Invasive Aspergillosis in Allogeneic Stem Cell Transplant Recipients: Increasing Antigenemia Is Associated with Progressive Disease

Fran cois Boutboul 0 Corinne Alberti 0 Thierry Leblanc 0 Annie Sulahian 0 Eliane Gluckman 0 Francis Derouin 0 Patricia Ribaud () 0 Laboratoire de Parasitologie-Mycologie 0 Departement de Biostatistique Medicale 0 Service d'Hematologie Pediatrique 0 Service de Greffe de Moelle 0 Hopital Saint-Louis 0 Paris 0 France 0 0 Received 19 July 2001; revised 14 November 2001; electronically published 20 February 2002. logie-Service de Greffe de Moelle , Hopital Saint-Louis 1, Avenue Claude Vellefaux, 75475 PARIS Cedex 10, France The kinetics of serum Aspergillus galactomannan, as determined by enzyme-linked immunosorbent assay, was examined in 37 allogeneic stem cell transplant (SCT) recipients treated for invasive aspergillosis (IA). Fiftyeight periods of response (response episodes) were evaluated. There were 42 response episodes that were considered treatment failures and 16 that were considered good (that is, complete or partial) responses. At baseline (the first day of each new response episode), the patients who experienced treatment failure and those who had good responses did not differ significantly with regard to median galactomannan index (GMI) value. Thereafter, GMI values significantly increased in the treatment failure group, whereas no significant changes were observed in the good response group (P p .002 ). An increase in the GMI value of 1.0 over the baseline value during the first week of observation was predictive of treatment failure with a sensitivity of 44%, a specificity of 87%, and a positive predictive value of 94%. We conclude that serial determination of serum GMI values is a useful tool for assessing prognosis of IA in allogeneic SCT recipients during treatment. - Invasive aspergillosis (IA) has become one of the leading infectious complications among allogeneic stem cell transplant (SCT) recipients, with a mortality rate of 80% [1, 2]. This dreadful prognosis is related to the severe and protracted immunosuppression experienced by these patients, as well as to the frequent delay in diagnosis and insufficient treatment efficiency and/or limiting toxicity of antifungal agents. Recently, the detection, by ELISA, of circulating Aspergillus galactomannan (GM) in samples of serum or other body fluids has received considerable attention as a valuable method for diagnosis of IA that can be used early in the course of infection [35]. The usefulness of GM monitoring for assessment of therapeutic response has been established in experimental models of IA [69], but its usefulness for humans is still debated. To complement a study of ours published elsewhere [3] about the determination of GM levels for the diagnosis of IA in SCT recipients, we observed the kinetics of GM serum levels in these patients and examined the relationship to outcome. PATIENTS, MATERIALS, AND METHODS Study population. We reviewed the medical records for all allogeneic SCT recipients who developed IA during the period from January 1995 through September 1998. All patients had undergone transplantation in the Bone Marrow Transplant Unit at Ho pital Saint-Louis (Paris, France). Diagnoses of IA were reviewed by the Hospital Aspergillus Study Group. Diagnoses were made on the basis of the findings of clinical charts, imaging studies (including CT scans), and mycological criteria. Patients with a definite diagnosis of pulmonary IA were those for whom fungal hyphae were observed directly in bronchial alveolar lavage specimens or in 2 consecutive sputum samples and/or in culture of these specimens, in conjunction with findings on a lung CT scan suggestive of IA (i.e., halo sign or air crescent). The diagnosis was considered probable if it was based only on findings of a CT scan suggestive of IA. A definite sinus infection was defined by a needle aspiration sample or biopsy specimen that was positive for Aspergillus species, in addition to consistent CT findings. A probable sinus or cerebral infection was defined by CT findings consistent with this diagnosis in the setting of another site positive for Aspergillus species. All other sites of infection were confirmed by examination of biopsy specimens. Cultures positive for Aspergillus species were not required to define a case when examination of smear or biopsy tissue specimens revealed fungal hyphae [2]. During the study period, IA was diagnosed in 43 patients. Thirty-seven patients met the criteria for confirmed (22 patients) or probable (15 patients) IA and had 2 serial determinations of serum GM levels. These patients constituted the study population. There were 23 male and 14 female patients. Median patient age was 26 years (range, 854 years). Eighteen patients had received a genoidentical graft and 19 had received an unrelated graft. Aspergillus fumigatus was isolated in 19 patients, Aspergillus flavus was isolated in 2, and Aspergillus niger was isolated in 1. The sites of infection were lung (n p 32), brain (n p 18), skin (n p 5), digestive tract (n p 2), sinus (n p 2), kidney (n p 1), heart (n p 1), and choroid (n p 1). None of the patients had brain lesion(s) without also having another site of infection, and 22 patients had 2 sites of infection. The crude mortality rate was 89% (33 patients died). The median duration of survival was 23 days (25th75th percentiles, 992 days) after IA diagnosis and 105 days (25th75th percentiles, 66271 days) after transplantation, as determined by a Kaplan-Meier estimate. IA was the main cause of death for 18 patients (54.5%), and it was an associated cause of death for 12 additional patients (36.4%). An autopsy was performed for 3 patients. Most patients received treatment with several sequential courses of antifungal agents. Patients were switched from one treatment to another for maintenance or because of either treatment failure or toxicity. In this series of 37 patients, 27 patients received amphotericin B deoxycholate (median duration of treatment, 3 days), 28 received a lipid formulation of amphotericin B (median duration, 17 days), 14 received itraconazole (median duration, 43 days), and 4 received voriconazole (median duration, 52 days). Twelve patients underwent adjunctive surgery (11 thoracic and 1 neurosurgical); resection of the lesion(s) was complete for 6 of these patients and incomplete for 6. Definition of responses and of response episodes. Response to treatment was assessed clinically and by imaging. Responses were classified as complete response (i.e., disappearance of all lesions visible on a CT scan), partial response (i.e., reduction of 150% of all lesions), or treatment failure (i.e., any other responses). Complete and partial responses were considered good responses. The minimum period required for evaluation of a response was 7 days, except for 6 patients with rapidly progressive disease who died within 1 week of starting treatment, who were considered to have had treatment failures. For the purpose of this study, evaluations of response were determined retrospectively, a (...truncated)