debate has centered on the “relevance” of in vitro synergistic actions. In this regard (Table VI), we have evaluated the effects of amikacin and cefsulodin in mouse intraperitoneal infections caused by two strains of P. aeruginosa. With both challenges, evidence for antimicrobial synergism using conventional criteria [23] was observed. At subtherapeutic doses of amikacin and cefsulodin, the best survival was observed in those mice treated with combination therapy. IN VITRO SUSCEPTlBlLlTY OF PSEUDOMONAS AERUGINOSA TO AMIKACIN AND GENTAMICIN Over the last 13 years, amikacin has been used at the University of California, Los Angeles, Medical Center, first as an investigational agent and now as initial therapy of presumed gram-negative infections on an unrestricted basis. A major question is whether emergence of resistance to amikacin has occurred. We have restudied 96 isolates of Pseudomonas collected in the base period of 1972 to 1974, several intervening years, and, more recently, from 1962 to 1964. In order to obtain as much controlled data as possible, these organisms were taken from TABLE VI
Therapy of Experimental lntraperitoneal Infections Caused by Two Stratns of Pseudomonas aeruginosa
DNB Concentrations (K2)
Strain Fisher Type I* Sutvivon (total = 21)
Strain Fisher Me V’ survivors (total = 21)
2 21 8 5 1 16
1 20 3 2 2 14
None Amikacin (1.5 mg) Cefsulodin (80 mg) l/4 Amikacin (A) l/4 Cefsulodin (CFS) 114 A+ l/4 CFS ‘Both CFS.
challenge
organisms
were
The American Journal ot Medicine
individually
susceptible
to A and
Volume 79 (suppl 1A)
25
a32 _
.
16 -
0.
00
0
00.
0
0..
.
000
.
.
.
.
.
.
.
14 -
12 -
Minimum Inhibitory Concentration (dmL) (i4W
,O
66, _
0 .
00..
. .
000 000
. .
. .
.
66,
figure 1. Susceptibility patterns of 96 I? aeruginosa isolates showed no significant difference in amikacin susceptibility between the base period from 1972 to 1974 and the last collection period from 1982 to 1984. The geometric mean minimal inhibitory concentration was 4 pglml. In contrast, the geometric mean minimal inhibitory concentration for gentamicin declined from 6 mlml to 4 @ml during the same two collection periods.
44,
2 2, Sl i, I 1975-76
I 1979-M
1 1962-64
Year
a32
16
Minimum Inhibitory Concentration
1o 0
WW
0
6 1
.
.
Figure 2. Susceptibility patterns of 98 Kiebsiella blood isolates showed no significant changes in geometric mean minimal inhibitory concentrations for either amikacin or gentamicin between 1972 and 1984.
1975-77
Year
25
July
l5,199!i
The Amerlcen
Journal
of Medlclne
Volume
79 (suppl
1A)
AMINOGLYCOSIDE
storage, purified, and run simultaneously with cation-supplemented Mueller-Hinton medium. Interestingly, when the isolates from the collection periods were compared, there was no significant difference in amikacin susceptibility (Figure 1). During the period from 1972 to 1974, the geometric mean minimal inhibitory concentration of amikacin against blood isolates of Pseudomonas was 4 Fg/ml. This remained the same for the period from 1982 to 1984. In the base period of 1972 to 1974, the geometric mean minimal inhibitory concentration of gentamicin against Pseudomonas isolates was 8 a/ml. Interestingly, the geometric mean minimal inhibitory concentration of gentamicin against Pseudomonas isolates during 1982 to 1984 declined to 4 pg/ml, but this was not statistically significant. However, the issue of susceptibility breakpoints bears strongly on interpretation of these data. It is well established that peak blood levels of amikacin usually range anywhere from 18 to 25 @g/ml; this is roughly four times the geometric mean minimal inhibitory concentration of the average gentamicin minimal inhibitory concentration for a Pseudomonas isolate. Similar observations have been made for 98 Klebsiella
SYMPOSIUM-YOUNG
blood isolates collected at our institution over the last 14 years (Figure 2). No significant changes in geometric mean minimal inhibitory concentrations have occurred. There are a few highly resistant isolates among the pseudomonads isolated in recent years, but the appearance of these strains is not a statistically significant trend. The important questions are whether these resistant isolates originated from patients who had received aminoglycosides and, if so, which type of therapy the patient was previously exposed to (implying the mechanism by which resistance developed). Usually, antecedent therapy consisted of gentamicin or tobramycin. CONCLUSIONS
Appropriate use of an aminoglycoside such as amikacin has not led to rapid emergence of resistance, a phenomenon that has been observed with some beta-lactam compounds. In immunocompromised hosts, the initial selection of the most broadly active compounds used in combination is still to be strongly recommended. Such a view is hardly new but is a reaffirmation of principles (with further supporting data) that have been presented in other publications [3,24,25].
REFERENCES
Young LS: The clinicalchallengeof Pseudomonasaeruginosa infections.Rev Infect Dis 1964; G(suppl3): s603-~607. SodeyGP, BolivarR, FainsteinV: Infectiouscomplicationsin leukemiapatients.Semin Hematol1962; 19: 193-226. YoungLS: Problemsin determiningthe efficacyof aminoglycosides. Rev Infect Dis 1963; 5: s250-~257. BundtzenRW, GerberAV,Cohn DL, Craig WA: Postantibiotic suppressionof bacterialgrowth. Rev InfectDis 1961;3: 2837. McDonaldPJ,CraigWA,KuninCM: Persistenteffectof antibioticson Staphylococcus aureusafter exposurefor limitedperiods of time. J Infect Dis 1977; 135: 217-223.
9.
10.
11.
12.
13.
14.
Wilson DA, Rolinson GN: The recovery period following exposure of bacteria to penicillins. Chemotherapy 1979: 25: 1422. Valdivieso M, Feed R, Rodriguez V, Bodey GP: Amikacin therapy of infections in neutropenic patients. Am J Med Sci 1975; 270: 453-463. Gribble MJ, Chow AW, Naiman SC, Smith JA, et al: Prospective trial of piperacillin monotherapy versus carboxymethyl penicillin/aminoglycoside combination regimens. Antimicrob Agents Chemother 1983; 24: 388-393. Anderson ET, Young LS, Hewitt WL: Antimicrobial synergism in the therapy of gram-negative rod bacteremia. Chemotherapy 1977; 24: 45-64. Klastersky J, Zinner SH: Synergistic combinations of antibiotics in gram-negative bacillary infections. Rev Infect Dis 1982; 4: 294-301. Reyes MP, Lerner AM: Current problems in the treatment of infective endocarditis due to Pseudomonas aeruginosa. Rev Infect Dis 1982; 24: 45-54. Lau WK. Young LS, Black R, et al: Comparative efficacy and toxicity of amikacin/carbenicillin versus gentamicin/carbeniciIlin in leukopenic patients. Am J Med 1977; 62: 959-966. Love LJ, Schimpff SC, Schiffer CA, Wiernik PH: Improved prognosis for granulocylopenic patients with gram-negative bacteremia. Am J Med 1980; 68: 643-648. Young LS: Amikacin: experience in a comparative clinical trial
July
15, 1995
15.
16.
17.
18.
19.
20.
21.
22.
23.
24. 25.
ThOA
with gentamicin in leukopenic subjects. In: Cunent chemotherapy. Washington: American Society for Microbiology and International Society of Chemotherapy, 1978; 246-248. SandersCC, Sanders WE: Emergence of resistance during therapy with the newer B-lactam antibiotics: role of inducible B-lactamases and implications for the future. Rev Infect Dis 1983; 5: 639-648. Winston DJ, Busuttil RW, Kurtz TO, Young LS: fvloxalactam therapy of nosocomial infections. Rev Infect Dis 1982; 4(suppl): s850-~655. Oblinger MJ, Bowers JT, Sande MA, Mandell GL: Moxalactam therapy vs. standard antimicrobial therapy for selected serious infections. Rev Infect Dis 1982; 4(suppl): s639-s649. Smith CR, Ambinder R, Lipsky JJ, et al: Cefotaxime compared with nafcillin plus tobramycin for serious bacterial infections. Ann Intern Med 1984; 101: 469-477. Gordon AJ, Phyfferoen M: Cefoperazone sodium in the treatment of serious bacterial infections in 2100 adults and children. Rev Infect Dis 1983; B(suppl 1): s188-~199. Cohen MS, Washton HE, Barranco SF: Multicenter clinical trial of cefoperazone sodium in the United States. Am J Med 1984; 77(suppl 1 B): 35-41. Bolivar R, Fainstein V, Eking L, Bodey GP: Cefoperazone for the treatment of infections in patients with cancer. Rev Infect Dis 1983; S(suppl 1): s181-~187. Winston DW, Barnes RC, Ho W, Young LS, et al: Moxalactam @us piperacillin versus moxalactam plus amikacin in febrile granulocytopenic patients. Am J Med 1984; 77: 442-450. Young LS: Combination or single drug therapy for gram-negative sepsis. In: Remington JS, Swartz MN, eds. Current clinical topics in infectious diseases, vol 3. New York: McGrawHill, 1982; 177-205. Hewitt WL, Young LS: Symposium on amikacin: symposium perspective. Am J Med 1977; 62: 863-865. Young LS, Meyer-Dudnik D, Hindler J, Martin WJ: Aminoglycosides in treatment of bacteremic infections in the immunocompromised host. J Antimicrob Chemother 1981; B(suppl A): 121-132.
.merican
Journal
of Medicine
Volume
79 (suppl
1A)
27
,
LOWELL S. YOUNG, M.D. Los Angeles,
California
From the Department of infectious Diseases, Univereity of California School of Medicine, Center for Health Sciences, Los Angeles, California. Requests for reprints should be addressed to Dr. Lowell S. Young, Department of Medicine, UCLA Medical School, Center for Health Sciences, 10883 La Conte, Los Angeles, California 90024.
For much of the last decade, combination therapy with aminoglycosides has been accepted as the therapeutic approach of choice in immunocompromised hosts. improved clinical results have also correlated with the presence of synergistic interactions between the aminoglycoside and beta-lactam components of a regimen. Differences between the aminoglycosides and beta-la&am agents remain a subject of controversy. Studies at the University of California, Los Angeles, Medical Center suggest that amikacin interacts more frequently in a synergistic manner with beta-lactams than do alternative aminoglycosides. Amikacin has been used experimentally and (following llcensure) without reservation at the University of Caiifornia, Los Angeles, Medical Center since 1973. Almost 199 blood isolates of both Pseudomonas aeruginosa and Klebsiella pneumoniae collected during the last 12 years have been retested, and no evidence of increased aminoglycoside resistance was found. A relatively new development is Interest in empiric therapeutic regimens that employ two beta-lactam agents. In a large, recently completed study, less satisfactory results were observed in P. aeruginosa infections treated wlth the “double beta-lactam” than in those treated with the regimen containing amikacin; furthermore, nephrotoxicity and eighth nerve damage occurred no more commonly in the group receiving amikacin than in recipients of the double beta-lactam regimen. Aminoglycosides have been used to treat infections in immunocompromised patients for many years. In the 1950% prior to the availability of gentamicin, agents such as streptomycin and kanamycin were used frequently as initial therapy in these patients. In immunocompromised patients, the most important organism in terms of mortality, although not necessarily the most prevalent, has always been Pseudomonas aeruginosa. In fact, among university, county, cancer, and even community hospitals, P. aeruginosa is associated with the highestcase-fatality ratio of all causes of gram-negative rod bacteremia [l]. Several years ago, Bodey and colleagues [2] observed that survival from the onset of Pseudomonas sepsis in cancer patients was very poor. Often a significant number of these patients died within one or two days after the first positive blood culture result. For these reasons, therapy in this high-risk group of patients must be empiric: treatment must be initiated before the infecting organisms have been identified and their antimicrobial susceptibilities determined. In order to deal with gram-negative pathogens that might be present, many physicians employ aminoglycosides in the initial therapeutic regimen. However, there are many contro-
July
15, 1995
The
American
Journal
oi Medicine
Volume
79 (suppl
1A)
21
AMINOGLYCOSIDE
TABLE
I
SYMPOSIUM-YOUNG
Controversies
Aminoglycosldes
about
Aminogiycosides
811) ineffective In neutropenic
patients.
ISSUe
Adequacy of dose and achievement of therapeutic dose levels in vivo Better results are achieved by continuous infusion. ISSUeS
Supporting evidence scanty More toxicity in some studies Larger doses actually given Will comblnatlon use llmlt emergence of resistance to betaIactamsi Evidence Shown in some studies and may be related to duration of therapy Amlnoglycosldes will be replaoed by single agents, such as new cephalosporlns. Issue Suggested by some studies but issue unresolved in neutropenic and other high-risk patients
versies concerning the use of aminogiycosides, not only in immunocompromised patients but in other patients as well. Some of these controversies are highlighted in Table I. MAJOR
CONCERNS
ABOUT
AMINOGLYCOSIDES
Many have claimed that aminoglycosides are ineffective in neutropenic patients. However, it should be remembered that the ineffectiveness of a drug or class of drugs in vivo can be substantiated only by compelling and convincing evidence that adequate doses of the drug have been given and that therapeutic levels have actually been achieved. Many studies that claim aminogiycosides are ineffective in neutropenic patients have used subtherapeutic doses and have failed to demonstrate adequate blood levels [3]. Another controversy relating to aminogiycoside therapy involves the concept that better therapeutic results can be achieved by continuous infusion. Although this may seem plausible, a review of the literature suggests that, in fact, the beta-lactam antibiotics rather than aminoglycosides may give better therapeutic results when given in continuous infusion [2,3]. Craig and his colleagues [4,5] have studied the post-antibiotic effects of a variety of drugs. They found that lethal antibiotic activity of ail aminogiycosides is so rapid that, in many cases, there is a post-antibiotic effect of one to four hours against clinically important gram-negative organisms. Craig’s work, among others [6], suggests that when therapeutic levels of aminogiycosides fall below the minimal inhibitory concentration, organisms such as P. aeruginosa will start to grow again unless there is a significant post-antibiotic effect. This regrowth phenomenon, however, is much more striking with the beta-lactam antibiotics when studied in vitro. Therefore, it would seem that the application of the continuous infusion technique has been improperly directed and 22
July 15,1985
The American Journal of Medlclne
should, in fact, be applied more to the beta-iactam agents than to the aminoglycosides. In several studies, researchers have attempted to compare the efficacies of continuous and intermittent infusion of aminogiycosides. in one study by Bodey and associates [A on the continuous infusion of amikacin in neutropenic patients, results showed that more than half of the patients in whom amikacin levels were maintained at greater than 20 pg/mi subsequently showed some form of ototoxicity or nephrotoxicity. These results suggest that continuous exposure of vital eighth nerve and kidney structures to high levels of aminoglycosides increases risk of toxicity. A second issue of the continuous infusion controversy involves the amount of drug actually delivered. in an attempt to achieve sustained high levels of aminoglycosides [3], the amount of drug actually delivered is considerably greater than the amount the patient would have received by a conventional intermittent-dosing regimen. Thus, in those situations in which better clinical results seem to be obtained with continuous infusion, the improvement is likely due to the greater total daily dosage of drug delivered. IS EMERGENCE OF RESISTANCE COMBINATION THERAPY?
LESSENED
BY
Much attention has been paid to the growing concern that prolonged therapy with beta-iactam agents may cause the emergence of antimicrobial resistance. The concept has been advanced that combining a beta-lactam plus an aminogiycoside may limit the development of resistance to the beta-lactam. One of the major questions in current studies concerning any limitation of beta-iactam resistance by aminogiycosides is the duration of therapy. In many of these studies, the combination therapy is administered for only 10 to 14 days, so that significant differences in the rate of emergence of resistance would be very difficult to appreciate. The issue of emergence of resistance is certainly difficult to evaluate in most of the studies comparing immunocompromised and non-immunocompromised patients. One of the best-designed multicenter studies that addresses this issue was conducted in Canada, by Gribble and colleagues [8]. in this study, piperaciiiin was used alone and in combination with an aminogiycoside, usually gentamicin or tobramycin. When piperaciiiin was used as a single agent, it was associated with a significantly higher incidence of resistance (Table ii).
We have assessed therapeutic results in a retrospective study of P. aeruginosa bacteremia [9]. Combinations of the antimicrobial agents that exert synergistic antimicrobial activity against the actual infecting strains were associated with a significantly better therapeutic response. For example, 10 of 12 patients in whom synergistic interaction between gentamicin and carbeniciliin could be demonstrated showed clinical response. Cf the six patients whose isolates were not synergistically affected, none
Volume 79 (suppl 1A)
AMINOGLYCOSIDE
showed resoonse. Similar findinas have been reported bv Klastersky and Zinner [lo]. An especially interesting study was reported by Lerner and Reyes [Ill, in which they evaluated non-immunocompromised hosts with P. aeruginosa endocarditis. In their experience, the only patients with medical cures were those in whom the infecting Pseudomonas organism was synergistically inhibited and killed by the combination of an aminoglycoside and a -_ broad-spectrum penicillin. In a prospective study of empiric therapy of neutropenic patients conducted at the University of California, Los Angeles, Medical Center [12], the protocol called for either amikacin plus carbenicillin or gentamicin plus carbenicillin. Almost concurrently, Love et al [13] at the University of Maryland employed a similar protocol using ticarcillin as their antipseudomonal penicillin. If the results of these two studies are pooled, it becomes evident that the clinical response rate is over 80 percent when the pathogen isolated from blood is susceptible to both antibiotics used in gram-negative bacteremia in neutropenic patients. Although the issue of synergism remains unsettled, additive effects do exist and are likely to be beneficial. Experience at the University of California, Los Angeles, Medical Center [14] in analyzing individual isolate response versus clinical therapeutic response has shown that there is a statistically significant difference between inhibition by only one drug and inhibition by two drugs that exhibit a synergistic effect. Our results indicate that both in vitro synergism and additive effects are associated with improved therapeutic effects. NEW AGENTS: SINGLE AND COMBINED The majority of newly introduced antibiotics for systemic use in the United States are cephalosporins and related agents. Because of claims that these newer agents are more potent and have a broader spectrum than previously available antibiotics, many studies of single-agent therapy with these cephalosporins have been undertaken. Nevertheless, monotherapy with these newer agents in the treatment of some gram-negative bacillary infections caused by such organisms as Pseudomonas, Enterobacter, Serratia, and even Citrobacter has, in the experience of many investigators [15], been associated with a rapid emergence of resistance. In our study [16], the emergence of resistance in P. aeruginosa to single-agent therapy with moxalactam reached 35 percent; a somewhat lower rate of emergence of resistance was found in another study with cefoperazone (unpublished data). Some of our patients did show clinical improvement with moxalactam therapy despite emergence of resistance. However, improvement may have been due to a variety of factors unrelated to antimicrobial treatment per se. For example, a patient with a leg ulcer may have been treated with both antibiotics and surgical debridement, so that clinical response could not be attributed solely to antibiotic monotherapy. July
15, 1996
TABLE II
Emergence ol Resistance* Using Plperkillin Alone and in Combination with Aminoglycosldes Rsglmen
Piperacillin Piperacillin
SYMPOSIUM-YOUNG
Penent
alone + aminoglycoside
Number
42 17
26 24
“Emergence of resistance accounted for five of nine episodes ment failure. Reproduced with permission from [a].
of treat-
To support claims that single-agent therapy with newer cephalosporins-because of their greater potency and broad spectrum-will supplant the aminoglycosides, there are some seemingly well-executed clinical studies [17,18] in which treatment with a third-generation cephalosporin has been compared with combination therapy. In a recent study by Smith et al [la], which was not a strict comparison of monotherapy versus the same agent plus an aminoglycoside, the effects of monotherapy with cefotaxime were compared with those of combination therapy that included an antistaphylococcal agent plus an aminoglycoside, i.e., cefotaxime versus nafcillin plus tobramytin. The overall response rates were 90 percent and 60 percent, respectively. However, upon careful review of these data and the methodology employed in collecting them, some disproportionalities become evident. In this double-blind, randomized, prospective study, there were relatively few gram-negative rod infections, particularly of the bloodstream [la]. The entire study included a total of nine Pseudomonas infections, seven Enterobacter infections, and 10 Klebsiella infections. Further analysis of these 26 infections shows that 20 were treated with nafcillin plus tobramycin, whereas only six received cefotaxime alone. This difference is statistically significant. Additionally, this study enrolled no neutropenic patients. Studies such as this one, however excellent the clinical design, have not convincingly established that third-generation cephalosporins like cefotaxime can replace combination therapy with aminoglycosides for serious gram-negative infections in neutropenic patients. The excellent results of recent studies of cefoperazone in gram-negative bacteremia would, by any measure, be considered impressive. Response rates for bacteremia range from 84 to 100 percent in summaries of case reports collected during the course of some clinical trials [19] conducted before licensure in the United States. It should be borne in mind, however, that evaluability criteria can affect the interpretation of summary data. For example, in a multicenter study conducted by Cohen et al [20], a case was not considered evaluable if a patient received cefoperazone therapy for fewer than five days. Thus, by eliminating those patients who received short-duration therapy, they achieved a response rate of 100 percent in the treatment of gram-negative bacteremia. In contrast, The American
Journal
of Medlclne
Volume
79 (suppl
1A)
23
AMINOGLYCOSIDE
TABLE
III
SYMPOSIUM-YOUNG
Significant interactions between Lactam Antibiotics Demonstrated and in Animal Models
Betain Vitro
DOUBLE
BETA-LACTAM
THERAPY
Bolivar and associates [21] studied cefoperazone in the treatment of bacteremia occurring in neutropenic patients and attained a clinical response rate of 50 percent for Enterobacter infections, 33 percent for Klebsiella, and 50 percent for PSeUdDmOnaSinfections. The major apparent difference between the results of the multicenter cefoperazone study summarized by Cohen et al [20] and the report of Bolivar and colleagues [21] is that the patients in the latter study were neutropenic.
The most recent innovation in empiric antimicrobial therapy in febrile neutropenic patients is “double beta-lactam” treatment. This therapy entails the use of a broad-spectrum penicillin plus a cephalosporin, rather than a betalactam agent with an aminoglycoside. In a study conducted at the University of California, Los Angeles, Medical Center [22], we compared moxalactam plus piperacillin with moxalactam plus amikacin, the latter being considered “conventional therapy.” In patients with Pseudomonas infections, only one of five receiving the double beta-lactam regimen showed a favorable response, as opposed to seven of nine treated with moxalactam plus amikacin. This difference misses attaining statistical significance by only one one-hundredth of a point (p = 0.06). Also in this study, two patients with Pseudomonas bacteremia who had initially had susceptible isolates of P. aeruginosa experienced a relapse of bacteremia associated with multiply beta-lactam-resistant isolates. This is highly suggestive of the phenomenon of inducible beta-lactamase resistance that has been well studied by Sanders and Sanders [15]. Mechanisms for this phenomenon are summarized in Table III. An important point regarding the treatment of immunocompromised patients with double beta-lactam therapy concerns the incidence and severity of neutropenia in these patients. In our study, there was significantly more severe neutropenia among patients receiving the double beta-lactam regimen. It is well established that beta-lactam agents cause neutropenia, which is almost always reversible. Therefore, we must be concerned that use of two beta-lactam agents may actually prolong neutropenia and augment risk of superinfection.
EORTC DATA
NEPHROTOXICITY
Data from the European Organization for the Research and Treatment of Cancer (EORTC) have been discussed at length elsewhere in these symposium proceedings. However, the relevance of their results to this discussion pertains only to one infection-P. aeruginosa occurring in Trial I and Trial III. The last study showed clearly that combination therapy using amikacin plus a broad-spectrum antipseudomonal penicillin, such as azlocillin or ticarcillin, is statistically significantly better than combination therapy with a cephalosporin-even one of the most popular of the new agents, cefotaxime.
In another study at the University of California, Los Angeles, Medical Center, we compared the combination of amikacin plus carbenicillin with the combination of netilmitin plus carbenicillin in more than 230 patients in order to examine the incidence and nature of nephrotoxicity (L.S. Young and W.L. Hewitt, unpublished data). We chose netilmicin because it had been hoped, on the basis of animal studies, that it would be a less nephrotoxic agent than the other aminoglycosides. Our results did not support this contention, however. Our definition of nephrotoxicity required a rise in serum creatinine level of at least 60 per-
Synergism
Antagonism
Mechanism Inactivation of beta-lactamase by inhibitor; enzyme-labile beta-lactam reaches lethal target in cell
Induction of beta-lactam hydrolyzing enzymes
Attack on different lethal targets (penicillin-binding proteins) in cell
Induction of beta-lactamasemediated barrier that blocks access to lethal targets in cell
Examples of Organisms Involved Staphylococcus aureus, gram-negative bacteria producing types II, Ill, IV, or V beta-lactamases
Enterobacteriaceae possessing inducible type I beta-lactamases (e.g., Enterobacter, Serratia, Ciirobacter)
Enterobacteriaceae
Enterobacteriaceae Pseudomonas inducible type lactamases
Adapted
with permission
TABLE
IV
from
(151.
Synergistic Inhibition and/or Killing of 30 Gentamlcin-Reslstant Pseudomonas Amikacin or Tobramycln Plus Any of Six Beta-Lactam Antlmicroblals Ticanillin
Amikacin Tobramycin Total synergism ‘30 strains
24
and possessing I beta-
Thienamycin
Cafaulodin
Moxalacfam
15
0
7 22
2 2
6 5 11
16 13 31
per group.
July 15, 1995
The American Joumel of Medicine
Volume 79 (suppl 1A)
aeruginosa’ Csfoparaxana 22 20 42
by Cefofaxlma 25 20 45
AMINOGLYCOSIDE
TABLE V
In Vitro Synergism against AminoglycosidcSusceptible
Number of strains tested Amikacin+ Ticarcillin Moxalactam Cefoperazone Cefotaxime Tobramycin+ Ticarcillin Moxalactam Cefoperazone Cefotaxime ‘All blood isolates were chosen because ‘K-E-S group: Klebsiella-Enterobacter-Serraia
SYMPOSIUM-YOUNG
Gram-Negative Rods*
pseudomonas aerusinosa
Eschsrichia coli
K-E-S’ group
Acinetobacter
Providencia
13
9
15
3
9
12 13 12 13 8 13 11 13
5 8 7 of piperacillin group
resistance.
cent to a concentration of 1.3 mg/dl or more. Of the 21 patients who met this criterion of nephrotoxicity, we found only two in whom the rise in serum creatinine was attributable only to the aminoglycoside with which they had been treated. In the great majority of cases, nephrotoxicity occurred in patients who were hypotensive, receiving amphotericin B, cisplatin, diuretics, radiologic contrast media, or other potentially responsible agents. Therefore, much of what has been attributed to aminoglycoside toxicity could be due to important cofactors. RELEVANCE OF SYNERGISTIC ACTIONS OF ANTIMICROBIAL COMBINATIONS Numerous synergism studies have been conducted both in vitro and in vivo. Recent interest has been focused on newer agents, such as imipenem and cefsulodin. It is interesting to note that despite the potency of imipenem, for example, it is quite difficult to demonstrate synergistic interaction with aminoglycosides. One explanation for this may be that the killing of P. aeruginosa by imipenem is so complete that there are no surviving subpopulations for the concomitant aminoglycoside to kill. In the study represented in Table IV, of 30 gentamicin-resistant strains of P. aeruginosa, more synergistic effects were observed with the less potent antipseudomonal compounds than with imipenem. Although there seems to be little synergism with imipenem, there is far greater synergism evidenced with cefoperazone or cefotaxime. This phenomenon may, in fact, be related to the incomplete killing with the less potent agents (cefoperazone, cefotaxime). It seems clear that, with these selected strains of P. aeruginosa (selected for gentamicin resistance), there is somewhat more synergism with amikacin plus the beta-lactams than there is with the tobramycin/beta-lactam combination. Correspondingly, in studies of aminoglycoside-susceptible gram-negative rods and organisms selected for piperacillin resistance, amikacin was occasionally more active than tobramycin (Table V). Juty 15,1985
One debate has centered on the “relevance” of in vitro synergistic actions. In this regard (Table VI), we have evaluated the effects of amikacin and cefsulodin in mouse intraperitoneal infections caused by two strains of P. aeruginosa. With both challenges, evidence for antimicrobial synergism using conventional criteria [23] was observed. At subtherapeutic doses of amikacin and cefsulodin, the best survival was observed in those mice treated with combination therapy. IN VITRO SUSCEPTlBlLlTY OF PSEUDOMONAS AERUGINOSA TO AMIKACIN AND GENTAMICIN Over the last 13 years, amikacin has been used at the University of California, Los Angeles, Medical Center, first as an investigational agent and now as initial therapy of presumed gram-negative infections on an unrestricted basis. A major question is whether emergence of resistance to amikacin has occurred. We have restudied 96 isolates of Pseudomonas collected in the base period of 1972 to 1974, several intervening years, and, more recently, from 1962 to 1964. In order to obtain as much controlled data as possible, these organisms were taken from TABLE VI
Therapy of Experimental lntraperitoneal Infections Caused by Two Stratns of Pseudomonas aeruginosa
DNB Concentrations (K2)
Strain Fisher Type I* Sutvivon (total = 21)
Strain Fisher Me V’ survivors (total = 21)
2 21 8 5 1 16
1 20 3 2 2 14
None Amikacin (1.5 mg) Cefsulodin (80 mg) l/4 Amikacin (A) l/4 Cefsulodin (CFS) 114 A+ l/4 CFS ‘Both CFS.
challenge
organisms
were
The American Journal ot Medicine
individually
susceptible
to A and
Volume 79 (suppl 1A)
25
a32 _
.
16 -
0.
00
0
00.
0
0..
.
000
.
.
.
.
.
.
.
14 -
12 -
Minimum Inhibitory Concentration (dmL) (i4W
,O
66, _
0 .
00..
. .
000 000
. .
. .
.
66,
figure 1. Susceptibility patterns of 96 I? aeruginosa isolates showed no significant difference in amikacin susceptibility between the base period from 1972 to 1974 and the last collection period from 1982 to 1984. The geometric mean minimal inhibitory concentration was 4 pglml. In contrast, the geometric mean minimal inhibitory concentration for gentamicin declined from 6 mlml to 4 @ml during the same two collection periods.
44,
2 2, Sl i, I 1975-76
I 1979-M
1 1962-64
Year
a32
16
Minimum Inhibitory Concentration
1o 0
WW
0
6 1
.
.
Figure 2. Susceptibility patterns of 98 Kiebsiella blood isolates showed no significant changes in geometric mean minimal inhibitory concentrations for either amikacin or gentamicin between 1972 and 1984.
1975-77
Year
25
July
l5,199!i
The Amerlcen
Journal
of Medlclne
Volume
79 (suppl
1A)
AMINOGLYCOSIDE
storage, purified, and run simultaneously with cation-supplemented Mueller-Hinton medium. Interestingly, when the isolates from the collection periods were compared, there was no significant difference in amikacin susceptibility (Figure 1). During the period from 1972 to 1974, the geometric mean minimal inhibitory concentration of amikacin against blood isolates of Pseudomonas was 4 Fg/ml. This remained the same for the period from 1982 to 1984. In the base period of 1972 to 1974, the geometric mean minimal inhibitory concentration of gentamicin against Pseudomonas isolates was 8 a/ml. Interestingly, the geometric mean minimal inhibitory concentration of gentamicin against Pseudomonas isolates during 1982 to 1984 declined to 4 pg/ml, but this was not statistically significant. However, the issue of susceptibility breakpoints bears strongly on interpretation of these data. It is well established that peak blood levels of amikacin usually range anywhere from 18 to 25 @g/ml; this is roughly four times the geometric mean minimal inhibitory concentration of the average gentamicin minimal inhibitory concentration for a Pseudomonas isolate. Similar observations have been made for 98 Klebsiella
SYMPOSIUM-YOUNG
blood isolates collected at our institution over the last 14 years (Figure 2). No significant changes in geometric mean minimal inhibitory concentrations have occurred. There are a few highly resistant isolates among the pseudomonads isolated in recent years, but the appearance of these strains is not a statistically significant trend. The important questions are whether these resistant isolates originated from patients who had received aminoglycosides and, if so, which type of therapy the patient was previously exposed to (implying the mechanism by which resistance developed). Usually, antecedent therapy consisted of gentamicin or tobramycin. CONCLUSIONS
Appropriate use of an aminoglycoside such as amikacin has not led to rapid emergence of resistance, a phenomenon that has been observed with some beta-lactam compounds. In immunocompromised hosts, the initial selection of the most broadly active compounds used in combination is still to be strongly recommended. Such a view is hardly new but is a reaffirmation of principles (with further supporting data) that have been presented in other publications [3,24,25].
REFERENCES
Young LS: The clinicalchallengeof Pseudomonasaeruginosa infections.Rev Infect Dis 1964; G(suppl3): s603-~607. SodeyGP, BolivarR, FainsteinV: Infectiouscomplicationsin leukemiapatients.Semin Hematol1962; 19: 193-226. YoungLS: Problemsin determiningthe efficacyof aminoglycosides. Rev Infect Dis 1963; 5: s250-~257. BundtzenRW, GerberAV,Cohn DL, Craig WA: Postantibiotic suppressionof bacterialgrowth. Rev InfectDis 1961;3: 2837. McDonaldPJ,CraigWA,KuninCM: Persistenteffectof antibioticson Staphylococcus aureusafter exposurefor limitedperiods of time. J Infect Dis 1977; 135: 217-223.
9.
10.
11.
12.
13.
14.
Wilson DA, Rolinson GN: The recovery period following exposure of bacteria to penicillins. Chemotherapy 1979: 25: 1422. Valdivieso M, Feed R, Rodriguez V, Bodey GP: Amikacin therapy of infections in neutropenic patients. Am J Med Sci 1975; 270: 453-463. Gribble MJ, Chow AW, Naiman SC, Smith JA, et al: Prospective trial of piperacillin monotherapy versus carboxymethyl penicillin/aminoglycoside combination regimens. Antimicrob Agents Chemother 1983; 24: 388-393. Anderson ET, Young LS, Hewitt WL: Antimicrobial synergism in the therapy of gram-negative rod bacteremia. Chemotherapy 1977; 24: 45-64. Klastersky J, Zinner SH: Synergistic combinations of antibiotics in gram-negative bacillary infections. Rev Infect Dis 1982; 4: 294-301. Reyes MP, Lerner AM: Current problems in the treatment of infective endocarditis due to Pseudomonas aeruginosa. Rev Infect Dis 1982; 24: 45-54. Lau WK. Young LS, Black R, et al: Comparative efficacy and toxicity of amikacin/carbenicillin versus gentamicin/carbeniciIlin in leukopenic patients. Am J Med 1977; 62: 959-966. Love LJ, Schimpff SC, Schiffer CA, Wiernik PH: Improved prognosis for granulocylopenic patients with gram-negative bacteremia. Am J Med 1980; 68: 643-648. Young LS: Amikacin: experience in a comparative clinical trial
July
15, 1995
15.
16.
17.
18.
19.
20.
21.
22.
23.
24. 25.
ThOA
with gentamicin in leukopenic subjects. In: Cunent chemotherapy. Washington: American Society for Microbiology and International Society of Chemotherapy, 1978; 246-248. SandersCC, Sanders WE: Emergence of resistance during therapy with the newer B-lactam antibiotics: role of inducible B-lactamases and implications for the future. Rev Infect Dis 1983; 5: 639-648. Winston DJ, Busuttil RW, Kurtz TO, Young LS: fvloxalactam therapy of nosocomial infections. Rev Infect Dis 1982; 4(suppl): s850-~655. Oblinger MJ, Bowers JT, Sande MA, Mandell GL: Moxalactam therapy vs. standard antimicrobial therapy for selected serious infections. Rev Infect Dis 1982; 4(suppl): s639-s649. Smith CR, Ambinder R, Lipsky JJ, et al: Cefotaxime compared with nafcillin plus tobramycin for serious bacterial infections. Ann Intern Med 1984; 101: 469-477. Gordon AJ, Phyfferoen M: Cefoperazone sodium in the treatment of serious bacterial infections in 2100 adults and children. Rev Infect Dis 1983; B(suppl 1): s188-~199. Cohen MS, Washton HE, Barranco SF: Multicenter clinical trial of cefoperazone sodium in the United States. Am J Med 1984; 77(suppl 1 B): 35-41. Bolivar R, Fainstein V, Eking L, Bodey GP: Cefoperazone for the treatment of infections in patients with cancer. Rev Infect Dis 1983; S(suppl 1): s181-~187. Winston DW, Barnes RC, Ho W, Young LS, et al: Moxalactam @us piperacillin versus moxalactam plus amikacin in febrile granulocytopenic patients. Am J Med 1984; 77: 442-450. Young LS: Combination or single drug therapy for gram-negative sepsis. In: Remington JS, Swartz MN, eds. Current clinical topics in infectious diseases, vol 3. New York: McGrawHill, 1982; 177-205. Hewitt WL, Young LS: Symposium on amikacin: symposium perspective. Am J Med 1977; 62: 863-865. Young LS, Meyer-Dudnik D, Hindler J, Martin WJ: Aminoglycosides in treatment of bacteremic infections in the immunocompromised host. J Antimicrob Chemother 1981; B(suppl A): 121-132.
.merican
Journal
of Medicine
Volume
79 (suppl
1A)
27
,