Invasive aspergillosis (IA) usually occurs among severely immunocompromised patients, such as hematopoietic stem-cell transplant (HSCT) recipients, those with prolonged neutropenia, and those undergoing treatment for graft-versus-host disease (1). The most commonly isolated species, Aspergillus fumigatus, causes approximately 90% of infections and is ubiquitous in the environment. Infection follows inhalation of spores. Pulmonary infection is the most common form of invasive disease, but dissemination to the central nervous system and other sites can occur (2). IA is associated with a high mortality; the overall case-fatality rate is 58% (3). Although the incidence of IA can be up to 15% among HSCT recipients, it is less common among solid-organ transplant recipients and rare among renal transplant recipients (RTRs), in whom the incidence is less than 1%. Outbreaks of IA among RTRs are uncommon (4).
On February 6, 2001, hospital X, a 350-bed facility that specializes in solid-organ transplantation in California, reported to the local health department four IA infections among RTRs in January and February 2001. Construction was ongoing at the time on the floor where RTRs were housed. Only three patients with IA had been reported in the previous 2 years, none among RTRs. Ninety percent of all transplants performed at hospital X are renal transplants; 295 were performed in the year 2000. To determine the extent of this outbreak, potential risk factors, and environmental sources for IA in the RTR population, we conducted an investigation that included cohort, environmental, and laboratory studies.
We conducted a retrospective cohort study among all RTRs hospitalized between January 1 and February 5, 2001 (study period), using a standardized data extraction form, reviewed medical records, radiographic, microbiologic, and pathologic reports, and an existing renal transplant database. Information was collected on patient demographic and clinical characteristics, location during hospitalization, transplant characteristics, complications after transplantation, medications, laboratory findings, and outcome. We defined a case of IA as any proven or probable Aspergillus infection in an RTR at hospital X during the study period, using a consensus case definition recently published (5). (Proven IA was defined by histology showing Aspergillus sp. hyphal tissue invasion in a site with evidence of tissue damage or positive Aspergillus sp. culture from a normally sterile but clinically abnormal site. Probable IA was defined by a combination of positive Aspergillus sp. culture or cytology from respiratory secretions in a patient with a clinically compatible picture.) In addition, descriptive data were collected on three patients who developed IA after the study period.
Environmental inspection of hospital X focused on the heating, ventilation, and air conditioning (HVAC) filtration systems, areas of ongoing construction and water damage, and infection-control precautions taken during construction. Environmental surface samples were obtained from the RTR ward, operating room, intensive care unit patient care areas, elevators, HVAC systems, and the areas of ongoing construction. Identification of environmental and clinical isolates of Aspergillus spp. was performed using standard methods. Molecular typing of isolates was performed by restriction fragment length polymorphism analysis of DNA extracts using a DNA probe (Afut 1) as previously described (6).
Univariate analysis was performed using SAS 8.0 (SAS Institute Inc., Cary, NC) and StatXact 4.0 (Cytel Software Co., Cambridge, MA). When possible, continuous variables following a linear distribution were dichotomized at the 75% quartile (e.g., days on corticosteroids and dialysis duration) to create a discrete variable. Stratified analysis was performed when feasible.
Forty RTRs were hospitalized during the study period; four cases were identified (attack rate=10%) (Table 1). No cases were identified among other transplant recipients. Three RTRs met the “proven” case definition, and one met the “probable” case definition. A. fumigatus was isolated from all four cases. The median age of patients was 44 years (range 33–55 years). Three were men, and three were Hispanic. The median time from last transplantation until diagnosis of IA was 71.5 days (range: 58–415 days); all were admitted two or three times during the study period. Three presented with pneumonia, and all showed evidence of renal dysfunction and received treatment for presumed rejection. All four patients died between 3 and 11 days after IA diagnostic culture collection.
No significant associations (P <0.05) were detected between many transplant-specific or host-related factors (Table 2). Moreover, environmental factors, such as patient proximity to the construction site, which was located directly opposite the RTR ward, and the duration of hospital stay, were not found to be significantly associated with an increased risk of IA.
The patients were significantly more likely to have received sirolimus, as determined by univariate analysis. Patients who received both mycophenolate and sirolimus, simultaneously for 2 to 3 days or sequentially during the study period, were also at increased risk for IA. Other host factors, indicative of severe immunosuppression, were found to be significantly associated with the development of IA (Table 2). The risk of disease associated with the use of mycophenolate and sirolimus during the study period persisted even after stratifying by prolonged corticosteroid use and certain induction agent usage (e.g., basiliximab [Simulect, Neoral, Novartis, East Hanover, NJ]).
Three additional RTRs (patients 5–7), not included in the cohort study, developed IA (two proven and one probable) in the 4 months after our investigation (May–August 2001). Two received sirolimus and mycophenolate, all required prolonged dialysis posttransplant, and two died. In April 2001, we recommended chemoprophylaxis using oral itraconazole solution (5 mg/kg/day) for RTRs with the risk factors for IA identified by our cohort study. The duration of prophylaxis depended on adequate recovery of immune function. Patients 5 through 7 did not receive itraconazole.
We observed that the barriers between the construction site and the RTR ward were inadequate (7). In addition, much trafficking of staff between the construction and patient care areas was occurring. Moreover, RTRs demonstrated little or no respiratory protection while outside their rooms, and RTRs, construction workers, and debris shared the same elevators.
The intensive care unit, where RTRs resided immediately posttransplant, was depressurized with air flowing into the patient area from the outside corridor. RTR patient rooms did not have high efficiency particulate air filtration. The transplant team reported that immunosuppression regimens were altered for RTR, on an individual basis, by the transplant surgeon. The use of sirolimus had increased fivefold in the 2 months before the outbreak. No other changes in patient care procedures or staffing were noted.
DNA fingerprinting analysis revealed that the Aspergillus isolates from patients 5 and 6 had indistinguishable profiles; two environmental isolates had genotype matches; however, no genotype matches were observed between environmental and clinical isolates (Fig. 1).
This investigation suggests that severely immunocompromised RTRs can be at an increased risk for IA. The small number of cases limited our ability to control for potential confounders using multivariate analysis or more extensive stratification. Nonetheless, known markers of immunosuppression, such as prolonged dialysis, high-dose corticosteroid treatment, and the use of sirolimus with mycophenolate (a regimen not approved by the Food and Drug Administration), was strongly associated with IA among RTRs. Six of seven RTRs (86%) with IA died, a rate similar to that seen for HSCT recipients who develop IA (3). This high mortality is striking when contrasted with the normal 1-year survival of RTRs (8).
As mold spores are ubiquitous, and construction in hospitals is a common occurrence, the implementation of established infection-control measures to decrease exposure to airborne spores has limited the number of outbreaks of IA among high-risk patients. Noncompliance with these guidelines during hospital construction may have exposed susceptible patients to high levels of Aspergillus spores (7). Health care-related outbreaks of IA have been associated with construction work or with defective air filtration (1). Because most of the RTRs who developed IA were discharged and readmitted multiple times from hospital X, exposure to Aspergillus in the community cannot be excluded (1). This study did not detect a significant association between construction-related factors and development of IA, which could be the result of many factors, including the small number of cases involved and the difficulty in measuring environmental exposures. However, the clustering of RTRs with IA and the indistinguishable DNA fingerprints from two patients suggests transmission from a common source, likely the hospital. The lack of genotypic matches between environmental and clinical isolates is consistent with the knowledge that A. fumigatus has a genetically diverse population (1).
Although few data exist regarding the effectiveness of IA prophylaxis among solid-organ transplant recipients, chemoprophylaxis with itraconazole was recommended for high-risk RTRs to prevent further episodes of IA in this outbreak. A recent study found that itraconazole solution may be effective in preventing IA in neutropenic patients (9). However, until data from controlled clinical trials are available, itraconazole chemoprophylaxis to prevent IA in RTRs should be considered in select situations, such as ongoing outbreaks.
We recommended compliance with proven environmental protective measures that follow the Healthcare Infection Control Practices Advisory Committee guidelines, such as appropriate placement of impermeable barriers, the use of high-efficiency particulate air filters in HVAC systems, N95 respirator use by patients when traveling through potentially contaminated areas, preventing traffic between construction and patient care areas, and designating one elevator for the exclusive use of construction workers and debris removal (7).
Larger studies are needed to determine if sirolimus has an independent role as a risk factor for IA; this specific association has not been previously described (10). However, our investigation suggests that RTRs, who generally are not considered to be at risk for IA, are indeed susceptible when immunosuppressed. These patients would benefit if hospital infection control personnel had advanced warning of construction so that established guidelines for prevention of IA could be implemented during construction activities in hospitals.