Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, May 2008, p. 1743-1750, Vol. 52, No. 5
0066-4804/08/$08.00+0 doi:10.1128/AAC.01388-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

The Old Court, Broadstairs, Kent, United Kingdom,1 Hospital de Cruces, Baracaldo Vizcaya, Spain,2 Pfizer Inc., New York, New York,3 Women's and Children's Hospital, Adelaide, Australia,4 Royal Perth Hospital, Perth, Australia,5 University of Tartu, Tartu, Estonia,6 Hematology Hospital Clinic, Barcelona, Spain,7 St. Vincent's Hospital, Sydney, Australia,8 Peter MacCallum Cancer Centre, Melbourne, Australia,9 Westmead Hospital, Sydney, Australia,10
Received 26 October 2007/ Returned for modification 29 December 2007/ Accepted 10 January 2008
|
|
|---|
28 days of therapy. Patients receiving primary therapy showed a 61% response versus 56% for the others. The best therapeutic responses were seen for skin/subcutaneous (91%) or bone (79%) infections, and the lowest for CNS infections (43%). Patients without major immune suppression (72%) or those with solid organ transplantation (63%) or various hematological conditions (60%) showed the best responses by underlying condition. Median known survival time was 133 days (therapy successes, 252 days; failures, 21 days). In all, 43 (40%) patients died, 73% due to scedosporiosis. Patients with Scedosporium prolificans infection had significantly reduced survival times (P = 0.0259) and were more likely to die from fungal infection (P = 0.002) than were Scedosporium apiospermum-infected patients. In a subset of 43 patients where voriconazole baseline MICs were available, response to voriconazole was higher for S. apiospermum-infected patients (54% response; MIC50, 0.25 µg/ml) than for S. prolificans-infected patients (40% response; MIC50, 4.0 µg/ml). Voriconazole demonstrated clinically useful activity in the treatment of both S. apiospermum and S. prolificans infections and was well tolerated. |
|
|---|
The broad-spectrum azole voriconazole exhibits activity in vitro against both Scedosporium species (4, 30, 45). Voriconazole may show fungicidal activity against S. apiospermum and is efficacious in both mouse and guinea pig models of scedosporiosis (4, 5, 10). In humans, numerous case reports have detailed its clinical effectiveness against Scedosporium infections, including those involving the central nervous system (CNS) and bone (6, 12, 13, 14, 20, 22, 23, 34, 35, 39, 40, 47). Most describe the use of voriconazole as the sole antifungal agent, sometimes combined with surgery. Successful control of infection with voriconazole in combination with terbinafine or an echinocandin has been described, particularly for S. prolificans, where voriconazole is less potent in vitro (3, 15, 19, 25, 32, 46, 50); however, no large series of cases have been published to date and the long-term efficacy of voriconazole in treating scedosporiosis has not been clearly established. One small (n = 10 patients) study of voriconazole use as salvage therapy noted a 30% response rate (38). Thus, treatment remains largely driven by anecdotal experience. This report summarizes the efficacy of voriconazole in the treatment of 107 Scedosporium infections in patients registered in the primary, salvage, and compassionate/named patient programs from the Pfizer global voriconazole clinical trials database and postmarketing cases from Australia. The in vitro susceptibilities of isolates from 43 of these patients to voriconazole and standard antifungal agents were also determined.
|
|
|---|
Definitions. Successful therapy was defined as a complete or partial clinical response at the end of treatment/last visit. All other responses were considered therapy failures (18). Stable disease was defined as "no significant change in clinical, imaging or microbiological results" (18). Known survival time was defined as the last date on which a patient was reported to be alive. In total, 117 cases were identified; however, only 107 cases that satisfied the criteria for "probable" or "definite" infection according to published European Organization for Research and Treatment of Cancer/Mycoses Study Group guidelines (1) were evaluated.
Microbiology. Scedosporium isolates were identified to species level (as S. prolificans or S. apiospermum) by standard colonial and mycological methods (9). In vitro susceptibility data were obtained for baseline isolates from 43 patients. MICs of amphotericin B, itraconazole, and voriconazole were determined using a broth microdilution method, according to the Clinical and Laboratory Standards Institute (formerly NCCLS) M38-A guidelines using a MIC endpoint of complete growth inhibition (33). Confirmation of species identification and susceptibility testing were carried out at the two Voriconazole Mycology Reference Laboratories (11) (Pfizer study isolates) or at the Mycology Unit of the Women's and Children's Hospital, Adelaide, South Australia (Australian isolates).
Statistical analysis. The overall global response to treatment (success or failure) and patient survival end points were summarized by site of infection and by underlying disease. In addition, survival and global responses to treatment were compared for subjects infected with S. prolificans and subjects infected with S. apiospermum. For these two groups, Kaplan-Meier and logistic regression analyses were performed and the impact of two covariates, site of infection and underlying disease, was evaluated. Logistic regression analyses, controlling for the above factors, were conducted for the global overall response, survival, and death due to fungal infection. Data were analyzed using SAS software (43).
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. Patient demographic data
|
All patients received voriconazole at standard doses, at least initially (6 mg/kg of body weight intravenously twice daily on day 1, followed by 4 mg/kg intravenously twice daily and then a switch to oral therapy at 200 mg twice daily). Dose escalation due to insufficient response to therapy was identified in 4/77 (5%) patients, but no information was available for the remainder. The overall duration of voriconazole therapy ranged from 1 to 802 days (median, 103 days) with 88 (82%) patients receiving >14 days of therapy, while 23 (21%) received treatment for a year or more (Tables 2 and 3).
|
View this table: [in a new window] |
TABLE 2. Scedosporium infection: clinical response and site of infection
|
|
View this table: [in a new window] |
TABLE 3. Scedosporium infection: clinical response and underlying conditiond
|
Cancer was a common underlying comorbidity (29/107 [27%]), with most cases affecting those with hematological malignancies (22/29 [76%]) (Table 3). Patients with underlying conditions categorized as "other" formed a large group (27%) and included 16 who had recent trauma or surgery, five with diabetes mellitus, and two with near-drowning incidents (Table 3). In total, 24/34 (71%) organ transplant recipients had received solid organ transplants (11 lung, nine kidney, three heart or heart-lung, and one liver) and 9% had hematopoietic cell transplants.
Scedosporium species. A total of 70 (65%) patients were diagnosed by the investigator as infected with S. apiospermum, and 35 (33%) were diagnosed as infected with S. prolificans. One patient was infected with both species (considered S. prolificans for analyses since this species is less susceptible to voriconazole and other antifungal agents and more difficult to treat), while for the other no species identification was available.
Response to therapy. A successful global response to therapy was achieved in 61 (57%) patients, with 33 (31%) considered to have a complete response, 28 (26%) to have a partial response, six (6%) to have stable disease, and 40 (37%) assigned as treatment failures. The median voriconazole therapy duration in treatment success cases was 180 days (range, 13 to 799 days). In contrast, the median was 19 days (range, 1 to 802 days) for the therapy failures (n = 46). There was no difference in success rates between patients diagnosed as having definite and those diagnosed as having probable infections (57% versus 56%, respectively). Some patients (28/107 [26%]) received no prior therapy or only fluconazole prophylaxis and were regarded as receiving primary voriconazole therapy (Table 1). Their response rate to voriconazole therapy was 61% compared with 56% for the other 79 patients who had received one or more antifungal agents (described above). Infections of the skin/subcutaneous tissues had the highest success rate (91%), followed by bone (79%) and lungs/sinuses (54%). For the other body sites, the success rate was <50%, including 43% for CNS infections (Table 2).
Response to voriconazole therapy varied by underlying condition. The highest success rate was observed for patients with conditions described as "other" (72%) (patients with relatively minor or no defined compromise), followed by solid organ transplantation (63%) and other hematological conditions (60%) (Table 3). Conversely, patients with cancer and hematopoietic stem cell transplantation had the poorest success rate (40 to 45%) (Table 3). Of note, 21/30 patients (70%) with skin/subcutaneous tissue or bone infections had "other" or solid organ transplantation as their underlying condition. Species (P = 0.0456) and site of infection (P = 0.0316) were shown to be statistically significant prognostic factors for overall global response.
Successful responses to therapy were observed for 45/70 (64%) of S. apiospermum- and 16/36 (44%) of S. prolificans-infected patients. This difference in rate of success between the two groups was barely not significant (P = 0.052); however, when adjusting for site of infection, the rate of success difference between the two groups became significant (P = 0.046; data not shown). All nine successfully treated CNS infections were due to S. apiospermum, although one patient was infected with both species.
Patient survival.
Known patient survival ranged from 1 to >2,000 days (median, 131 days). It exceeded 1 year in 21% of patients, including six with cerebral, four with disseminated, and five with pulmonary infection. Successfully treated patients had a median survival time of 252 days (range, 13 to >2,000 days), but in therapy failures the median was only 21 days (range, 1 to 802 days) (P = 0.001). The all-cause mortality was 40% (43/107) with 74% of deaths attributed to Scedosporium infection. All 19 patients who received
14 days of therapy died, 79% from fungal infection. Of 64 patients not known to have died, the length of the observation period varied from 18 to >2,000 days (median, 246 days) from start of therapy. It exceeded 365 days in 21 (33%) cases.
Patients with infection identified at "other" body sites (275 days), skin/subcutaneous tissue (178 days), or bone (147 days) had the longest median survival times, while those with lung, disseminated, or CNS infection had the lowest (101 to 116 days; Table 2). Of cancer patients, those with hematological malignancy had substantially shorter median survival times than those with other cancer types (median of 34 versus 108 days). Patients with underlying conditions assigned as "other" (221 days) and solid organ transplantation (140 days) had the best median survival times (Table 3).
Both species (P = 0.0247) and underlying disease (P = 0.0440), but not site of infection, were shown to be statistically significant prognostic factors for survival. The odds ratios are shown in Table 4, and a Kaplan-Meier analysis by species is shown in Fig. 1. There were significant differences in the overall number of patients surviving, both with (P = 0.0247) and without (P = 0.0259) adjustment for the effect of underlying disease. Similar results were observed for the number of patients dying due to fungal infection (P = 0.0020/P = 0.0029).
|
View this table: [in a new window] |
TABLE 4. Logistic regression analysis of survival by Scedosporium species for primary underlying disease
|
![]() View larger version (8K): [in a new window] |
FIG. 1. Kaplan-Meier plot of all-cause death within the first year for S. apiospermum versus S. prolificans.
|
1.0 µg/ml. For amphotericin B, 26/28 (93%) S. apiospermum isolates had MICs of
2.0 µg/ml, while all 14 S. prolificans isolates had MICs of
4.0 µg/ml. Voriconazole was more active than itraconazole or amphotericin B against S. apiospermum (MIC90, 1 µg/ml) and was as active as, or more active than, both comparator agents against S. prolificans (5/15 isolates had MICs of <4.0 µg/ml). Not only were S. apiospermum isolates more susceptible to voriconazole in vitro than were S. prolificans isolates, but the 28 patients infected with these S. apiospermum isolates had a higher rate of response to voriconazole therapy (MIC50, 0.25 µg/ml; successful response rate, 54%) than did the 15 S. prolificans-infected patients (MIC50, 4.0 µg/ml; response rate, 40%) (Table 5). |
View this table: [in a new window] |
TABLE 5. In vitro susceptibilities of incident baseline Scedosporium isolates to voriconazole, itraconazole, and amphotericin B and responses to voriconazole therapy
|
Safety and toleration. Safety data were available for 86 (80.3%) patients (70 from the global clinical trials database and 16 from the Australian epidemiology study). Overall, 84 patients experienced 655 adverse events (77 posttreatment) including three Australian patients. Of the adverse events, 220 were serious and 175 severe. All events labeled as "serious" met the regulatory definition of serious adverse events. Only 133 (20.3%) events were considered possibly voriconazole related. In 15 patients voriconazole was discontinued temporarily or the dose was reduced and in 10 patients it was stopped permanently. Treatment-emergent, voriconazole-related serious adverse events included elevations in liver function tests (six), hallucinations (one), visual disturbances (one), abdominal pain (one), hypotension (one), eosinophilia (one), and relapse of acute myelogenous leukemia (one), accounting for 4 of the 10 permanent study drug-related discontinuations.
|
|
|---|
Pharmacokinetic data show that voriconazole is distributed throughout the body, including the brain, eyes, and other tissues (17, 26, 42). These data are supported by the results of this study, where voriconazole was used successfully to treat a wide range of Scedosporium infections. The findings are consistent with efficacy results for voriconazole for other fungal infections of the CNS (44), bone (31), and other body sites (18, 38). Moreover, we note that for CNS, pulmonary, bone, and subcutaneous/skin infections, efficacy is better than or similar to that published for invasive aspergillosis (18, 31, 44).
Overall, 57% of patients responded successfully to therapy, a higher percentage than that reported in other studies with salvage or compassionate therapy in patients with various fungal infections (28, 38, 49). Possible explanations include the differences in patient mix or causative fungal species or the fact that 26% of patients in the current study received voriconazole as primary therapy and thus may have been less seriously ill. The overall mortality rate (40%) in this study also compares favorably with those reported previously (2, 3, 12).
Unsurprisingly, patients with superficial infection (involving skin without dissemination) or single primary sites of infection (bone) showed better responses and survival times than those with dissemination or involvement of deep organ sites. Brain and/or disseminated disease due to both S. apiospermum and S. prolificans has previously been associated with high mortality rates (60 to 100%) (2, 3, 12, 24, 34); however, in the current study, survival times exceeding 1 year were recorded in some patients with cerebral, disseminated, or deep organ infections. To determine if these favorable survival data extend into the longer term (>2 years) requires a prolonged period of observation post-commencement of therapy. Although not uniform in our study, this period extended to a median duration of >8 months (246 days).
A significant challenge in determining therapeutic efficacy in this complex patient population is the high mortality related to underlying illness (e.g., cancer) and the influence of host immunity. The poorest responses were seen in hematology patients receiving cancer chemotherapy and in hematopoietic stem cell transplant recipients, consistent with the high mortality reported previously (24, 27, 45). Conversely, patients with no or relatively low levels of immune suppression had superior responses to voriconazole therapy, suggesting that both an effective drug and effective host immunity are required for a favorable response. However, solid organ transplant patients also responded well (63% success rate); in two other reports, voriconazole showed promise as an antifungal agent in a small number of transplant patients (6, 20). We also noted longer survival times in solid organ transplant recipients than in patients who were otherwise immunosuppressed, adding to the experience of Husain et al. (20), where receipt of voriconazole was associated with a trend toward better survival in this patient group. The role of host immunological recovery, reduction of immunosuppression, or surgery was not assessed in this series.
Importantly, the causative Scedosporium species influenced both response to voriconazole therapy and patient survival. Patients with S. apiospermum showed a higher response to voriconazole than did those with S. prolificans infection (66% versus 44%, respectively; P = 0.052). Others have noted that S. apiospermum infection is, in general, more likely to respond to antifungal therapy than S. prolificans infection (20, 24); however, differences in response rates to voriconazole between infections with the two species have largely been limited to individual case reports. Although the differences in the species-specific response rates in this study were not statistically significant, after adjusting for body site of infection, patients with S. prolificans infection were significantly less likely to achieve a successful clinical response. In addition, survival times were significantly lower for infection due to this species (P = 0.026) and the attributable mortality was higher (P = 0.002) than that for S. apiospermum infection.
It is noteworthy that all but one of the 61 patients who had a successful therapeutic response received 28 days or more of voriconazole therapy and that these patients also achieved significantly longer survival times than the 46 therapy failures, the majority of whom received <28 days of therapy. Successful responders included those with CNS, systemic, or deep organ infections. Since patients who failed to achieve successful responses were more likely to die within the first 28 days, and because of the presence of other confounding factors, we were unable to draw any firm conclusions regarding the impact of duration of therapy on outcome. Other limitations of the study include the inclusion of cases enrolled in different clinical (salvage/compassionate and postmarketing) settings and the use of investigator-based reported outcomes. Every effort was made to include only data that were comparable between studies and to provide guidelines for assigning outcome responses according to established clinical trial practice (18).
The in vitro susceptibility results confirm that voriconazole has good activity against S. apiospermum isolates recovered in different patient populations from a wide geographic distribution and that the drug was more potent than amphotericin B and itraconazole (current study) (8). Although MIC breakpoints to define resistance to voriconazole have not yet been published for molds, the MIC90 for S. apiospermum of 1 µg/ml is well within achievable serum voriconazole concentrations in humans (41). As observed previously (8, 30), S. apiospermum was more susceptible than S. prolificans to all drugs tested, including voriconazole, for which MICs for S. prolificans were slightly lower than those of amphotericin B. As a result, some authors have questioned the efficacy of voriconazole in treating S. prolificans infections (8); however, the results from the current study indicate that voriconazole has clinically useful activity against this species. In parallel with the greater in vitro potency of voriconazole against S. apiospermum, there was a higher response rate for S. apiospermum infection (54%) than for S. prolificans infection (40%). Further studies correlating in vitro susceptibility with in vivo outcome are warranted.
Given the inherent resistance of S. prolificans to currently available antifungal agents, combination therapy of an azole with terbinafine has been recommended by some authors (3, 7, 15, 19, 25, 32, 45, 50). Successful use of voriconazole with an echinocandin has also been reported (46). We are unable to support these recommendations as data on combination therapy were not available for the majority of patients in this data set.
Finally, where the data were available, voriconazole was well tolerated in these seriously ill patients despite prolonged treatment (median duration, 103 days), with 14% receiving more than 1 year of therapy. We acknowledge the absence of safety data for just under 20% of patients, including almost two-thirds of Australian patients. However, overall, only four patients discontinued therapy because of clinically significant serious adverse events possibly related to voriconazole.
In conclusion, this large study reporting the efficacy of voriconazole treatment for invasive scedosporiosis is encouraging. Efficacy was shown for S. apiospermum but also for the less-susceptible S. prolificans. Many patients had relatively long-term, successful outcomes. This was particularly notable for S. apiospermum infections, including sites considered difficult to treat such as the CNS and eye. Whether an alternative new azole, posaconazole, will prove as effective as or more effective than voriconazole in treating scedosporiosis remains to be established. Independently of the antifungal agent selected for therapy, early diagnosis to species level remains of paramount importance in overall patient management.
We also thank the many other infectious diseases physicians, clinical microbiologists, and hospital scientists for their assistance in case contribution and submission of isolates.
This study was sponsored by Pfizer Inc. P.T. was previously an employee of, and is currently a consultant to, Pfizer Inc. He received an honorarium from Pfizer in connection with the development of the manuscript. C.A. is currently an employee of Pfizer Inc. C.H.H., M.S., and D.E. are serving/have served on Antifungal Advisory Boards for Gilead Sciences, Pfizer, Australia; Merck, Sharp and Dohme, Australia; and Schering-Plough, Australia. I.L. was previously an employee of, and is currently a consultant to, Pfizer Inc. S.C.A.C. is/has been a member of the Antifungal Advisory Board of Gilead Sciences and Pfizer Inc., Australia. Editorial assistance was provided by T. Wetter of Parexel and was funded by Pfizer Inc.
Published ahead of print on 22 January 2008. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»