1210

Value of

antigen detection in predicting invasive pulmonary aspergillosis

Two ELISAs were used to detect serum and urinary aspergillus antigen in 121 patients who were profoundly neutropenic after leukaemia therapy or bone marrow transplantation. The presence of antigen correctly predicted development of invasive pulmonary aspergillosis (IPA) in 16 patients. In 2 other cases antigen appeared after the clinical diagnosis had been made, while in only 1 case was antigen not detected. In 11 of 13 episodes of clinically suspected fungal infection antigen was detected before clinical diagnosis was made. By contrast, antigen was detected in only 1 of 90 patients who had no evidence of IPA. Both ELISAs gave positive and negative predictive values for IPA of greater than 95%, demonstrating the value of antigen detection in early diagnosis of aspergillus infection and the assay’s ability to predict subsequent development of IPA. We conclude that neutropenic patients should be screened for aspergillus antigen, and propose that initial detection of fungal antigen justifies commencement of empirical antifungal therapy. Such an approach should improve the survival of patients who are at risk of developing this usually fatal infection.

Introduction Infection is an almost invariable complication of remission induction therapy of leukaemia, and bone marrow transplantation. Traditionally, bacterial infections have predominated; however, in patients who are severely neutropenic for prolonged periods, fungi, notably Aspergillus spp, are emerging as significant pathogens, 1-3 and invasive pulmonary aspergillosis (IPA) has become the most frequent cause of death due to infection in our bone marrow

transplant patients.4 Despite a better awareness of the need for empirical antifungal therapy during persistent fever, death is once infection becomes established.6 This is largely a consequence of the difficulty of establishing a clinical or microbiological diagnosis of IPA, and the ensuing delay in starting antifungal therapy. Clinical signs and common

symptoms are usually non-specific,7 as are radiological findings,8 while culture of the organism from upper respiratory sites does not always indicate invasive infection.9 Furthermore, antibody response to acute infection is generally poor or non-existent in these patients. 10 Definitive diagnosis of IPA rests on the histological demonstration, or culture, of the organism from open lung or transbronchial biopsy specimens," however, these procedures are associated with significant morbidity and occasional mortality. Demonstration of aspergillus antigen in serum ’12-17 urine,18 or bronchoalveolar lavage fluid19 provides an alternative approach to diagnosis. To establish the value of

antigen detection in the diagnosis of aspergillus infection, we undertook a study in high-risk neutropenic patients using an inhibition enzyme-linked immunosorbent assay (ELISA) method.

Patients and methods Patients 121 consecutive patients (59 male, 62 female; mean age 17’8years, range 5 months to 63 years) undergoing bone marrow transplant (73) or remission induction therapy (48) at Westminster Hospital and Westminster Children’s Hospital were studied. Their underlying diseases were: haematological malignancies (78), inborn errors of metabolism (39), solid tumours (2), aplastic anaemia (1), and congenital immunodeficiency (1). The frequency of bacterial and viral infections was similar to that reported by other unitS,20 although there was less cytomegalovirus infection in our patients due to their lower age. 19 patients had proven IPA (9 male, 10 female; mean age 152 years, range 6 months to 63 years). In 14 cases, diagnosis was confirmed by demonstration of septate branched hyphae and culture of Aspergillus sp from deep tissue specimens. Diagnosis of the remaining 5 patients depended on a temperature greater than 38°C for more than 5 days and unresponsive to systemic antibacterial agents, and repeated isolation of the same Aspergillus sp from sputum, bronchoalveolar lavage, or arterial blood, in the presence of pulmonary infiltrates. The organisms causing IPA were Aspergillusfurnigatus (14 patients), mixed A fumigatuslaflavus (2), Aflavus (1), and A glaucus group (1). In I case the fungus was not

speciated. infection (CFI) was seen in 12 as temperature greater than defined patients (13 episodes). 38°C for more than 5 days which was unresponsive to antibacterial agents with no evident aetiology. Pulmonary infiltrates were documented in 9 patients. Other systemic fungal infections (OFI) developed in 8 patients (7 invasive candidosis [all Candida albicans], 1 invasive mucormycosis [Rhizopus oryzae]). No evidence of fungal infection (NEFI) was seen in the remaining 82 patients.

Clinically suspected fungal It

was

Clinical specimens Serum was collected from all 121 patients at admission and as frequently as possible thereafter (usually twice weekly) until death or discharge from hospital. Sera were stored at - 70°C until just before assay when they were thawed, diluted 1:5 in phosphate buffered saline (PBS), and heated at 80°C for 30 min to dissociate immune complexes. Urine samples were available from 8 patients with IPA, 9 with CFI, 2 with OFI, and 75 of the NEFI group. Urines were dialysed extensively against distilled water before storage at — 70°C. They were not heated before assay. 100 sera and 81 urines from apparently healthy individuals were also collected and assayed to establish "negative" control values for the two

ELISAs.

ADDRESS: Department of Medical Microbiology, Charing Cross and Westminster Medical School, London SW1P 2AR, UK. (T R. Rogers, FRCPath, FRCPI, K A. Haynes, BSc, R. A Barnes*

MRCPath). Correspondence to

Dr T R

Rogers.

*Present address Department of Medical Microbiology, Wales, Cardiff, UK

University of

1211

TABLE I-NUMBER OF SAMPLES ASSAYED BY EACH ELISA

ELISA Two inhibition ELISAs

were

used. The

anti-aspergillus

antibody was either a high-titre human anti-aspergillus serum (pAb ELISA), or a rat IgM monoclonal antibody (Eco-Bio, Genk, Belgium) to A fumigatus galactomannan (GM) (mAb ELISA). The ELISA method was essentially the same for each assay.15 Briefly, specimens (180 ul) were pre-incubated with one of the inhibiting antibodies (20 ul), then added in duplicate to microtitre plates which had been freshly coated with a crude extract of A

fumigatus. After 60 min at 37°C plates were washed and 100 pl of an appropriate secondary antibody, conjugated with horseradish peroxidase, was added to each well. After a further washing step, colour was developed with H202/0-phenylenediarnine dihydrochloride and absorbance read at 492 nm (A492)’Results were calculated using the formula: percentage inhibition = (1- A492 test/A492 antibody alone) x 100. A standard curve was included with each ELISA so that concentrations of aspergillus antigen in test specimens could be calculated. For the pAb ELISA this was a concanavalin-A-binding fraction of a water-soluble extract of A fumigatus NCPF2109,21 and for the mAb ELISA, GM prepared from A fumigatus NCTC1028 was used (gift from J. P. Latge, Mycology Unit, Institute Pasteur, Paris). These antigens were titrated from 1250 ng/ml to 0-6 ng/ml in either dialysed negative control urine or negative control serum diluted 1:5 in PBS, depending on the specimens being assayed. Limits of sensitivity for each assay in urine and serum were determined from 30 replicate standard curve experiments. The mean limit of detection for the pAb ELISA was 13-0 ng/ml of antigen in urine and 14-0 ng/ml in serum. The corresponding limits for the mAb assay were 10 ng/ml of GM in urine and 7-0 ng/ml in serum. For each patient group, the number of sera and urines assayed by each ELISA are shown in table i. On completion of treatment each patient’s ELISA results were correlated with the laboratory and clinical information. However, during treatment no ELISA results were communicated to the

attending physicians; therefore, they could management of any patient.

not

have influenced

Results

pAb and mAb ELISAs were capable of distinguishing patients with proven IPA from those with OFI and NEFI (fig 1). 18 of 19 patients (95%) with IPA were antigen positive in at least one specimen by both ELISAs. Furthermore, 8 (67%) of the CFI patients were positive by the pAb ELISA, and all 12 were positive by the mAb assay. Only 1 patient (1-5%) from the NEFI group was positive by the pAb ELI SA; no positives were seen with the mAb assay in this group, and no OFI patients were positive by either assay. The sensitivity, specificity, positive predicitive value, and negative predictive value for the diagnosis of IPA compared with the OFI and NEFI groups were 95%, 99%, 95%, and 99%, respectively, for the pAb ELISA, and 95%, 100%, 100%, and 99%, respectively, for the mAb ELISA. Significantly more urine samples than sera from patients with IPA were positive for antigen: in the pAb ELISA, 34 of 173 urines (20 %) were positive compared with 61 of 482 sera (13%) (&khgr;2 = 5.02; p

Value of antigen detection in predicting invasive pulmonary aspergillosis.

Two ELISAs were used to detect serum and urinary aspergillus antigen in 121 patients who were profoundly neutropenic after leukaemia therapy or bone m...
522KB Sizes 0 Downloads 0 Views