Print E-mail
Volume 49, Number 1, February 2016

Threshold of galactomannan antigenemia positivity for early diagnosis of invasive aspergillosis in neutropenic children 

Veronique Dinand, Madasu Anjan, Jaswinder Kaur Oberoi, Shilpi Khanna, Satya Prakash Yadav, Chand Wattal, Anupam Sachdeva


Corresponding author:

Anupam Sachdeva
Corresponding author. Pediatric Hematology Oncology and Bone Marrow Transport Unit, Department of Pediatrics, Institute of Child Health, Sir Ganga Ram Hospital, New Delhi 110 060, India. 


Background and purpose: 

Invasive aspergillosis (IA) is an important cause of morbidity and mortality in immunocompromised patients. Pediatric data on the accuracy and optimal cutoff of galactomannan antigen detection to diagnose IA is sparse and controversial. We evaluated the utility and optimal serum galactomannan assay (GA) cutoff in children. 



Children with febrile neutropenia due to malignancy, hematopoietic stem cell transplant, aplastic anemia, or congenital neutropenia, were prospectively included from 2007 to 2011. All new episodes of febrile neutropenia were recorded. In case of a previous diagnosis of IA, subsequent episodes were excluded. One to four GA were tested by enzyme immunoassay during each episode. Bronchoalveolar lavage and other relevant samples for mycological diagnosis, and computed tomography of chest/sinus were performed wherever appropriate. IA was classified as “proven”, “probable”, and “possible” as per the 2008 European Organisation for Research and Treatment of Cancer and Mycoses Study Group Guidelines. The optimal cutoff value was determined using receiver operating characteristic curves in episode-wise analysis. 



There were 145 patients with 211 febrile episodes included: hematopoietic stem cell transplant (n = 15), oncological (n = 113), and hematological disorders (n = 17). Forty-five children (31.0%) developed IA (5 proven, 15 probable, and 25 possible). Cutoff value of single GA ≥ 0.7 for proven/probable/possible IA offered the best combination of sensitivity (82.2%)/specificity (82.5%), and 94.4% negative predictive value. Two consecutive positive GA ≥ 0.7 had a sensitivity/specificity of 75.0%/91.0%. Index GA ≥ 1.9 was associated with significantly higher mortality in children with IA and overall. 



Serum GA is sensitive to diagnose IA in pediatric patients with excellent negative predictive value at an optimal cutoff of ≥0.7. Considering two consecutive values ≥0.7 increases specificity to 91.0%.


Key words:

Aspergillosis, Child, Hematology/oncology, Infections in immunocompromised hosts, Neutropenia, Stem cell transplantation