There is little specific information on the fundamental epidemiology of pediatric IA. In one large study of 621 patients with IA done in the greater Paris area, the mean age was 40.3 years (range 6 days – 89.7 years) and there were no specific pediatric data (Cornet et al, 2002).
Risk factors
Another study examined risk factors for mould infection in 230 bone marrow transplant patients and enrolled patients from age 4 months – 54 years but the mean age was approximately 29 years and there was no specific pediatric analysis (Yuen et al, 1997). One study examined 409 patients for early infections after HSCT and enrolled patients from ages 6 months – 65 years (mean age 32 years), but again there was no pediatric analysis (Kruger et al, 1999). Another study of 173 allogeneic HSCT patients after non-myeloablative conditioning analyzed patients with a mean age of 53 years (range 0-72 years) and did not comment on pediatric disease (Fukuda et al, 2003). One study of 327 patients with IA from 1985-1999 did stratify patients into three age groups (Marr et al, 2002). A total of 13% of patients were < 19 years old, with 34% between 19-40 years old and 53% of patients > 40 years. The number of transplants performed in that youngest age group was not reported so a true incidence of pediatric disease cannot be calculated. The true incidence of IA in children therefore remains unknown despite numerous epidemiologic studies of the disease.
Even large-scale studies of infections in pediatric HSCT patients do not answer the fundamental questions. In a study of 148 pediatric HSCT patients from 1986-1996 with 8 proven IA patients there were also 48 patients with suspected invasive fungal infection, but the results were not stratified between Candida and Aspergillus infection and there were no specific IA analyses (Hovi et al, 2000). A larger report reviewed 510 pediatric HSCTs in 485 patients from 1990-1998 (Benjamin et al, 2002). There were 26 IA cases (4.79% of infections) in 584 culture-proven infections during the first year post-transplant. This pediatric report was the first to employ tools beyond descriptive statistics to analyze pediatric IA. A multivariable analysis showed that IA was more likely to be associated with severe graft-versus-host disease (GVHD) (RR 7.5%; 95% CI 3.0 – 18.4). Additional analyses revealed that in the first 30 days there were 10 cases of IA, 13 cases from 31-100 days post-transplant, and 3 cases from days 101-365 post-transplant.
Autopsy data from Japan for IA patients from years 1989, 1993, and 1997 analyzed the patients based on age into single decade blocks (Kume et al, 2003). From a total of 412 IA patient autopsies there were 14 IA cases in children aged 0-9 years, and 24 in those aged 10-19 years. By comparison, there were 92 cases from age 50-59 years and 102 cases from ages 60-69 years. The study did not report the total number of children in each range who underwent autopsy so it is again impossible to calculate an incidence of pediatric IA.
Case-fatality rate
In an excellent review of IA case fatality rates pooled from 1,941 patients from clinical trials, cohort or case-control studies, and case series of ³ 10 patients with definite or probable IA from 1995-1999, there was some stratification of case fatality rate by decade of life (Lin et al, 2000). The mean age of all patients was 44.2 years (range 3-91 years), and the youngest cohort ( £ 20 years) had a case fatality rate of 68.2% (15/22). The next highest case fatality rate was 59.3% in the age group of 21-30 years old. The investigators concluded that there was little variation in mortality by age, but the one pediatric case fatality rate was considerably higher than the other age cohorts. This suggests the epidemiologic possibility that pediatric and adult IA differ in outcome.
Disease manifestations
There have been two published reviews of childhood cases and both suffer from the limitations of time because diagnostic and therapeutic tools available to today’s pediatric clinician have improved substantially. In 1993 the Hospital for Sick Children in Toronto reviewed 39 cases of pediatric IA from 1979-1988 (Walmsley et al, 1993). Of those 39 cases, 24 were proven IA and 15 were probable IA. The median age of the cases was 10 years (range 22 days – 18 years), and 74% had a hematologic malignancy or were a bone marrow transplant recipient. Thirty-one of 36 patients had an absolute neutrophil count (ANC) less than 500 cells/ m l with the mean duration of an ANC < 1000 cells/ m l of 20 days. All except one patient had an immunocompromised underlying condition.
In 41% (16/39) of the total patients, the Aspergillus infection was cutaneous disease which was first suspected as a skin lesion. These lesions were described as tender, erythematous macules or vesicles which frequently progressed to necrotic eschars. The lesions typically presented at sites of trauma related to armboards or intravenous sites (69%). Skin lesions resolved in 56% (9/16) of patients, but in all cases resolution was coincident with recovery from neutropenia. In another 41% of patients, IA was first suspected based on pulmonary findings of fever and an abnormal chest radiograph (15 cases) or pleuritic pain (1 case) despite broad spectrum antibiotics. Patients were hospitalized for a mean of 47 days (range 0-180 days) in 6 months preceeding diagnosis. The overall survival rate was only 23.1% (9/39), which is a similar survival rate to the 31.8% in the age group of patients £ 20 years from the large case review (Lin et al, 2000), and generally lower than many adult studies.
The second pediatric IA study was a review of 66 cases of proven IA from approximately 9,500 children treated from 1962-1996 at St. Jude’s Children’s Hospital in Memphis (Abbasi et al, 1999). The median age was 11.2 years (range 1.3 – 21.6 years). The ANC for these patients was < 500 cells/ m l for a median of 14 days (1-402 days), and the interval between onset of underlying disease and IA was a median of 16 months (0-180 months). Sixty-six percent of patients had been hospitalized for a median of 36 days (1-52 days) before the onset of clinical disease, and clinical symptoms were present for a median of 11 days (0-69 days) before the diagnosis of IA. This report did calculate a true incidence of IA in specific pediatric subpopulations, and found that 8% (2/25) of the patients with myelodysplastic syndrome had IA, followed by an IA incidence of 7% (1/14) in chronic granulomatous (CGD) patients, 6% (1/16) in choriocarcinoma, 4.6% (2/43) in aplastic anemia, 4% (26/647) in acute myelogenous leukemia, 4% (1/24) in chronic myelogenous leukemia, and 1% (29/2659) in acute lymphoblastic leukemia. Patient survival was 58% at one month, 25% after two months, and 15% after 10 months. That is a lower survival rate than most adult studies have reported, and the authors found that pulmonary IA patients fared worse than non-pulmonary patients with an overall median time of 29 days (3-312 days) between diagnosis and death.
Species distribution
The species distribution of Aspergillus isolates for pediatric and adult patients is different in some studies. A large National Institute of Allergy and Infectious Diseases Bacteriology and Mycoses Study Group study reviewed 256 isolates of Aspergillus species from patients with IA from 24 medical centers (Perfect et al, 2001) and A. fumigatus yielded 67% (171/256) of isolates while A. flavus was the second most common isolate at 16% (41/256). This parallels the species distribution in the large voriconazole randomized clinical trial (Herbrecht et al, 2002) where 77% (85/110) were A. fumigatus and 6% (7/110) were A. flavus . However, in both the Toronto (Walmsley et al, 1993) and St. Jude (Abbasi et al, 1999) studies, A. flavuswas the predominant pathogen. In Toronto 65% (17/26) of isolates were A. flavus , followed by 15% (4/26) A. fumigatus as the second most common pediatric Aspergillus isolate. At St. Jude, 72% (28/39) isolates were A. flavus , followed by 38% (15/39) A. fumigatus isolates.
However, in two more recent studies the pediatric epidemiology paralleled previous adult studies. In the pediatric voriconazole compassionate release study, the species distribution was predominantly with patients infected with A. fumigatus (26/42), followed by A. flavus(6/42), and A. nidulans (3/42) (Walsh et al, 2002). In a French pediatric study with amphotercin B lipid complex (ABLC) the most common isolates were A. fumigatus (11/23), A. flavus (6/23), A. niger (1/23), and unspeciated Aspergillus spp. (5/23) (Herbrecht et al, 2001). This limited few pediatric experience concludes with two studies of 65 patients indicating thatA. flavus is the predominant species in children, while two more recent studies of 65 patients cite A. fumigatus as the more common species. The true answer might lay with the site of infection, as the earlier studies with A. flavus predominance cited a large percentage of cutaneous disease, (Walmsley et al, 1993, Abbasi et al, 1999) while the later treatment studies contained mostly patients with pulmonary aspergillosis.(Walsh et al, 2002, Herbrecht et al, 2001).
Treatment
ABLC treatment
There has not been a dedicated, prospective, large-scale investigation into treatment of pediatric IA. One large dataset analyzed adult and pediatric IA outcomes separately. The adult data were an analysis from the open-label, multi-center clinical study of emergency use of amphotericin B lipid complex (ABLC) at 5 mg/kg/d from 1990-1995 in the treatment of patients with proven or probable invasive fungal infections who either failed to respond to previous systemic antifungal therapy (including AmB at a cumulative dosage of at least 500mg), developed severe nephrotoxicity (serum creatinine ³ 2.5 mg/dL in adults or ³ 1.5 mg/dL in children), had pre-treatment renal insufficiency, or developed severe acute toxicity (Walsh et al, 1998). Patients were excluded if a systemic antifungal or investigational drug was administered concurrently with the ABLC. This study evaluated 551 patients with 556 courses of ABLC therapy, of which 291 patients fulfilled criteria for evaluation of efficacy. The mean age of all enrolled patients was 37.2 years (range 21 days – 93 years). There were 130 cases of IA, with a complete or partial response rate of 42%, stable response in 12%, and failure in 45% of patients. The complete or partial response rate of pulmonary IA (n=74) was 38%, disseminated IA (n=27) 30%, sinusitis (n=14) 64%, and single-organ extrapulmonary IA (n=15) 67%. The analysis of 556 treatment episodes did not stratify pediatric responses, but a subsequent analysis was performed on 54 of 111 pediatric (< 18 years old) patients from that study (Walsh et al, 1999). The mean patient age was 9.3 years (range 21 days – 16 years). There were 25 cases of IA with a complete or partial response rate of 56%, stable response in 8%, and failure in 36% of patients. The complete or partial response rate of pulmonary IA (n=10) was 50%, disseminated IA (n=7) 29%, sinusitis (n=5) 100%, and single-organ extrapulmonary IA (n=3) 67%.
In a retrospective French study of 46 pediatric patients treated with ABLC for invasive fungal infections from 1994-1997 (Herbrecht et al, 2001) the mean age of 23 cases of IA was 9.7 years (3 months – 18 years). Eighteen of 23 (78%) patients showed cure (52%) or improvement (26%), with 22% failing therapy. Three patients who initially improved later relapsed, dropping the cure or improvement rate to 15/23 (65%).
Voriconazole treatment
An analysis of the compassionate open label use of voriconazole in children for refractory IA with clinical or radiologic progression of disease after ³ 7 days of systemic antifungal therapy in children < 16 years included 42 patients with proven or probable IA (Walsh, 2002). The mean age of all 58 children with invasive fungal infections, including IA and 16 patients with other systemic fungal disease, was 8.2 years (range 9 months – 15 years). Analysis of the 42 patients with IA revealed a complete or partial response rate of 43%, stable response in 7% of patients, 10% of patients intolerant to therapy, and 40% failing therapy. The complete or partial response rate of pulmonary IA (n=12) was 33%, CNS (n=6) 50%, disseminated IA (n=7) 86%, sinusitis (n=7) 29%, and single-organ (bone, liver, or skin) IA (n=10) 30%.
The second largest series of pediatric aspergillosis patients who received voriconazole is seven patients with a mean age of 5 years (range 2-13 years), all with hematologic malignancies. They underwent liposomal AmB therapy for a mean of 6 weeks (range 2-18 weeks) before switching to voriconazole (Cesaro et al, 2003). There was a complete response in 2 patients, a partial response in 2 patients, stability in one patient, and failure in two patients. Complete or partial response coincided with recovery from neutropenia and hematologic remission of underlying diseases. This yielded an overall response rate of 57% and a 100-day survival rate of 42%.
Pharmacokinetics of antifungal drugs in children
Fundamental to treating aspergillosis is an understanding of the pediatric pharmacokinetics and pharmacodynamics of new antifungal drugs in children. Complete reviews of the newer antifungals and immunomodulatory strategies for IA in adults have already been published, (Steinbach et al, 2003, Steinbach, Stevens & Denning, 2003) but there are important differences in children. Voriconazole metabolism is non-linear in adults with an approximately 3-fold increase in the area under the concentration-time curve (AUC) following a 33% increase in dosage. Elimination of voriconazole is linear in children following doses of 3 and 4 mg/kg every 12 hours. This linearity was based on a eleven patient, single dose, open, two-center study in the United Kingdom of children ages 2-11 years (mean age 5.9 years) and a 28 patient multiple-dose, open, multicenter study in two age cohorts (ages 2-6, 6-12 years) (mean age 6.4 years) from eight centers (Walsh et al, 2004).
Drug exposures based on mean AUC at 4 mg/kg every 12 hours in children were similar to doses of 3 mg/kg every 12 hours in adults, suggesting that pediatric patients have a higher capacity for elimination on a weight basis than do adult healthy volunteers. Based on extrapolated plasma pharmacokinetics for the pediatric population at dosages of 5 – 12 mg/kg every 12 hours compared with adult data at 4 mg/kg every 12 hours, the pediatric dosage of approximately 11 mg/kg every 12 hours is equivalent to the adult dosage of 4 mg/kg every 12 hours in terms of AUC and mean plasma concentration. This statement is valid only when the linear pharmacokinetics of voriconazole were maintained throughout the dosage range (Walsh et al, 2004). The correct pediatric dosage is unknown but is likely much higher than that for adult patients. A new pharmacokinetic study that will enroll 36 children and test dosages of 6 and 8 mg/kg every 12 hours might clarify the situation (I. Lutsar, personal communication).
Caspofungin therapy in adults begins with a single daily loading dose of 70 mg followed by a 50 mg dose once daily. Pharmacokinetics in healthy men revealed linear pharmacokinetics and dose-proportional AUC concentration data, but pharmacokinetics appeared slightly different in children. The initial pediatric pharmacokinetic study enrolled 39 patients (ages 2-17) and data were obtained using both a weight-based (1mg/kg/d) and body surface area (70 mg/m 2 /day or 50 mg/m 2 /day) approach (Walsh et al, 2002). The weight-based approach resulted in sub-optimal plasma concentrations in all children relative to adults. The 50 mg/m 2 /day regimen yielded similar plasma concentrations and increased AUC to adult patients (50mg/d). Caspofungin had a beta half-life reduced 32-43% in children, with pharmacokinetic projections suggesting that dosing at 50 mg/m 2 /day appears to be more appropriate in children than using 1 mg/kg/day.
William Steinbach
Division of Pediatric Infectious Diseases
Box 3499
Duke University Medical Center
Durham, NC 27710 USA
Tel +1 (919) 684-6335
Fax +1 (919) 416-9268
Email: stein022@mc.duke.edu
November 2004
