Supported by an unrestricted educational grant from Pfizer, Inc.
Volume III, Issue 1 - April 2000
Supported by an unrestricted educational grant from Pfizer, Inc.
The risk factors for candidemia are well known (Table 1), and the most important of these is the presence of an indwelling intravascular catheter. Other significant risk factors include cancer chemotherapy with the attendant compromise of the mucosal barriers, abdominal surgery, treatment with antibacterial agents, candiduria, and injection drug use. The vast majority of candidemias are nosocomial infections, and Candida species account for between 5-10% of nosocomial blood stream infections. The source for many candidemias is the patient's endogenous flora. However, the transmission of Candida species from health care workers to patients has been demonstrated. True "community-acquired" candidemias in the absence of an IV catheter or recent chemotherapy should prompt consideration of either injection drug use or a complicated candiduria with hematogenous seeding.
Candidemias are associated with significant attributable mortality and prolonged hospitalization. Although there are many patients who had complete recoveries from candidemia without any specific therapy, we cannot predict which candidemic patients will not require antifungals. Therefore, since this is a serious infection and we now have safe, alternative therapies, in our opinion all candidemic patients should receive antifungal therapy (Fig. 1).
In practical terms, the present choice of antifungal is between fluconazole and amphotericin B. Both of these antifungals have solid clinical data supporting their use to treat candidemia, and there are important reasons detailed below why the more toxic amphotericin B should be used preferentially over the more convenient fluconazole. Itraconazole is another orally available azole antifungal, but experience using this particular azole to treat candidemias is limited. Flucytosine has a significant role in treating candidemias when used in combination with amphotericin B, and this combination should be considered for candidemic patients who are immunocompromised or have evidence of deep-seated infection. However, the bone marrow suppression associated with flucytosine and its lack of an intravenous preparation significantly limit its use, especially in critically ill cancer patients.
Using the combination of fluconazole and amphotericin B to treat candidemia is controversial. There are concerns that these two agents may be antagonistic, although there are some animal data and limited human data supporting the combination. Clinical trials investigating this combination are currently underway, and until these studies are analyzed the combination of the two agents to treat candidemias should be reserved for patients failing high-dose amphotericin B.
Using cytokine agents, such as granulocyte colony-stimulating factor or granulocyte-macrophage colony-stimulating factor, as adjuncts in the treatment of candidemia is recommended in neutropenic patients with candidemia. There are also some data showing that these cytokine agents can be helpful in non-neutropenic patients with candidemia. It appears that these agents augment macrophage-mediated killing of Candida, clinical trials using these agents in non-neutropenic patients are in progress.
Several important factors need to be considered in order to determine the choice of antifungal of therapy in candidemic patients. One important consideration in choosing the antifungal is the patient's clinical presentation. Patients presenting with hypotension and the systemic inflammatory response syndrome secondary to candidemia should be treated initially with amphotericin B given its broad-spectrum fungicidal activities. Significant renal dysfunction is a reasonable indication for the use of one of the lipid-associated amphotericin B preparations. Fluconazole can be given as initial therapy to candidemic patients presenting with a stable clinical course at the time of diagnosis. Patients who are neutropenic or have some other significant immunocompromising condition such as malignancies or organ transplantation should probably also receive amphotericin B as the initial treatment regardless of their clinical presentation. Switching to fluconazole in a stable patient that has infection due to a susceptible isolate of Candida is recommended. The reasons for using amphotericin B as initial therapy are the same as those given for the patients presenting with the sepsis syndrome. It may be especially important to have the fungicidal activities of amphotericin B (rather than the fungistatic properties of fluconazole) in the presence of significant host immune defects. Candidemia occurring in the setting of recent recovery from neutropenia should raise the possibility of chronic disseminated candidiasis, and should prompt imaging studies to look for the associated multiple abscesses in the liver, spleen and/or kidneys.
The species of Candida isolated from the blood is another important determinant of therapy since there are some species associated with resistance to fluconazole. Data correlating in vitro susceptibilities of yeast to clinical outcome are still limited, so the term "resistant" when applied to yeast is imprecise. However, generalizations about specific therapy for certain species can be made. When Candida species are found in blood cultures, the microbiology laboratory can rapidly make the distinction between C. albicans and non-albicans Candida species using the germ tube assay. Most candidemias are due to C. albicans, and almost all of these isolates are susceptible to fluconazole. Candidemias due to non-albicans Candida species are growing in number, and are significant because these isolates have variable fluconazole susceptibilities. Ideally, all candidemias due to non-albicans Candida species should empirically be treated with amphotericin B until final identification of the isolate is made. Candida krusei is intrinsically resistant to fluconazole, and infections due to this organism should only be treated with amphotericin B. Candida glabrata may be clinically resistant to fluconazole, and amphotericin B should be considered as the drug of choice in the treatment of candidemias due to this yeast. However, often patients are started on empiric fluconazole, and the species is subsequently identified as C. glabrata. If these patients are stable and have cleared the bloodstream on fluconazole, it is reasonable to finish their course of therapy with this drug. Isolates of C. tropicalis and parapsilosis are usually susceptible to fluconazole based on in vitro testing, and clinical response of these isolates to this antifungal has, generally, been very good. Candida lusitaniae is a rare cause of candidemia, and is relatively resistant to amphotericin B.
Azole antifungals are sometimes given as prophylactic medications in patient populations at high-risk for candidemias. When candidemias are diagnosed in patients who are either receiving an azole, or who have recently received one, amphotericin B should be used as empiric therapy. The rationale for this is that many of these "breakthrough" candidemias are caused by azole-resistant isolates such as C. krusei.
The antifungal doses used to treat candidemia mainly depend on whether the patient is clinically stable or is deteriorating with persisting fevers, hypotension and other signs of infection. For stable patients being treated with amphotericin B, we recommend using 0.5 mg/kg/day. Unstable patients should be given higher doses of 0.6-0.8 mg/kg/day. If one of the lipid-associated preparations is to be used, we would recommend a dose of 3-5 mg/kg/day. Stable patients being treated with fluconazole can receive 400 mg a day, assuming normal renal function. If the patient is deteriorating and there are reasons against using amphotericin B, we would give 800 mg of fluconazole a day. Fluconazole should be given intravenously to unstable patients and to patients with impaired gastrointestinal absorption. A reasonable dose of flucytosine is 100 mg/kg/day divided in four doses, and adjusted for renal dysfunction. Clinicians having the ability to measure serum flucytosine levels should aim for peak levels of 50-80 µg/mL.
All candidemic patients should have nonsurgically implanted intravascular catheters removed as part of their therapy. Removing catheters appears to be associated with more rapid clearance of yeasts from the bloodstream. Owing to the difficulties in managing surgically implanted catheters such as those used for hemodialysis or chemotherapy, it may be reasonable to attempt sterilizing the blood with the catheters in place as long as the patient is not clinically deteriorating or does not have persistently positive blood cultures.
An important factor in determining the duration of therapy is the duration of candidemia. Many candidemias can be transient and not associated with significant sequelae, while others are associated with prolonged seeding of the blood. We recommend that all candidemic patients have a repeat set of blood cultures drawn at the time of diagnosis, preferably from a peripheral venous puncture. If the follow-up blood cultures are negative and the patient has a stable clinical course, the patient may require only a short course of therapy such as seven days. If the follow-up cultures are positive, deep foci of infection should be considered such as abscesses, endopthalmitis, urinary tract involvement and septic thrombophlebitis. The blood cultures should be repeated for surveillance and the duration of therapy should be timed after the last positive culture.
Non-neutropenic patients with uncomplicated and transient candidemia (as documented with negative surveillance blood cultures) can be treated with short courses of antifungal agents. One group has used a regimen of 200 mg of amphotericin B given over 5-7 days to treat this select patient population, and a reasonable alternative would be 400 mg of fluconazole given for 7-10 days (assuming the isolate is sensitive). Non-neutropenic patients having persistent fungemia should be treated for at least two weeks after the last positive blood culture, and until all clinical manifestations of infection have resolved. Neutropenic patients diagnosed with candidemia should be given antifungals for the duration of neutropenia, and for at least two weeks after recovery from neutropenia.
Gary M. Cox, MD
John R. Perfect, MD
Department of Medicine
Duke University Medical Center
Durham, North Carolina