Objective: To evaluate the value of (1, 3)-β-D-glucan (G test) in bronchoalveolar lavage fluid (BALF) in diagnosing and determining the therapeutic effects among the critically ill patients suspected with invasive pulmonary fungal infection (IPFI)., Methods: Study was conducted between February 2010 and August 2011 in medical intensive care unit (MICU) of the Affiliated Hospital of Guiyang Medical College. Patients suspected of suffering from IPFI were enrolled. G tests of BALF and serum, and culture and microscopic examination of BALF were performed twice weekly.Clinical feature, mycological evidence and the results of G test were recorded. The results of IPFI were defined as "proven", "probable", "possible" and "non-IPFI". G test in BALF and serum was done with tachypleus amebocyte lysate. Positive G test was defined as cut-off index ≥20 ng/L in two consecutive measurements. Furthermore, the patients with positive G test received preemptive antifungal therapy with fluconazole or itraconazole. G tests of BALF and serum were respectively done on 7th day and 14th day of treatment., Results: Ninety-eight patients were included. Among them, 10 patients were "proven" in whom the positive rate of BALF G test was 90.0%; 29 patients the results were considered as "probable" in whom the positive rate of BALF G test was 82.8%; in 32 patients the results were "possible" IPFI in whom the positive rate of BALF G test was 71.9%, 27 patients were "non-IPFI" in whom the positive rate of BALF G test was 7.4%. The positive rate of BALF G test was 84.6% (33/39), the positive rate of serum G test was 59.0% (23/39), the positive rate of culture of BALF was 41.0% (16/39), the positive rate of microscopic examination of BALF was 38.5% (15/39) in "proven" cases and "probable" cases. G test (cut-off ≥20 ng/L) of BALF had shown to have sensitivity, specificity, positive predictive value (PPV) of 84.6% (33/39), 92.6% (25/27), 94.3% (33/35), respectively, and negative predictive value (NPV) of 80.7% (25/31). The G test detection (cut-off ≥20 ng/L) in serum had shown to have sensitivity, specificity, PPV of 58.9% (23/39), 88.9% (24/27), 88.5%(23/26), respectively, and NPV of 60.0% (24/40), and the differences in sensitivity were statistically significant (P < 0.05). BALF G test assay tended to become positive earlier than the culture for 2-8 days with mean of (5.35 ± 2.26) days.Forty out of 56 G test positive patients were given preemptive antifungal therapy for 2 weeks, and there was a good response in 31 patients, but no response in 9 cases with 22.5% mortality. After treatment, the result of G test (ng/L) was lowered in patients with a good response in treatment group (BALF: 245.13 ± 43.84, 174.00 ± 13.01, 28.52 ± 7.38; serum: 93.26 ± 18.75, 72.15 ± 12.90, 37.37 ± 10.45, all P < 0.05). On the other hand, an elevated value suggested an unsatisfactory result in ineffective group (BALF: 267.58 ± 54.63, 309.71 ± 82.47, 486.72 ± 98.21; serum: 101.58 ± 12.75, 98.07 ± 27.45, 112.07 ± 19.21, all P < 0.05). There were significantly differences in the results of G tests on 7th day and 14th day between BALF G test and serum G test in both groups (all P < 0.05)., Conclusions: It is suggested that BALF G test may be an useful test for early diagnosis of IPFI. Moreover, the dynamic change in G test values could be useful for assessing therapeutic response.