Diagnostic value of galactomannan in bronchoalveolar lavage fluid for chronic respiratory disease with pulmonary aspergillosis

Objective To explore the diagnostic value of the bronchoalveolar lavage fluid galactomannan (BALF GM) test for chronic respiratory disease with pulmonary aspergillosis, and establish the optimal cut-off value. Methods A total of 180 chronic respiratory disease patients seen at the Respiratory Medicine Department of Peking University Shenzhen Hospital from September 2017 to September 2018 were analyzed. According to the diagnostic criteria, we divided the patients into the case group (n = 70, comprising 5, 20, and 45 proven, probable, and possible cases, respectively) and the control group (n = 110). Bronchoalveolar lavage fluid was collected, and the BALF GM test results were analyzed. Results A non-parametric rank-sum test showed that the mean rank of the case group was 140.80, which was higher than that of the control group (58.49). The Z-value was 10.335 (P = 0.000), indicating that the general distribution of BALF GM differed between the two groups. A BALF GM cut-off value of 0.485 showed the highest diagnostic efficacy for pulmonary aspergillosis. The sensitivity, specificity, positive predictive value, and negative predictive value were 92.9%, 100%, 92.8%, and 100%, respectively. As the cut-off value increased, the specificity and sensitivity of the BALF GM test increased and decreased, respectively. Conclusions The BALF GM test can be used confirm the diagnosis of patients with pulmonary aspergillosis and chronic respiratory disease. The optimum BALF GM cut-off value was 0.485. Antifungal therapy is important for treating pulmonary aspergillus infection in patients with chronic respiratory disease.

risk of pulmonary infection (imaging examination revealed the presence of lung lesions); (2) no contraindications 58 for bronchoscopy and provision of informed consent; and (3) no antifungal use before admission. The study 59 protocol was approved by the Ethics Committee of our hospital. 60

Diagnostic classification 61
According to the 2016 IDSA guidelines on Aspergillus infection diagnosis and treatment [6] , together with the 62 diagnostic criteria for invasive fungal infections used in China [7] , the cases of chronic respiratory disease with 63 pulmonary aspergillosis were classified as follows: 64 (1) Proven (confirmed diagnosis): a lower respiratory tract specimen positive for Aspergillus, with confirmation by 65 molecular, immunological, and/or culture methods that the observed hyphae were of Aspergillus. 66 (2) Probable (clinical diagnosis): frequent use of hormones or repeated use of broad-spectrum antibiotics for 67 3 months; suggestive pulmonary imaging changes within 3 months; and positive for Aspergillus or a positive result 68 of a blood or BALF GM test. 69 (3) Possible: patients with chronic respiratory diseases who frequently use hormones or have been treated with 70 broad-spectrum antibiotics for 3 months, and who recently developed acute exacerbations and presented with 71 indicative changes in lung images within 3 months (chest X-ray or computed tomography [CT]), but in whom 72 Aspergillus was not cultured from lower respiratory tract specimens and the results of microscopic and serological 73 tests were negative. 74 (4) Aspergillus colonization: lower respiratory tract specimens of patients with chronic respiratory diseases positive 75 for Aspergillus, but without acute exacerbation, airway spasm, or new pulmonary infiltration. 76

Group assignment 77
total of 180 patients with chronic respiratory diseases were enrolled (92 males and 88 females), comprising 58 78 patients with chronic obstructive pulmonary disease, 59 with bronchial asthma, and 63 with bronchiectasis. 79 According to the clinical symptoms and the results of auxiliary examinations, the patients with proven, probable, or 80 possible pulmonary aspergillosis were included in the case group according to the above-described diagnostic 81 standards (n = 70; 5, 20, and 45 proven, probable, and possible cases, respectively). Patients with non-pulmonary 82 aspergillosis constituted the control group (n = 110). There was no significant difference in basic characteristics 83 between the case and control groups (all, P > 0.05; Table 1 with diffuse lesions in both lungs [9,10] . (2) Injection of saline:First, the bronchoscope was embedded in the opening 94 of the bronchial segment or subsection at the lavage site. Next, at the bronchial segment or sub-segment opening, 95 60-80 mL of physiological saline was injected as a 20 mL bolus, at 37°C or room temperature. (3) Vacuum suction: 96 10-15 s after infusion of normal saline, BALF was obtained at a negative pressure of < 100 mmHg (total 97 recovery, > 50%). Five milliliters of BALF were used for GM testing and a further 5 mL for fungal culture and 98 other tests. (4) Finally, bronchial biopsy, lung biopsy, or bronchial brushing was performed according to the 99 characteristics of the lesion. 100

BALF GM assay 101
GM in BALF was quantified by sandwich assay using an antibody against Aspergillus GM. The absorbance 102 value was determined in accordance with the instructions of the Enzyme-linked Immunosorbent Assay Kit for 103 Aspergillus, and the galactomannan index (GMI) value was calculated. Negative and positive controls were 104 performed in each experiment. 105

Data analysis
Data were analyzed using SPSS software (ver. 23.0; SPSS Inc., Chicago, IL, USA). Measurement data are 107 presented as means ± standard deviation, and count data as percentages or numbers of cases. The data were tested 108 for normality; normally distributed data were evaluated by independent sample t-test and chi-squared test, and 109 non-normally distributed data were subjected to non-parametric tests. The diagnostic efficacy of BALF GM was 110 evaluated by calculating the sensitivity, specificity, positive predictive value (PPV) and negative predictive value 111 (NPV). Finally, a receiver operating characteristic (ROC) curve was constructed to determine the optimum cut-off 112 value of BALF GM. 113 2 Results 114

GM status and pathogens in BALF in the case group 115
Using a GM cut-off value of 0.5, the GM status and pathogens detected in the case group are shown in Table 2. 116 The BALF GM test had a higher positive rate than the serum GM test, and BALF pathogen culture showed some 117 diagnostic utility. 118

BALF GM 122
By non-parametric rank-sum test, the mean rank of the case group was 140.80, which was higher than that of 123 the control group (58.49) ( Table 3). The BALF GM value of the case group was higher than that of the control 124 group. Mann-Whitney U and Wilcoxon W tests (Table 4)

Diagnostic efficacy of the BALF GM test 132
We constructed a ROC (Figure 1)

Treatment status of patients in the case group 144
Seventy patients in the case group received antifungals, such as voriconazole and fluconazole. After treatment 145 with antifungals, 20 cases were cured and 40 were improved (efficacy rate, 86%). Ten cases experienced adverse 146 effects (adverse event rate, 14%), one of whom died (Table 7). 147

Discussion 150
The incidence of pulmonary aspergillosis in patients with chronic respiratory diseases is increasing [10] . Because 151 of its atypical clinical manifestations, pulmonary aspergillosis can be obscured by the symptoms of chronic 152 respiratory diseases. Without treatment, the mortality rate increases significantly [11] , and patients with acute 153 exacerbations of chronic respiratory diseases suffer cough and shortness of breath. Such patients may also have 154 concomitant aspergillosis, which can be detected by chest imaging and is characterized by multiple plaques; when 155 it cooccurs with IPA, the symptoms are likely to be confusing. Therefore, it is important to improve the diagnostic 156 accuracy for chronic respiratory diseases cooccurring with IPA [12] . Pathogen culture takes some time to perform and 157 has a low positive rate, so is of limited utility for early diagnosis of pulmonary aspergillosis. Furthermore, the 158 decision for histopathological examination must be based on the patient's tolerance and economic resources, 159 limiting its clinical utility. GM, in serum, BALF, or cerebrospinal fluid is a biomarker of Aspergillus. The 160 diagnostic efficacy of GM testing is high, and GM is now considered a diagnostic marker for aspergillosis [13] . 161 Sulahian et al. [14] reported that the results of GM tests are available 5-8 days earlier than those of imaging 162 modalities. BALF can also be analyzed for Aspergillus, and BALF originates from the site of infection [15] . The 163 positive rate of fungal cultures obtained by bronchoscopy is 46-75% [16,17] , and varies depending on disease course, 164 disease severity, and whether antifungals have been administered. 165 Only seven patients in this study were positive for GM in serum; Aspergillus fungi did not invade the blood vessels, suggesting that these patients may still have been in the early stage of the disease. The serum GM-positive 167 rate was relatively low, suggesting that blood and BALF GM tests may be more suitable for the diagnosis of 168 chronic respiratory diseases and pulmonary aspergillosis. Meersseman et al. [18] reported that among 25 patients 169 with confirmed COPD and pulmonary aspergillosis, only 12 were positive for serum GM, but all were positive for 170 BALF GM. In a prospective, single-center study [19] , among 110 patients with non-granulocyte deficiency, GM in 171 BALF was more sensitive as an early diagnostic marker of pulmonary aspergillosis than GM in serum; the 172 sensitivity values were 88% and 42%, respectively, with comparable specificity. At present, there is no standard 173 cut-off value for the BALF GM test [20][21][22][23] . In a meta-analysis of the diagnostic performance of GM, Junling et al. [24] 174 showed that, with cut-off values of 0.5, 0.8, and 1.0, the sensitivity was 0.75, 0.69, and 0.68, and the specificity was 175 0.89, 0.94, and 0.96, respectively. 176 BALF GM cut-off values are set for patients with pulmonary aspergillosis with or without a variety of 177 underlying diseases, and may not be suitable for patients with chronic respiratory disease [25] . Therefore, a Huiqing et al. [26] found that early empiric antifungal therapy significantly improved the prognosis compared with 200 antifungal therapy after diagnosis. The antifungal treatment cycle should be ≥ 6-12 weeks, and should be 201 determined according to the patient's drug tolerance, immune status, clinical manifestations, laboratory test results, 202 and imaging findings. 203 The diagnostic utility of GM in BALF for pulmonary aspergillosis is an important area of respiratory disease 204 research. However, most prior studies focused on pulmonary aspergillosis in isolation. The BALF GM level is 205 affected by various factors; in this study of patients with chronic respiratory diseases, the optimum BALF GM 206 cut-off value for pulmonary aspergillosis was 0.485, which was lower than that reported by other studies [27] . There 207 are many risk factors for pulmonary Aspergillus infection, such as neutropenia, organ transplantation, blood tumors, 208 granular and non-granular deficiencies [28,29] , organ and non-organ transplantation [30,31] , and hematological 209 tumors [32,33]. the best cut-off value of BALF-GM in pulmonary Aspergillus patients with different risk factors 210 should be set According to the results of BALF-GM. further increasing the diagnostic efficacy of BALF GM for 211 pulmonary aspergillosis cooccurring with other conditions. For diagnosis of chronic respiratory diseases 212 cooccurring with pulmonary aspergillosis, in addition to the BALF-GM test, the diagnostic efficacy of pathogens in 213 BALF, GM in serum, and imaging modalities warrants further investigation. 214 In summary, Our study demonstrates that The BALF GM test has diagnostic utility for chronic respiratory 215 diseases cooccurring with pulmonary aspergillosis. But For pulmonary Aspergillus patients, it is perhaps that best 216 cut-off value of BALF-GM may be different according to different risk factors. More clinical data are needed to 217 explore BALF-GM values in patients with pulmonary aspergillus in different immune conditions. 218

Declarations 219
Ethics approval and consent to participate 220 The study was approved by the Ethics Committee of Peking university Shenzhen Hospital. 221

Consent for publication 222
All presentations of the study have consent to publish. 223

Availability of data and material 224
The materials described in the manuscript are available from the corresponding author on reasonable request. 225 Funding Natural Science Foundation of Guangdong Province (No 2017030313830)