OBJECTIVETo evaluate the diagnostic value of serum galactomannan antigen (GM) and (1→3)-β-D-glucan antigen (BG) assay in invasive fungal infections (IFI) in the patients with hematologic malignancies and the role in monitoring therapeutic response.

METHODSFifty one patients with hematological malignancies met the criteria for inclusion: (1) body temperature above 38°C for 48 hours, (2) failure to respond to broad-spectrum antibiotic treatment, or (3) temperature rose again after the responded drop. Blood samples were collected twice at the first week, then once a week in at least four weeks. The double antibody sandwich enzyme-linked immunosorbent assay (ELISA) and colorimetric assay were used for detecting GM and BG. The positive GM test is defined as two consecutive tests at different time GM value > 0.5 or > 0.8 and the positive G test is defined as BG value > 80 pg/ml. The patients were assigned into four groups as proven, probable, possible, and non-fungal infection respectively, and 21 normal volunteers were as controls.

RESULTSTwo hundred and forty serum samples were collected from 51 patients including 2 of proven IFI, 26 probable IFI, 17 possible IFI and 6 non-fungal infection. The true-positive group including the proven and probable groups, and true negative group was the non-fungal infection group. GM tests were positive in 21 of 28 cases in true positive group, and only one of 6 cases in non-fungal infection. The sensitivity, specificity, positive predictive value and negative predictive value were 75%, 83.3%, 95.5% and 41.7%, respectively. G tests were positive in all 28 cases of the true positive group, and 4 in 6 non-fungal infection cases. The sensitivity, specificity, positive predictive value and negative predictive value were 100%, 33.3%, 87.5% and 100%, respectively. G test is more sensitive than GM test (P = 0.015), but there was no significant difference in specificity of the two tests (P = 0.242). In 19 of 21 patients with GM test positive, anti-fungal treatment was effective, and GM value gradually decreased to negative, two invalid patients were persistent with GM test positive. After two weeks treatment, the average GM value was significantly lower in the effective group than in the ineffective group (P < 0.05). BG values in the responded patients showed a gradual decline similar to that of GM values, but not to negative. The changes of BG value in ineffective group varied with a trend upward. The changes in BG value had no relation with treatment effectiveness.

CONCLUSIONSSerum GM and BG antigens detection provides strong evidence for early diagnosis of IFI. Combination of GM and G tests can improve the diagnostic specificity and reduce the false positive GM test seems superior to G test for monitoring GM and BG values during treatment.