The misleading effect of serum galactomannan testing in high-risk haematology patients receiving prophylaxis with micafungin.

OBJECTIVES

To investigate the performance of the routine serum galactomannan (sGM) assay in the diagnosis of invasive aspergillosis (IA) in high-risk haematology patients receiving prophylaxis with micafungin.

METHODS

Retrospective study including all haematological patients who received prophylaxis with micafungin during high-risk IA episodes (neutropenic patients after chemotherapy for acute myeloid leukaemia/myelodysplastic syndrome; allogeneic haematopoietic stem-cell transplantation during early neutropenic phase or graft-versus-host disease requiring high prednisone doses) and for whom at least one sGM result was available. Episodes were classified as follows: true-positive (positive GM in the context of IA), false-positive (positive GM result in patients who had no evidence of IA), true-negative (negative GM test results and no IA), or false-negative (negative GM test in the context of IA). Non-evaluable patients were excluded.

RESULTS

Among 146 evaluable episodes, four were true-positive in the context of probable breakthrough IA (incidence of breakthrough IA, 2.7%); 111/146 high-risk episodes (76%) were considered true-negative and 31/146 (21.2%) were considered false-positive. No false-negative episodes were detected. All but one of the false-positive episodes were detected in surveillance GM tests, leading to high-resolution CT scans in eight cases (8/31; 25.8%), all of which were negative. The positive predictive and negative predictive values of sGM for surveillance and diagnostic approaches were 3.2% (1/31) and 100% (110/110) and 75% (3/4) and 100% (1/1), respectively.

CONCLUSIONS

Surveillance of asymptomatic patients receiving prophylaxis with micafungin using sGM is unnecessary, because the results are either negative or false-positive. However, sGM remains useful in the diagnosis of breakthrough IA in symptomatic patients during prophylaxis.