Dose-response relationships of three amphotericin B formulations in a non-neutropenic murine model of invasive aspergillosis

New lipid-associated formulations of amphotericin B (AmB) have been developed in order to reduce toxicity and enhance the efficacy of AmB by allowing administration of higher doses of the drug. We determined the in vivo dose-response relationships of 1 day and 7 day treatment of AmB, Ambisome (AmBi) and Abelcet (ABLC) in a non-neutropenic murine model of invasive aspergillosis by using survival as an endpoint. Female CD-1 mice were infected intravenously 48 h prior to start therapy with Aspergillus fumigatus (1 x 10(7) conidia/mouse). Groups of 10 mice were treated iv for 1 day or 7 days with increasing 2-fold doses of AmB, ABLC and AmBi up to a maximum of 20 mg/kg/day. Mortality was determined twice daily until day 15. Results were analyzed using product-moment survival analysis and by determining the dose response relationships on day 15. Survival at day 15 of mice with 7 day AmBi or ABLC treatment was significantly better than that of controls or AmB. The ED50s of AmBi and ABLC were 0.06 (95% CI: 0.03-0.127) mg/kg and 0.21 (0.06-0.66) mg/kg respectively. In addition, the maximum effect was higher for AmBi than ABLC, 90% survival versus 68%, respectively. Most of the effects of treatment with AmBi were reached after 1 day of treatment, indicating that the first dose given is most important in predicting survival. This study shows that AmBi and ABLC were significantly more efficacious than AmB in a non-neutropenic murine model of invasive aspergillosis, and that the effect observed was primarily dependent on the first dose administered.

Interlaboratory reproducibility of a microsatellite-based typing assay for Aspergillus fumigatus through the use of allelic ladders: proof of concept

An interlaboratory study was performed with the aim of investigating the reproducibility of a multiplex microbial microsatellite-based typing assay for Aspergillus fumigatus in different settings using a variety of experimental and analytical conditions and with teams having variable prior microsatellite typing experience. In order to circumvent problems with exchange of sizing data, allelic ladders are introduced as a straightforward and universally applicable concept for standardization of such typing assays. Allelic ladders consist of mixtures of well-characterized reference fragments to act as reference points for the position in an electrophoretic trace of fragments with established repeat numbers. Five laboratories independently analysed six microsatellite markers in 18 samples that were provided either as DNA or as A. fumigatus conidia. Allelic data were reported as repeat numbers and as sizes in nucleotides. Without the use of allelic ladders, size differences of up to 6.7 nucleotides were observed, resulting in interpretation errors of up to two repeat units. Difficulties in interpretation were related to non-specific amplification products (which were resolved with explanation) and bleed-through of the different fluorescent labels. In contrast, after resolution of technical or interpretive problems, standardization of sizing data by using allelic ladders enabled all participants to produce identical typing data. The use of allelic ladders as a routine part of molecular typing using microsatellite markers provides robust results suitable for interlaboratory comparisons and for deposition in a global typing database.

Disseminated Rhizopus microsporus infection in a patient on oral corticosteroid treatment: a case report

A 71-year-old male with mild steroid-induced hyperglycaemia was diagnosed with a lethal invasive Rhizopus microsporus infection. Disseminated zygomycosis is a rare entity and is most frequently found in neutropenic patients with haematological malignancies, post-transplants or in patients on deferoxamine therapy. Infection is characterised by tissue infarction and necrosis due to angioinvasive hyphae. Culture of Zygomycetes is necessary for species determination but histology is a must to prove the infection. Ante-mortem diagnosis and culture is challenging and therefore mortality approaches 100%. Apart from amphotericine B, most anti-fungals have no activity against Zygomycetes but posaconazole might offer new possibilities as a first-line agent. Timely diagnosis, rapid surgery of infected tissue, correction of underlying disorders and correct anti-fungal therapy might be life-saving. Due to the increasing use of potent immunosuppression, stem cell and organ transplants and possibly selection for Zygomycetes by prior treatment with broad-spectrum antifungal therapy, the incidence of zygomycosis is rising. Therefore, clinicians might encounter an increasing number of zygomycosis cases in the near future.

Molecular typing of Aspergillus species

Aspergillus species are widely distributed fungi that release large amounts of airborne conidia, which are dispersed in the environment. Several Aspergillus species have been described as human pathogens. Molecular techniques have been developed to investigate the epidemiological relation between environmental and clinical isolates. Several typing methods have been described for Aspergillus species, most of them with reference to Aspergillus fumigatus. Here, we summarise all the different available molecular typing techniques for Aspergillus. The performance of these techniques is evaluated with respect to their practical feasibility, and their interpretation and discriminatory power assessed. For A. fumigatus isolates, a large extent of genetic variability is demonstrated and therefore fingerprinting techniques with high discriminatory power and high reproducibility are required for this species. Afut1-restriction fragment length polymorphism and microsatellite typing showed the highest discriminatory power. In addition, the microsatellites show excellent reproducibility. Other typing techniques are still useful for smaller epidemiological problems and for less well-equipped laboratories.

[An outbreak of psittacosis at a bird-fanciers fair in the Netherlands]

Within 2 weeks after a bird-fanciers fair in the Netherlands in November 2007 11 patients presented at our hospital with fever, shivers and severe headache. Dyspnea and dry cough were less common, although the chest X-ray showed a consolidation in 9 out of 11 patients. The clinical diagnosis of psittacosis was quickly confirmed using real-time PCR, although the sensitivity of this test was low (20%). In 9 patients the diagnosis was later confirmed by a rise in complement fixing antibodies in paired sera. None of the patients needed intensive care treatment. All patients recovered uneventfully with antibiotic treatment. Psittacosis is an avian zoonosis, caused by Chlamydophila psittaci. Humans are infected by inhalation of the bacterium that is shedded by excreta or dust from feathers of different sites of either sick or asymptomatic, mostly psittacine, birds. The clinical picture ranges from asymptomatic or mild, flue-like symptoms to severe illness. A timely diagnosis is necessary for successful outbreak management. The realtime PCR is an adequate test in that respect.

Efficacy and pharmacodynamics of flucytosine monotherapy in a nonneutropenic murine model of invasive aspergillosis

The therapeutic efficacy of flucytosine (5FC) monotherapy and the pharmacodynamic index predictive of efficacy were evaluated in a nonneutropenic mouse model of acute invasive aspergillosis. Mice were infected intravenously with an Aspergillus fumigatus isolate (the median MICs of 5FC were 128 mug/ml under the standard condition, 0.5 microg/ml at pH 6.0, and 0.031 microg/ml at pH 5.0) 2 h prior to the start of therapy and were treated for 7 days with different 5FC dosing regimens. The total doses ranged from 50 to 800 mg/kg of body weight/day and were administered at 6-, 12-, and 24-h intervals. The efficacy was assessed by means of survival. The survival rates of the treatment groups ranged from 40 to 90%, while the survival rate of the control group was 20%. The efficacy found depended primarily on the total daily dose. However, the power of our sample size may have been too low to exclude an effect of dose fractionation. The pharmacodynamic index that most strongly correlated with the efficacy was the area under the serum concentration-time curve and MIC ratio (R(2) = 0.86). We conclude that 5FC monotherapy is efficacious in a murine Aspergillus fumigatus infection model.

A case of imported paracoccidioidomycosis: an awkward infection in the Netherlands

Paracoccidioidomycosis is an important endemic mycosis in South America. In Europe the disease is very rare and only found as infections in travelers to Latin America. We report here the first case encountered in the Netherlands for which the appropriate diagnosis was not attained for several months. A Dutch 60-year-old man presented with a painful ulceration in the buccal mandibular vestibular mucosa of three months duration. While his medical history was uneventful, he had worked, until 8 years prior to his presentation, as a carpenter for 25 years in the jungles of Peru and Ecuador. An aberrant chest radiograph, CT-scan of the lungs and increased erythrocyte sedimentation rate were suggestive of sarcoidosis or a bronchiolitis obliterans organizing pneumonia. There was no improvement in the patient's symptoms despite the use of budesonide and prednisone medication, as well as tuberculosis prophylaxis with isoniazide and rifampicin, and local use of miconazole. Quite to the contrary, as an irritated, irregular hyperemic mucosa and gingiva with ulceration were noticed during this period of time. These precipitated an incisional biopsy through which a mixed inflammatory cellular infiltrate and large yeast cells were found on histopathologic examination. Based on the patient's travel history and the multiple budding yeastlike cells revealed in the biopsy tissue, the diagnosis of paracoccidioidomycosis was finally made. This was supported by the isolation of Paracoccidioides brasiliensis in culture. Antimycotic oral therapy with itraconazole was started and continued for 15 months. At two and five year follow-ups, the patient was asymptomatic. In Europe, it may be expected that diseases that are endemic in other areas will be seen more frequently in countries where the diseases are not routinely encountered. It is most likely that the use of corticosteroid medication, with its inherent immunosuppressive effect, resulted in the reactivation of an infection acquired many years before in Latin America. The etiologic agent then disseminated from the initial focal point to cause the ensuing oral mucous membrane lesions. The importance of the patient's prolonged residence in Latin America was overlooked. The very long latency of endemic mycoses emphasizes the need for a meticulous history which should include not only recent trips, but also past residence in foreign countries.

Relationship between in vitro activities of amphotericin B and flucytosine and pH for clinical yeast and mold isolates

In this study, we investigated the pH dependency of the in vitro activities of amphotericin B (AMB) and flucytosine (5FC) against Candida spp., Cryptococcus neoformans, Aspergillus fumigatus, Rhizopus spp., and Scedosporium prolificans in RPMI 1640 buffered with citrate buffer (pH 4.0, 5.0, 5.4, and 6.0), citrate-phosphate buffer (pH 5.4, 6.0, 6.4, and 7.0), and 3-[N-morpholino]propanesulfonic acid (MOPS) (pH 6.4, 7.0, 7.4, and 7.9). For 5FC, no significant differences were found between MICs obtained with the different buffers, while for AMB, significant differences were found. The MICs obtained with citrate-phosphate buffer were approximately 1 twofold-dilution step higher than the MICs obtained with MOPS. We demonstrated that the in vitro activities of AMB and 5FC against yeast and mold isolates were pH dependent. The in vitro activity of AMB decreased when the pH was lowered, while the in vitro activity of 5FC increased. The effect of the pH on the in vitro activities was dependent not only on the antifungal agent tested but also on the microorganism. For AMB, there was a nonlinear relationship (median r(2), 0.864) for Candida spp., C. neoformans, A. fumigatus, and Rhizopus spp. over the pH range tested. The mean MICs ranged from 0.5 to 2.52 microg/ml at pH 7.0 and from 20.16 to 32 microg/ml at pH 5.0. For S. prolificans, there was no relationship. For 5FC, there was a linear relationship for Candida spp. (median r(2), 0.767) and a nonlinear relationship for C. neoformans and A. fumigatus (median r(2), 0.882) over the pH range tested. The mean MIC values ranged from 0.125 to 1,024 microg/ml at pH 7.0 and from 0.02 to 4 microg/ml at pH 5.0. For Rhizopus spp. and S. prolificans, the relationship could not be determined, since the MIC was >1,024 microg/ml over a pH range of 4.0 to 7.9.

Effect of pH on the in vitro activities of amphotericin B, itraconazole, and flucytosine against Aspergillus isolates

The in vitro susceptibilities of 21 Aspergillus isolates were tested against three antifungal agents in RPMI 1640 and yeast nitrogen base at pH 5.0 and 7.0 by a broth microdilution format of the NCCLS method. The MICs of amphotericin B and itraconazole were higher, while those of flucytosine were lower, at pH 5.0 than at pH 7.0. The poor correlation between in vitro results and clinical outcome could be due to a difference in pH between the in vitro susceptibility test and at the site of infection.

In vitro interactions between amphotericin B, itraconazole, and flucytosine against 21 clinical Aspergillus isolates determined by two drug interaction models

Combination therapy of flucytosine (5FC) with other antifungal agents could be of use for the treatment of invasive aspergillosis. However, interpretation of the results of in vitro interactions is problematic. The fractional inhibitory concentration (FIC) index is the most commonly used method, but it has several major drawbacks in characterizing antifungal drug interaction. Alternatively, a response surface approach using the concentration-effect relationship over the whole concentration range instead of just the MIC can be used. We determined the in vitro interactions between amphotericin B (AMB), itraconazole, and 5FC against 21 Aspergillus isolates with a broth microdilution checkerboard method that employs the dye MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide]. FIC indices based on three different MIC endpoints (MIC-0, MIC-1, and MIC-2) and the interaction coefficient alpha were determined, the latter by estimation from the response surface approach described by Greco et al. (W. R. Greco, G. Bravo, and J. C. Parsons, Pharmacol. Rev. 47:331-385, 1995). The value obtained for the FIC index was found to be dependent on the MIC endpoint used and could be either synergistic, indifferent, or antagonistic. The response surface approach gave more consistent results. Of the three combinations tested, the AMB-5FC combination was the most potent in vitro against Aspergillus spp. We conclude that the use of the response surface approach for the interpretation of in vitro interaction studies of antifungals may be helpful in order to predict the nature and intensity of the drug interaction. However, the correlation of these results with clinical outcome remains difficult and needs to be further investigated.

An adult case of oral infection with Kingella kingae

An exceptional case of microbiologically confirmed oral infection with Kingella kingae in an immunocompetent adult (30-year-old woman) is presented and the pathogenesis is discussed and related to known literature data.K. kingae is a rather common but yet relatively unknown commensal corroding bacterium from the oro- and nasopharynx in healthy children, which might turn into a human pathogen causing osteomyelitis, arthritis, spondylitis, endocarditis and intervertebral diskitis in young children and rarely endocarditis, septic arthritis, meningitis, epiglottitis, diskitis and bacteraemia in adults. Sofar K. kingae associated stomatitis was reported in children and a few adults, however, with concomitant herpes simplex virus infections, and without microbiological confirmation. In the described case no viral infection was found. The proven K. kingae stomatitis represents an extension of the pathogenic spectrum and suggests that the breach of the oral mucosal barrier can be caused by the bacterial pathogen itself. Whether a concomitant viral infection is necessary forK. kingae to actually invade the bloodstream remains to be considered.

Evaluation of a polymerase chain reaction reverse hybridization line probe assay for the detection and identification of medically important fungi in bronchoalveolar lavage fluids

An assay system in which polymerase chain reaction (PCR) amplification of the ITS-1 region of ribosomal DNA (rDNA) is combined with a reverse-hybridization line probe assay (LiPA) was used for the identification of six Candida species and four Aspergillus species in pure cultures of clinical isolates, as well as in bronchoalveolar lavage (BAL) fluid samples from 42 patients with various underlying diseases. The results were compared with the results obtained with conventional routine identification methods as well as with a commercial enzyme-linked immunosorbent assay (ELISA) galactomannan detection assay and an Aspergillus-specific PCR. No discrepancies between the PCR-LiPA system and routine methods were found for pure cultures of Candida and Aspergillus species except in the case of Aspergillus versicolor. In BAL fluid samples in which Candida species were cultured, the PCR-LiPA system identified more species than did the routine methods. When routine analyses of patient samples were supplemented by adding data obtained by repurifying and re-identifying cultures and by taking isolates obtained from other body sites into account, the results agreed with PCR-LiPA system results in 81% of the cases (34/42). Most of the remaining discrepancies (6/8) involved cases in which such supplementary data were not available. In BAL fluid samples from which A. fumigatus was cultured, the agreement between the PCR-LiPA system and the routine methods was low. Only 2 of 11 BAL samples shown to contain A. fumigatus in ELISA and genus-specific PCR assays were positive in PCR-LiPA system. The PCR-LiPA system enables the simultaneous detection and identification of different fungal species present in pure or mixed populations within 6 h in a single assay. Optimization is required, however, before it is useful as a diagnostic tool in the clinical microbiology laboratory.

[Primary meningococcal-monoarthritis (Neisseria meningitidis serotype C) in a child]

A 26-month-old girl had a painful, swollen right knee, accompanied by fever and vomiting. She had had an upper respiratory tract infection for a number of days. Following a positive culture of synovial fluid, monoarthritis caused by Neisseria meningitidis serotype C without other signs of illness was diagnosed. She recovered completely after one joint aspiration and intravenous and oral antibiotic therapy. N. meningitidis serotype C infections without meningitis or septicaemia are rare, but should form part of the differential diagnosis of septic monoarthritis.

Duration of antifungal treatment and development of delayed complications in patients with candidaemia

The optimal duration of antifungal treatment of candidaemia is unknown. Prolonged treatment has been advocated to prevent late-onset complications caused by metastatic infectious foci. To evaluate the relationship between duration of antifungal therapy and the development of delayed complications in patients with candidaemia, a retrospective descriptive study was performed in a tertiary care centre. The study included 180 consecutive patients with candidaemia identified by at least one positive blood culture. Development of late-onset complications of candidaemia during long-term follow-up was monitored. Of the patients treated with antifungal agents, 55% completed treatment without apparent complications, 33% died during therapy, 12% developed disseminated disease and one patient continues to receive ongoing treatment. Of the 81 patients who completed treatment as scheduled, 25% received treatment for less than 2 weeks, 31% for 2-4 weeks and 38% for more than 4 weeks; duration of therapy was unknown in 6%. Only three (2%) patients developed delayed complications; in these patients, the duration of antifungal treatment had been 3 weeks, 5 weeks and 22 weeks, respectively. In conclusion, despite theoretical concerns, there seems to be no correlation between the duration of antifungal treatment and the development of late complications in patients with candidaemia. This suggests that treatment of 2 weeks or less may be sufficient, provided the initial response to therapy is favourable.

Risk factors in surgical patients with intra-abdominal Candida infection in intensive care

Background

Candida species have been isolated with increasing frequency during intra-abdominal infection, yet their role as pathogens or co-pathogens and as determinants for outcome, especially in patients who are frequently reoperated on, is not well understood. Furthermore, the value of antimycotic therapy is the subject of discussion.

Methods

Candida cultures from the abdomen of surgical patients over the past 10 years (1989–1999) were documented. Medical records of the patients were reviewed for risk factors for morbidity and death.

Results

Some 72 patients (45 men, 27 women), of mean age 60 (range 18–88) years, had Candida cultured from the abdomen during operation. Indications for surgery included duodenal ulcer perforation in 15 patients (21 per cent), gastrointestinal malignancy in 15 (21 per cent), pancreatitis in nine (12 per cent) and various disorders in the remaining 33. A total of 361 relaparotomies was performed in 72 patients (mean 5, range 0–48). Cultures were positive at the first laparotomy in 19 patients, at the first relaparotomy in 20 patients and in 33 patients at the second or more relaparotomy. All but one patient were treated in the intensive care unit. Some 45 (62 per cent) of 72 patients died a mean of 44 (range 3–234) days after the first laparotomy. Risk factors for death, which appeared not to be significant, are shown in the Table.

Conclusion

The number of relaparotomies among patients with Candida-positive cultures was high. No significant risk factors for death could be identified in surgical patients with intra-abdominal Candida infection in intensive care.

Role of cytokines of the tumour necrosis factor family in the immune response to disseminated Candida albicans infection

Background

Tumour necrosis factor (TNF) α, lymphotoxin (LT) α, CD40L and FasL are members of the TNF family that play a crucial role in modulation of the immune response. Their role in the defence against infection with Candida albicans was investigated in mice deficient in either TNF-α and LT-α (TNF–/– LT–/– mice), CD40L (CD40L–/– mice) or Fas (MRL/lpr mice).

Methods

Mortality rates were compared in mice infected intravenously with 106 colony-forming units of C. albicans per mouse.

Results

After infection with C. albicans the TNF–/– LT–/– mice had a significantly increased mortality rate compared with control mice (100 versus 40 per cent; P < 0·01). This was due to a 10–1000-fold increased outgrowth of the yeasts in the kidneys and liver of TNF–/– LT–/– mice (P < 0·01). Defective recruitment and phagocytosis, but not Candida killing, were responsible for these effects. CD40L–/– mice were also more susceptible to systemic candidiasis than the wild-type controls (mortality rate 80 versus 50 per cent; P < 0·05), and the growth of Candida in the kidneys was one order of magnitude higher in the deficient than in control mice (P < 0·05). Neutrophil function in the CD40L–/– mice was normal, whereas decreased Candida killing by macrophages through nitric oxide-dependent mechanisms was responsible for the effect of CD40/CD40L interactions. In contrast, Fas-defective MRL/lpr mice were significantly more resistant to disseminated candidiasis (mortality rate 50 versus 100 per cent; P < 0·01); this was mediated by the facilitation of neutrophil migration to the site of infection.

Conclusion

Cytokines of the TNF family play a crucial role in the modulation of host defence against fatal C. albicans infection. Their effects are exerted selectively at the level of neutrophil or macrophage function.

In vitro interaction of flucytosine combined with amphotericin B or fluconazole against thirty-five yeast isolates determined by both the fractional inhibitory concentration index and the response surface approach

Combination therapy could be of benefit for the treatment of invasive yeast infections. However, in vitro interaction studies are relatively scarce and the interpretation of the fractional inhibitory concentration (FIC) index can be contradictory due to various definitions used; not all information on the interaction study is used in the index, and different MIC end points exist for different classes of drugs. Fitting an interaction model to the whole response surface and estimation of an interaction coefficient alpha (IC(alpha)) would overcome these objections and has the additional advantage that confidence intervals of the interaction are obtained. The efficacy of flucytosine (5FC) in combination with amphotericin B (AB) and fluconazole (FCZ) was studied against 35 yeast isolates in triplicate (Candida albicans [n = 9], Candida glabrata [n = 9], Candida krusei [n = 9], and Cryptococcus neoformans [n = 8]) using a broth microdilution checkerboard method and measuring growth after 48 h by a spectrophotometer. The FIC index and IC(alpha) were determined, the latter by estimation from the response surface approach described by Greco et al. (W. R. Greco, G. Bravo, and J. C. Parsons, Pharmacol. Rev. 47:331-385, 1995) by using a computer program developed for that purpose. For the 5FC-FCZ combination, the interactions determined by the IC(alpha) generally were in concordance with the interactions determined by the FIC index, but large discrepancies were found between both methods for the 5FC-AB combination. These could mainly be explained by shortcomings in the FIC approach. The in vitro interaction of 5FC-AB demonstrated variable results depending on the tested Candida isolate. In general, the 5FC-FCZ combination was antagonistic against Candida species, but for some Candida isolates synergism was found. For C. neoformans the interaction for both combinations was highly dependent on the tested isolate and the method used. Response surface approach is an alternative method for determining the interaction between antifungal agents. By using this approach, some of the problems encountered with the FIC were overcome.

Aspergillus galactomannan antigen levels in allogeneic haematopoietic stem cell transplant recipients given total parenteral nutrition

False-positive tests for Aspergillus galactomannan have been reported in neutropenic patients. We failed to detect any circulating antigen during the 2 weeks following allogeneic haematopoietic stem cell transplantation of 12 patients who had severe mucositis but were unable to eat.

Comparison of fractional inhibitory concentration index with response surface modeling for characterization of in vitro interaction of antifungals against itraconazole-susceptible and -resistant Aspergillus fumigatus isolates

Although the fractional inhibitory concentration (FIC) index is most frequently used to define or to describe drug interactions, it has some important disadvantages when used for drugs against filamentous fungi. This includes observer bias in the determination of the MIC and no agreement on the endpoints (MIC-0, MIC-1, or MIC-2 [> or = 95, > or = 75, and > or = 50% growth inhibition, respectively]) when studying drug combinations. Furthermore, statistical analysis and comparisons are troublesome. The use of a spectrophotometric method to determine the effect of drug combinations yields quantitative data and permits the use of model fits to the whole response surface. We applied the response surface model described by Greco et al. (W. R. Greco, G. Bravo, and J. C. Parsons, Pharmacol. Rev. 47:331-385, 1995) to determine the interaction coefficient alpha (ICalpha) using a program developed for that purpose and compared the results with FIC indices. The susceptibilities of amphotericin B (AM), itraconazole (IT), and terbinafine (TB) were tested either alone or in combination against 10 IT-susceptible (IT-S) and 5 IT-resistant (IT-R) clinical strains of Aspergillus fumigatus using a modified checkerboard microdilution method that employs the dye MTT [3-(4,5-dimethyl-2-thiazyl)2,5-diphenyl-2H-tetrazolium bromide]. Growth in each well was determined by a spectrophotometer. FIC indices were determined and ICalpha values were estimated for each organism strain combination, and the latter included error estimates. Depending on the MIC endpoint used, the FIC index ranged from 1.016 to 2.077 for AM-IT, from 0.544 to 1.767 for AM-TB, and from 0.656 to 0.740 for IT-TB for the IT-S strains. For the IT-R strains the FIC index ranged from 0.308 to 1.767 for AM-IT, from 0.512 to 1.646 for AM-TB, and from 0.403 to 0.497 for IT-TB. The results indicate that the degree of interaction is not only determined by the agents themselves but also by the choice of the endpoint. Estimates of the ICalpha values showed more consistent results. Although the absolute FIC indices were difficult to interpret, there was a good correlation with the results obtained using the ICalpha values. The combination of AM with either IT or TB was antagonistic in vitro, whereas the combination of IT and TB was synergistic in vitro for both IT-S and IT-R strains. The use of response surface modeling to determine the interaction of drugs against filamentous fungi is promising, and more consistent results are obtained by this method than by using FIC indices.