422
Editorial Response: Should All Patients with Candidemia Be Treated with Antifungal Agents?
See the article by Fraser et al. on pages 414-21. In the past, many clinicians have stratified candidemic patients according to the probability of their having disseminated candidiasis and have withheld antifungal therapy from those individuals for whom the probability was low. This strategy was utilized mainly for two reasons: (1) two studies [I, 2] were frequently quoted as having established Candida as a relatively non consequential organism and (2) the toxicity of amphotericin B was considered formidable. In 1974, on the basis oftheir study of70 cases offungemia in immunocompromised hosts, Young et al. [3] recommended treating all candidemic patients who were neutropenic. However, the practice of treating all candidemic patients, regardless of whether they are neutropenic, has not been generally advocated until recently, when data on the epidemiology of Candida became available. Extremely interesting data exist in the careful study of Fraser et al. During the year between 1988 and 1989,28% of the patients with candidemia did not receive any antifungal therapy. Comparable data from other medical centers for the same period are not readily available. However, this percentage of untreated patients is probably representative of trends in most centers. The overall mortality for candidemic patients described by Fraser et al. was 57%, and the overall
Received I June 1992. Reprints or correspondence: Dr. John E. Edwards. Jr., Division of Infectious Diseases. Department of Medicine. Harbor-UCLA Medical Center. 1000 West Carson Street. Torrance. California 90502.
Clinical Infectious Diseases 1992;15:422-3 © 1992 by The University of Chicago. All rights reserved. 1058-4838/92/1503-0006$02.00
mortality for those patients who had candidemia but did not receive therapy was 63%. While the data for attributable mortality due to Candida in untreated candidemic patients were not presented in detail, extrapolation from the studies of Wey et al. [4] and Komshian et al. [5] suggests that the attributable mortality was at least 40%. In fact, it was likely higher, since the patients selected for the group in which antifungal therapy was withheld were undoubtedly considered to be at low risk for death because of having underlying illnesses of lesser severity. These patients were also undoubtedly considered to be at low risk for disseminated candidiasis. Two facts emerge from retrospective analysis ofthese data: (I) a 63% mortality rate is unacceptably high for a group of patients from whom therapy was withheld and (2) predicting which patients for whom the probability of having disseminated candidiasis was low was generally unsuccessful. Furthermore, mortality was the only focus of this analysis; the incidence of late complications of candidemia-such as renal infection, osteomyelitis, and endophthalmitis-was not determined. Long-term follow-up was not included in the experimental design. The high mortality rate among the untreated candidemic patients in the study by Fraser et al. vividly underscores both the dangers in not treating patients with candidemia and the difficulty in attempting to stratify patients according to their likelihood of having infection of deep organs with Candida. Until methods to stratify patients more accurately are developed, it is prudent to treat all patients who have candidemia, regardless of whether they are neutropenic and regardless of whether the candidemia is associated with the presence of a catheter. Some form of antifungal chemotherapy should be given, unless there is a specific contraindication to the available antifungal agents. This more aggressive approach may be modified when better serodiagnostic, radiographic, and culture techniques are developed to improve patient stratification. Until such time, however, amphotericin B should be given to those patients for whom the probability of having disseminated candidiasis is high or moderate. Fluconazole has considerably less toxicity than amphotericin B, and preliminary, noncomparative studies of its use in treatment of candidemia are favorable. However, there are no published comparative studies offluconazole and amphotericin B. Within the perspective of the data of Fraser et al., the large majority of candidemic patients should be treated with amphotericin B until comparative data WIth fluconazole and amphotericin B are available for analysis. A large, multicenter, national cooperative study comparing fluconazole to amphotericin B for treatment ofcandidemic patients who are not neutropenic is now in progress. Data from this study will
Downloaded from http://cid.oxfordjournals.org/ at University of Glasgow on June 24, 2015
In this issue of CID. Fraser et al. have presented a careful, in-depth analysis of the risk factors, therapy, and outcome for patients with candidemia at a large urban medical center during a l-year period between 1988 and 1989. In their discussion they state, "The significant mortality associated with most episodes of candidemia underscores the need to diagnose and treat all candidemic patients quickly. The total dose and duration of antifungal therapy should be determined several days into therapy, after the patient's candidemia has been identified as either transient or sustained." This statement suggests an approach to candidemia that is nearly opposite to the conventional, less-aggressive therapeutic approach ofprevious years. In light ofthe developing epidemiological data on candidemia, the recommendation of Fraser et al. is highly meritorious.
GO 1992;15 (September)
Editorial
John E. Edwards, Jr. Division ofInfectious Diseases. Harbor/ UCLA Medical Center. and Department of Medicine. UCLA School ofMedicine. Los Angeles. California
References I. Ellis CA. Spivack ML. The significance of candidemia. Ann Intern Med
1967;67:511. 2. Toala P. Schroeder SA, Daly AK. Finland M. Candida at Boston City Hospital. Clinical and epidemiological characteristics and susceptibility to eight antimicrobial agents. Arch Intern Med 1970; 126:983. 3. Young RC, Bennett JE. Geelhoed GW. Fungemia with compromised host resistance: a study ono cases. Ann Intern Med 1974;80:605-12. 4. Wey SB. Mori M. pfaller MA, Woolson RF. Wenzel RP. Hospital-acquired candidemia: the attributable mortality and excess length of stay. Arch Intern Med 1988; 148:2642-5. 5. Komshian SY. Uwaydah AK. Sobel JD. Crane LR. Fungemia caused by Candida species and Torulopsis glabrata in the hospitalized patient: frequency. characteristics. and evaluation of factors influencing outcome. Rev Infect Dis 1989; II :379-90.
Downloaded from http://cid.oxfordjournals.org/ at University of Glasgow on June 24, 2015
be available for analysis in --1 year. Until then, amphotericin B will remain the mainstay of therapy both for patients with candidemia and for patients who may not be candidemic but have all the predisposing factors for disseminated candidiasis and have not responded to antibacterial agents. However, fluconazole represents a nontoxic alternative to amphotericin B that can be used in patients who are candidemic but for whom the probability of having disseminated disease is very low. The introduction of this nontoxic agent has facilitated formulation of the recommendation that all candidemic patients receive antifungal agents rather than having a subset of patients receive no therapy. Fraser et al. appropriately state that in regard to the question of which antifungal therapy should be given to specific patients with suspected disseminated candidias, "future researchers of candidemia should stratify patients by severity of illness and duration of candidemia as well as by prognosis of underlying illness to determine optimum treatment regimens."
423
Editorial Response: Should All Patients with Candidemia Be Treated with Antifungal Agents?
See the article by Fraser et al. on pages 414-21. In the past, many clinicians have stratified candidemic patients according to the probability of their having disseminated candidiasis and have withheld antifungal therapy from those individuals for whom the probability was low. This strategy was utilized mainly for two reasons: (1) two studies [I, 2] were frequently quoted as having established Candida as a relatively non consequential organism and (2) the toxicity of amphotericin B was considered formidable. In 1974, on the basis oftheir study of70 cases offungemia in immunocompromised hosts, Young et al. [3] recommended treating all candidemic patients who were neutropenic. However, the practice of treating all candidemic patients, regardless of whether they are neutropenic, has not been generally advocated until recently, when data on the epidemiology of Candida became available. Extremely interesting data exist in the careful study of Fraser et al. During the year between 1988 and 1989,28% of the patients with candidemia did not receive any antifungal therapy. Comparable data from other medical centers for the same period are not readily available. However, this percentage of untreated patients is probably representative of trends in most centers. The overall mortality for candidemic patients described by Fraser et al. was 57%, and the overall
Received I June 1992. Reprints or correspondence: Dr. John E. Edwards. Jr., Division of Infectious Diseases. Department of Medicine. Harbor-UCLA Medical Center. 1000 West Carson Street. Torrance. California 90502.
Clinical Infectious Diseases 1992;15:422-3 © 1992 by The University of Chicago. All rights reserved. 1058-4838/92/1503-0006$02.00
mortality for those patients who had candidemia but did not receive therapy was 63%. While the data for attributable mortality due to Candida in untreated candidemic patients were not presented in detail, extrapolation from the studies of Wey et al. [4] and Komshian et al. [5] suggests that the attributable mortality was at least 40%. In fact, it was likely higher, since the patients selected for the group in which antifungal therapy was withheld were undoubtedly considered to be at low risk for death because of having underlying illnesses of lesser severity. These patients were also undoubtedly considered to be at low risk for disseminated candidiasis. Two facts emerge from retrospective analysis ofthese data: (I) a 63% mortality rate is unacceptably high for a group of patients from whom therapy was withheld and (2) predicting which patients for whom the probability of having disseminated candidiasis was low was generally unsuccessful. Furthermore, mortality was the only focus of this analysis; the incidence of late complications of candidemia-such as renal infection, osteomyelitis, and endophthalmitis-was not determined. Long-term follow-up was not included in the experimental design. The high mortality rate among the untreated candidemic patients in the study by Fraser et al. vividly underscores both the dangers in not treating patients with candidemia and the difficulty in attempting to stratify patients according to their likelihood of having infection of deep organs with Candida. Until methods to stratify patients more accurately are developed, it is prudent to treat all patients who have candidemia, regardless of whether they are neutropenic and regardless of whether the candidemia is associated with the presence of a catheter. Some form of antifungal chemotherapy should be given, unless there is a specific contraindication to the available antifungal agents. This more aggressive approach may be modified when better serodiagnostic, radiographic, and culture techniques are developed to improve patient stratification. Until such time, however, amphotericin B should be given to those patients for whom the probability of having disseminated candidiasis is high or moderate. Fluconazole has considerably less toxicity than amphotericin B, and preliminary, noncomparative studies of its use in treatment of candidemia are favorable. However, there are no published comparative studies offluconazole and amphotericin B. Within the perspective of the data of Fraser et al., the large majority of candidemic patients should be treated with amphotericin B until comparative data WIth fluconazole and amphotericin B are available for analysis. A large, multicenter, national cooperative study comparing fluconazole to amphotericin B for treatment ofcandidemic patients who are not neutropenic is now in progress. Data from this study will
Downloaded from http://cid.oxfordjournals.org/ at University of Glasgow on June 24, 2015
In this issue of CID. Fraser et al. have presented a careful, in-depth analysis of the risk factors, therapy, and outcome for patients with candidemia at a large urban medical center during a l-year period between 1988 and 1989. In their discussion they state, "The significant mortality associated with most episodes of candidemia underscores the need to diagnose and treat all candidemic patients quickly. The total dose and duration of antifungal therapy should be determined several days into therapy, after the patient's candidemia has been identified as either transient or sustained." This statement suggests an approach to candidemia that is nearly opposite to the conventional, less-aggressive therapeutic approach ofprevious years. In light ofthe developing epidemiological data on candidemia, the recommendation of Fraser et al. is highly meritorious.
GO 1992;15 (September)
Editorial
John E. Edwards, Jr. Division ofInfectious Diseases. Harbor/ UCLA Medical Center. and Department of Medicine. UCLA School ofMedicine. Los Angeles. California
References I. Ellis CA. Spivack ML. The significance of candidemia. Ann Intern Med
1967;67:511. 2. Toala P. Schroeder SA, Daly AK. Finland M. Candida at Boston City Hospital. Clinical and epidemiological characteristics and susceptibility to eight antimicrobial agents. Arch Intern Med 1970; 126:983. 3. Young RC, Bennett JE. Geelhoed GW. Fungemia with compromised host resistance: a study ono cases. Ann Intern Med 1974;80:605-12. 4. Wey SB. Mori M. pfaller MA, Woolson RF. Wenzel RP. Hospital-acquired candidemia: the attributable mortality and excess length of stay. Arch Intern Med 1988; 148:2642-5. 5. Komshian SY. Uwaydah AK. Sobel JD. Crane LR. Fungemia caused by Candida species and Torulopsis glabrata in the hospitalized patient: frequency. characteristics. and evaluation of factors influencing outcome. Rev Infect Dis 1989; II :379-90.
Downloaded from http://cid.oxfordjournals.org/ at University of Glasgow on June 24, 2015
be available for analysis in --1 year. Until then, amphotericin B will remain the mainstay of therapy both for patients with candidemia and for patients who may not be candidemic but have all the predisposing factors for disseminated candidiasis and have not responded to antibacterial agents. However, fluconazole represents a nontoxic alternative to amphotericin B that can be used in patients who are candidemic but for whom the probability of having disseminated disease is very low. The introduction of this nontoxic agent has facilitated formulation of the recommendation that all candidemic patients receive antifungal agents rather than having a subset of patients receive no therapy. Fraser et al. appropriately state that in regard to the question of which antifungal therapy should be given to specific patients with suspected disseminated candidias, "future researchers of candidemia should stratify patients by severity of illness and duration of candidemia as well as by prognosis of underlying illness to determine optimum treatment regimens."
423
