Systemic Fungal Infection

C. albicans
Other Candida

Systemic fungal infection is becoming more and more common in modern hospitals. In this article we will concentrate on candidiasis and aspergillosis, and hardly mention other systemic fungal infections such as Histoplasmosis, Blastomycosis, Coccidioidomycosis and so on. (Cryptococcus neoformans deserves a whole page on its own)! Severe systemic fungal infection in hospitals is commonly seen in three major settings:
  1. Neutropaenic patients following chemotherapy, and other oncology patients with immune suppression;
  2. Persons immune compromised due to Acquired Immune Deficiency Syndrome caused by HIV infection;
  3. Patients in intensive care (ICU), who are not necessarily neutropaenic, but are compromised due to the presence of long-term intravascular lines or other breaches in their integument, severe systemic illness or burns, and prolonged broad-spectrum antibiotic therapy.

Other (quoted) predisposing factors are:

Systemic fungal infections cause ~ 25% of infection-related deaths in leukaemics. Infections due to Candida species are the fourth most important cause of nosocomial bloodstream infection. In certain other circumstances, fungal infections are also a major problem. Serious fungal infections may cause 5-10% of deaths in those undergoing lung, pancreas or liver transplantation Acquired fungal sepsis occurs in up to 13% of very low birthweight infants. Small retrospective studies of neonates suggest an association between prolonged third-generation cephalosporin use and Candida infection.

There is a strong suggestion that invasive fungal infections have become more common in recent years, with a nearly 500% increase in the incidence of blood-stream infection with Candida spp. since the 1980s [J Hosp Infect 1995 Jun;30 Suppl:329-38]. Most systemic fungal infections are in fact due to Candida, but Aspergillus infections are also seen. Other causative organisms are less common, although in people with AIDS, a host of different fungi are important causes of disease and death - for example, Cryptococcus neoformans, as well as Histoplasma capsulatum, and strange organisms such as Penicillium marneffei, trichosporonosis, and fusariosis.
One should also not forget that Pneumocystis carinii, so common in AIDS that prophylaxis is not only warranted but mandatory, is now regarded by many as a fungus, although it is not susceptible to many conventional antifungals because it contains cholesterol and not ergosterol in its cell membrane!

Candida albicans

Practical Points
  1. Do not assume that the Candida infection you have encountered is C. albicans
  2. Resistance to azole antifungals is on the increase!
  3. Health care workers commonly transmit yeasts by hand.

C. albicans is an asexual, diploid, dimorphic fungus that is widespread on humans and in their environment. We still don't understand why this common commensal sometimes becomes pathogenic, although impaired host defence mechanisms seem crucial. A variety of fungal virulence factors is being actively explored by researchers, but fungal adherence may be most important. In the future it may even be possible to disrupt this adherence, but at present, we're stuck with antifungal agents.

Much of this page will touch on Candida infection, as this is by far the most important fungal infection in modern hospitals. Candidaemia is said to be the fourth leading cause of bloodstream infection in the USA. [CDC: Am J Infect Control 1999 27 520-32]. Colonization usually precedes candidaemia. Health care workers commonly transmit yeasts between patients by hand, as shown by molecular typing of strains. Candida spp. have been isolated from 15-54% of hands of health care workers in the ICU setting.

There is a large amount of information on the web about C. albicans. Some good sites are listed below .

A note on terminology

Fungal terminology is confusing to the uninitiated (me)! Here are a few terms to whet your appetite:

How do we distinguish C. albicans from other fungi?

This depends on:

  1. Morphology: When the organism is grown on cornmeal agar, chlamydospores are indicative of C. albicans, but take up to 4 days to develop. Pseudohyphae without chlamydospores indicate another Candida sp.; arthrospores suggests Trichosporon. C. glabrata does NOT form pseudohyphae on cornmeal agar.

    C. albicans colonies at 30 o on glucose peptone agar are white or cream, smooth and glistening or occasionally dull and rough, features which are not really helpful in distinguishing it from other spp. True hyphae, pseudohyphae and chlamydospores, with clustered blastospores along the lengths of hyphae, are seen on cornmeal agar.
    C. tropicalis is similar (also with true hyphae, but no chlamydospores. There may also be internodal blastospores. Biochemistry is useful, as C. tropicalis can ferment sucrose. C. dubliniensis is morphologically and biochemically very similar to C. albicans.
    C. krusei doesn't have true hyphae (although the pseudohyphae may look very hypha-like), with ellipsoidal blastospores.
    C. lusitaniae has long pseudohyphae, few branches, prolific small oval blastospores, and characteristically assimilates rhamnose.
    C. parapsilosis has pseudohyphae that often terminate in a swollen cell, with occasional roundish blastoconidia.
    C. glabrata doesn't form pseudohyphae (as noted above), for which reason it was formerly known as Torulopsis. It has small round to oval blastoconidia. (C. famata and C. inconspicuosa are similar, as is Saccharomyces cerevisiae). }

  2. Biochemistry. Sugar fermentation can give a rapid presumptive diagnosis. This is not good enough on its own (at present) - morphology should always be examined too.

  3. "The serum germ tube test" is a quick presumptive test for C. albicans! A light inoculum of cultured yeast is incubated in horse serum for 2-3 hours at 37 o C. The test is positive if there are short hyphae without a constriction where the hypha joins the parent cell . False negatives occur.

  4. Molecular techniques will probably revolutionise the identification of Candida species.

    Sketch showing a variety of different species, some branching, others not; some even yeast-like
    Candida morphology on cornmeal agar
    a = albicans, t=tropicalis, k=krusei, l=lusitaniae, g=glabrata, p=parapsilosis; chl=chlamydospore
    (After Campbell et al.)

    Non-albicans Candidiasis

    Not all fungal infections are due to C. albicans - more than a third are due to other organisms such as C. glabrata, C. krusei, C. parapsilosis, C. tropicalis, C. pseudotropicalis, and occasionally very resistant species like C. lusitaniae. In the European SENTRY study, only 53% of 170 episodes of candidaemia were attributable to C. albicans [Diagn Microbiol Infect Dis 1999 Sep;35(1):19-25]. Prominent organisms are:

    Invasive Aspergillosis

    Practical Points
    1. Aspergillosis is a major cause of infective mortality in oncology patients in some centres
    2. It is often missed.
    3. Do you know that it's not causing deaths in your hospital?

    Substantial outbreaks of this disease may occur in immune compromised patients following on inhalation of Aspergillus spores. See for example [Am J Hematol 2001 Apr;66(4):257-62]. In addition, the disorder may be more common than previously thought in patients with chronic obstructive pulmonary disease [Intensive Care Med 2001 Jan;27(1):59-67]! In a Spanish study of AIDS patients, 1.12% had invasive pulmonary aspergillosis - antiretroviral treatment increased survival substantially [Eur J Clin Microbiol Infect Dis 2000 Sep;19(9):688-93].

    Invasive aspergillosis (IA) is often fatal. One review reported an overall fatality rate of 58%, with rates of 87% in bone marrow transplant patients [Clin Infect Dis 2001 Feb 1;32(3):358-66]. Another review of 595 patients with probable or confirmed IA showed complete responses to amphotericin B in only one quarter of patients, with similar high death rates, especially in the sicker patients who received AB therapy alone, rather than itraconazole +- AB [Medicine (Baltimore) 2000 Jul;79(4):250-60].

    Diagnosis of Systemic Fungal Infection

    Practical Points
    1. If you don't suspect it, you'll miss it!

    The problem with diagnosis and treatment of systemic fungal infection can still often be summed up as "too little too late". Conventional diagnosis of these infections, based on blood cultures or culture of the offending organism from multiple sites, often delays therapy. Patients have frequently been on many different antibiotics for long periods of time in the vain hope that the systemic signs of infection or 'sepsis' will subside. When antifungal therapy is instituted, inappropriate dosing seems quite common.

    There is controversy about when you should treat for apparent systemic fungal infections, especially Candida "infection". Most but perhaps not all authors would treat based on isolation of the organism from the bloodstream. There is not even agreement that intravascular lines should be replaced if fungaemia is detected, although most would also regard this as imperative. It has been said that density of infection (some talk about the "colonisation index"), number of positive cultures, isolation from non-contiguous sites, type of organism (C. tropicalis), and isolation from usually sterile fluids all predict the likelihood of severe systemic fungal infection. [Int J Antimicrob Agents 2000 Jul;15(2):83-90]

    Candiduria is also controversial, although most would agree that it should be treated once (a) confirmed and (b) risk has been stratified appropriately. This is unfortunately seldom done in clinical practice [Mycoses 1999;42(4):285-9].

    Newer tests that have been advocated for early diagnosis of systemic fungal infection include:

    Diagnosis of Invasive Aspergillosis

    Invasive aspergillosis may be difficult to diagnose, especially if one is not aware that it might be present. The presence on high-resolution CT scan of a "halo" or "crescent air" sign is thought to be diagnostic, but in one study the crescent sign was only seen in 10/21 patients [J Comput Assist Tomogr 2001 Mar-Apr;25(2):305-10]; others allege that the halo sign is reliable with early CT, but disappears later on [J Clin Oncol 2001 Jan 1;19(1):253-259]! Screening the blood for galactomannan may be very valuable [Blood 2001 Mar 15;97(6):1604-10] with a sensitivity of up to 90% and specificity of 98%. In one prospective study, 14% of 215 patients on chemotherapy appear to have developed invasive pulmonary aspergillosis [Haematologica 2000 Jul;85(7):745-52]. Polymerase chain reaction may also be of value in early diagnosis [Br J Haematol 2000 Jan;108(1):132-9].

    Antifungal therapy

    Practical Points
    1. Do NOT use amphotericin mixed into intralipid!
    2. If you've neglected the nutritional status of your patient, all the antifungals in the world probably won't help!
    3. If the granulocytopaenia doesn't resolve, then cure of the fungal infection is unlikely!
    4. Fluconazole may antagonise the effect of amphotericin B, at least with Candida infection.

    We will first discuss the drugs used in treatment, and then treatment of specific fungal infections:
    Amphotericin B
    Other drugs
    Rx of Candidiasis
    Rx of Aspergillosis

    Amphotericin B

    The mainstay of antifungal therapy for severe systemic mycoses is still Amphotericin B (often abbreviated to AB). This predominantly fungicidal agent is available in various forms:

    Because of the cost of the lipid formulations, and little evidence of increased efficacy (despite better tolerance), most would reserve these for cases where AB desoxycholate is causing severe toxicity. Some have advocated mixing AB in "Intralipid" but the mixture does not appear to be very stable. I would now avoid this practice.

    Amphotericin B is useful in treatment of infection with Blastomyces, Coccidioides, Histoplasma, Paraoccidioides, Candida, Cryptococcus, but does have substantial associated toxicity. It is a "polyene", and works on fungi by binding to ergosterol in the fungal cell membrane, disrupting the membrane and killing the fungus. Note that AB appears to have a marked post-antifungal effect (PAFE) of up to ten or more hours, although the clinical relevance of this is controversial. . AB may also have immune stimulatory effects! The drug doesn't work reliably for Trichosporon beigelii, Pseudoallescheria boydii, certain Fusarium spp., and Candida lusitaniae.

    Trials of therapy have often been poorly designed, but there is a fair amount of recent evidence that conventional Amphotericin B does work (although you would appear to have to treat 25 patients to save one life), as does liposomal AB. Meta-analyses [Cochrane Database Syst Rev 2000;(4):CD000026] were based on 8 trials showing that IV A/B lowers mortality (RR 0.72, CI 0.51 - 1.02), and three trials of AmBisome. Another analysis [Cochrane Database Syst Rev 2000;(2):CD000026] suggests that routine prophylaxis with antifungals does not influence mortality, a suggestion that few oncologists will happily swallow!

    AB's oral bioavailability is under 5%. Systemic (intravenous) administration is associated with a long list of adverse effects, including marked fever (in about 50%), anaphylaxis (about 1 in 100), nausea, vomiting, and nephrotoxicity. Nephrotoxicity includes lowered GFR, increased loss of potassium and magnesium, and even distal renal tubular acidosis. Nephrotoxicity is said to be irreversible if therapy is prolonged [Antimicrob Agent Chemother 1978 13 271-6]. Other nephrotoxic agents may worsen toxicity. It is alleged that nephrotoxicity may be diminished by sodium-loading. Rarely, severe hypertension has been reported in association with AB therapy! The desoxycholate form has a long half life (24-48 hours), and a volume of distribution of about 4 L/kg.

    Because it precipitates in normal saline, AB must be mixed in 5% dextrose water, and given via a central line to prevent phlebitis. It is not necessary to adjust dosage in liver failure, but many lower the dose in renal failure(?). The drug is not dialysable. CSF penetration is poor (2-4% of serum levels), but meningeal penetration may be better, and the drug enters the CSF of infants extremely well. We often tend to underdose our patients with (say) 0.5mg/kg - one should probably aim for closer to 1mg/kg or more, especially with more resistant fungi such as Aspergillus. Such doses are achieved with fewer side-effects with newer formulations such as liposomal amphotericin, but the cost is immense. Far higher doses have been used with lipid formulations of AB (even 7mg/kg).

    The old tradition of giving a test dose of AB has been questioned, and some would ignore this approach. (Dose: 1mg in 100ml 5% D/W over 30min). Conventionally the daily dose is given over about 6 hours, but some argue that it should be given over a longer period.

    Fluconazole and other Azoles

    Practical Points
    1. Beware of drug interactions with azoles, especially with, for example, cyclosporin.

    Azoles (such as fluconazole) inhibit fungal growth by preventing formation of ergosterol, vital for fungal cell membrane integrity. They do this by inhibiting fungal cytochrome P-450 (14-a sterol demethylase), so it is not surprising that these agents have powerful inhibitory effects on some human CYPs. (See our CYP page ). Fluconazole is usually regarded as a "first generation azole", and may well be supplanted in the next few years by newer azoles such as vorizonazole, which have a wider spectrum of activity. Azoles generally have "non-concentration dependent activity" against most fungi.

    Fluconazole is mainly used for C. albicans infection (and some other susceptible Candida spp. but NOT C. krusei, and has inconstant activity against C. glabrata). C. albicans may acquire resistance, especially with chronic or recurrent treatment in AIDs patients. Fluconazole may be effective against Cryptococcus neoformans meningitis, and Coccidioidomycosis.

    Fluconazole has good GIT absorption, and is renally excreted. CVVHD causes substantial removal, even more than normal kidneys, and far greater than with CVVH or intermittent dialysis [Mycoses 1999 Apr;42(1-2):17-9]. There is a substantial potential for drug interactions owing to inhibition of CYP2C9 , and also 3A4. Volume of distribution may be substantially increased in critically ill patients. There is good cerebrospinal fluid penetration. Adult doses used for severe infections are often in the range of 200mg to 400mg/day, although 800mg/day(+) has sometimes been used. Half life (at 200mg/day) is about 27-37 hours, with a V D of 0.7 to 1 L/kg.
    { Susceptibility testing may be of value in guiding dosing, with the AUC:MIC ratio perhaps a good predictor of response. Fluconazole has no post-antifungal effect, unlike AB. Both human serum and GM-CSF may enhance killing of fungi by fluconazole.

    For a review of fluconazole and itraconazole in C. albicans infection see [ J Antimicrob Chemother 2000 Apr;45(4):555]. In neonates, fluconazole has been given at 5 to 6 mg/kg, but long dosing intervals are required in very low birthweight infants [Eur J Med Res 2000 May 23;5(5):203-8].

    Details of azoles other than fluconazole are provided below:

    Other agents


    Voriconazole is an azole antifungal. Its plasma protein binding is about 70%, it is mainly cleared by the liver (? CYP 3A4), and has good, rapid oral bioavailability. There is substantial first-pass metabolism, which is saturable. Good for Candida spp, Aspergillus. This includes fluconazole resistant organisms such as C. glabrata. Voriconazole is fungicidal for aspergillus at twice MIC. It is well absorbed orally (about 90%). Voriconazole is also effective against:

    It is NOT useful against Sporothrix schenckii and zygomycetes.

    Dosage should probably be ~200 to 400 mg/day. In one study, the loading dose was 6mg/kg BD for 24 hours, followed by 3mg/kg BD. Side effects include (reversible) mild-to-moderate visual disturbance in about 14% of people (enhanced brightness or blurred vision), and raised bilirubin/alkaline phosphatase in some. Rash occurs in about 4%, and photosensitivity and erythema have been reported.

    Treatment of Systemic Candida Infection

    The mainstay of therapy is still amphotericin B, but this may change with the advent of newer azoles such as voriconazole. C. lusitaniae, which is often resistant to AB, is fortunately still fairly uncommon. The major problem is the toxicity of AB, which is substantial. The expense of lipid formulations of AB practically precludes their use.

    Some have advocated the use of fluconazole, often in high doses, for putative or confirmed Candida infection. The problem with this is that not only are C. krusei and C. glabrata often resistant to fluconazole, but with the advent of repeated or long-term fluconazole therapy/prophylaxis (in, for example, AIDS patients), even C. albicans may be resistant. Dosing of fluconazole is highly controversial.

    Treatment of Invasive Aspergillosis

    Amphotericin B is probably still the drug of choice for invasive aspergillosis, although voriconazole may be very effective, and posaconazole has shown promise in experimental models of the disease, as has ravuconazole. Some would even recommend surgical resection for selected cases of invasive pulmonary aspergillosis. A review of 2121 published cases suggests that amphotericin B combined with either rifampicin or flucytosine may be better than amphotericin B monotherapy [Denning DW & Stevens DA: Rev Infect Dis 1990 12(6) 1147-201]. The 'evidence' is not cast in stone.

    Treatment of Aspergillus infection in the immune-compromised is urgent , and should be based on clinical or radiological criteria, without waiting for microbiological confirmation.

    A fascinating article by Clemons et al. shows that in IL-10 deficient knockout mice, survival after infection with A. fumigatus is better than with wild-type mice that have IL-10. IL-10 promotes a 'Th-2' phenotype, as is seen in asthmatics! [Clin Exp Immunol 2000 Nov;122(2):186-91]

    Sequential Therapy

    This has become quite popular - initial aggressive amphotericin B therapy may be followed by e.g. long-term fluconazole. This approach is best-documented in cryptococcal meningitis in AIDS patients, but a similar approach (AB followed by itraconazole) has also been used with Aspergillus infection.

    Treatment Failures

    These are still very common with all systemic mycoses. Often they are thought to reflect the severity of the underlying disease, but there is some concern that more often than not, we underdose our patients with antifungal therapy, use the wrong agents, or start too late. Secondary resistance may also arise. Treatment failure occurs in about twenty to thirty percent of patients with candidaemia, and the death rate in patients is high. In one study of 415 cases of candidaemia, overall mortality was 45%, [CMAJ 1999 Feb 23;160(4):493-9], being lower in children. In invasive aspergillosis, treatment failures are still more common (about 50%).

    Secondary resistance is usually seen in more chronic situations, especially in AIDS patients where there is recurrent or prolonged therapy with fluconazole.

    Assessment of resistance to antifungal therapy is difficult. The National Committee for Clinical Laboratory Standards (NCCLS) has defined guidelines for testing of azole antifungals, but standards for amphotericin B are not clear. (For details, see Dodds et al.).

    For fungal resistance see also [New Horiz 1996 Aug;4(3):338-44]. Primary resistance to fluconazole is common in C. glabrata, and often emerges (or may be primarily present) in C. krusei. C. lusitaniae is primarily resistant to amphotericin, but fortunately, such infections are uncommon.