Effects of postoperative antifungal therapy on the recurrence of Aspergillus infection after pulmonary aspergilloma resection: a prospective, randomized, controlled, single-center, 24-month, open-label, parallel group trial

Abstract

Background

Surgical resection is considered the most effective therapeutic method for pulmonary aspergilloma (PA), while the necessity of postoperative antifungal therapy remains controversial. Previous studies have several shortcomings such as retrospective design and inadequate follow-up durations. In this study, we aim to address these evidence gaps by evaluating the effects of postoperative antifungal therapy on both Aspergillus infection recurrence and drug-related side effects.

Methods

We conducted a prospective, randomized, controlled, single-center clinical trial. Eligible patients with PA who underwent surgical resection at our institute were enrolled and randomly assigned to receive either a 3-month course of voriconazole (VRZ) tablet postoperatively or symptomatic treatment alone. Participants were evaluated for Aspergillus infection recurrence and drug-related side effects every 3 months for 2 years. Ultimately, 36 patients in the VRZ group and 14 patients in the control group were included in the analysis, with balanced baseline characteristics between groups.

Results

After a median follow-up of 27 months, no recurrence of Aspergillus infection was observed in either group. Symptomatic improvements (e.g., hemoptysis and white sputum) were comparable between groups. Although serum IgG level significantly increased by 2.22 g/L after VRZ treatment (p < 0.001), other parameters indicating immune activity (CD4⁺%, CD8⁺%, CD4⁺/CD8⁺ ratio, IgA, IgM and Aspergillus-specific IgG) and pulmonary function showed no significant intergroup differences at the final follow-up. Regarding safety, VRZ intervention was associated with a significant increase in aspartate aminotransferase level (20.9 ± 6.2 U/L to 26.5 ± 7.8 U/L, p = 0.023). Hypokalemia incidence was numerically higher in the VRZ group (21.4% vs. 0, p = 0.530), though not statistically significant.

Conclusions

Our findings demonstrated that postoperative antifungal therapy did not reduce Aspergillus recurrence rate but posed potential safety risks. For immunocompetent patients undergoing complete resection without intraoperative fungal spillage, routine postoperative antifungal agents are not recommended.

Trial registration

Chinese Clinical Trial Registry Number: ChiCTR1800019990, registration date: December 11th, 2018.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-025-11267-w.

Introduction

Chronic pulmonary aspergillosis (CPA) is a severe respiratory disease caused by Aspergillus spp., currently divided into five subtypes: simple aspergilloma, chronic cavitary pulmonary aspergillosis (CCPA), chronic fibrosing pulmonary aspergillosis (CFPA), Aspergillus nodule and chronic necrotizing pulmonary aspergillosis (CNPA) [1]. Among the various subtypes, pulmonary aspergilloma (PA) represents the most common manifestation, characterized by the development of Aspergillus balls within pre-existing pulmonary cavities, typically occurring in immunocompetent hosts [1]. Since the original classification by Belcher and Plummer in 1960, PA has been categorized into either simple PA or complex PA (synonymous with CCPA) [1, 2]. Simple PA usually localized to thin-walled lung cavities without surrounding lung parenchyma involvement. Complex PA is characterized by thick-walled cavities with invasion of adjacent lung tissue and pleura. These cavities typically arise secondary to underlying pulmonary pathologies, such as tuberculosis (TB), bronchiectasis, chronic obstructive pulmonary disease (COPD), necrotizing pneumonia, sarcoidosis, or lung cancer [3]. The fungal ball itself consists of Aspergillus hyphae (primarily Aspergillus fumigatus, Af), conidia, mucus and necrotic debris [4]. PA represents the end-stage manifestation of CPA and is often asymptomatic initially, delayed diagnosis can lead to severe complications, including malnutrition, life-threatening hemoptysis and progressive lung destruction. According to estimations conducted by Denning, PA accounts for 86.3% of CPA cases, with an estimated annual incidence exceeding 200 thousand globally [5, 6]. Mortality rates reach 23% for complex PA and 11% for simple PA [7].

Antifungal therapy demonstrates limited efficacy due to the avascular nature of cavities and fungal balls, compounded by drug-related toxicity [8]. Despite bronchial artery embolization (BAE) may temporize hemoptysis, 50% of patients experience recurrence, relegating BAE to a bridge to definitive surgery [9]. For symptomatic patients with adequate pulmonary reserve, especially those appeared massive or recurrent hemoptysis, surgical resection is a suggested therapeutic option [1]. Advances in surgical techniques have substantially reduced postoperative mortality and complications [10]. Nevertheless, recurrence rates range from 5 to 41%, with complex PA posing a higher risk than simple PA [11–13]. Relapse may originate from intraoperative dissemination of fungal material into adjacent tissues, the pleural cavity, or bloodstream. Despite this, the role and duration of postoperative antifungal therapy remain contentious and data on drug safety are scarce [14]. Existing studies are predominantly retrospective, with conflicting conclusions [1, 12, 13, 15–17]. To address this gap, we conducted a prospective, randomized, controlled trial to assess postoperative voriconazole’s effects in preventing Aspergillus recurrence and associated adverse events. Comprehensive assessments were performed to evaluate antifungal therapeutic efficacy across three domains: symptomatic manifestation, immune function (both cellular and humoral immunity markers) and pulmonary function. This study provides robust clinical evidence to inform therapeutic decision-making and guide the development of evidence-based guidelines for postoperative management of PA. In addition, our follow-up protocol established in this study may be advisable for PA patients.

Methods

Trial design

This prospective, randomized, controlled, single-center, 24-month, open-label, parallel group trial was conducted in Shanghai Pulmonary Hospital. The target population comprised patients with suspected PA who underwent surgical resection at our institute between January 2020 and September 2024. Final enrollment was restricted to immunocompetent subjects with PA confirmed by imaging findings and histopathological analysis of surgically resected lung tissue (Fig. 1). Patients’ baseline characteristics, defined as conditions after surgery but prior to intervention, were systematically monitored and documented. Subsequently, participants were randomly allocated to the voriconazole (VRZ) group or control group via simple randomization method. During the first three months after surgery, patients in the VRZ group received VRZ tablets (Vfend^®) 0.2 g, twice daily, while those in the control group receive no antifungal therapy simultaneously. Both groups were permitted symptomatic treatments (e.g., antitussives, expectorants) as clinically indicated. Participants were assessed for Aspergillus infection recurrence quarterly for 24 months after surgery (Fig. 2A and B). Both hospital and telephone follow-up were permitted, with the former strongly recommended when feasible. For patients undergoing in-hospital follow-up, comprehensive assessments including symptoms, biomarkers for fungus infection, inflammatory parameters, cytokine levels, immunity status, biochemical indicators (liver function, renal function and electrolyte levels), electrocardiograph (ECG), pulmonary function and chest computed tomography (CT) were evaluated. For telephone follow-ups, structured interviews focused on recurrence-related symptoms were conducted. Detailed information for follow-up protocol is presented in Additional file Appendix 1.

Fig. 1.

CT scan of pulmonary aspergilloma (PA). A Simple PA in a 36-year-old woman, “air-crescent” sign was observed in a left lower lobe thin-walled cyst, which appeared as a solid oval mass surrounded by a crescent of air. B Complex PA in a 30-year-old man, two fungus ball-like shadows were seen in left upper lobe and right lower lobe, respectively. Strands of Aspergillus were confirmed by pathologic examination of excised lung tissue in both cases

Fig. 2.

Research workflow. A Patients were recruited and randomly assigned to voriconazole group and control group. Oral voriconazole together with usual treatment were prescribed for the voriconazole group, while control group merely received usual treatment. B During follow-up period, series of examinations were performed in participants every 3 month for 2 years. C The actual times of hospital follow-up and follow-up duration from surgery to last telephone contact were listed, respectively

Sample size

According to the published experiments, the recurrence rate of VRZ group is expected to be less than 1%, and that of observation group is about 23% [11, 14, 18]. The sample size was calculated using PASS 15 software with two sided α = 0.10 and β = 0.20, which suggested a minimum of 40 patients in total (20 per group). In consideration of approximately 10% missing rate, 44 patients are planned to be included totally.

Participants

Eligible participants were adults meeting all of the following criteria. (1) Patients received surgical resection at Shanghai Pulmonary Hospital with a confirmed diagnosis of PA, which required the appearance of fungal ball in pulmonary cavity on CT and histopathological evidence for Aspergillus infection in resected lung tissue. (2) Patients were immunocompetent at baseline, defined as neutrophil count > 1.5 × 10⁹/L and CD4⁺ T lymphocyte count > 500 cells/µL [19]. (3) Written informed consent was provided. Patients were excluded if diagnosed as malignant tumor or severe organ dysfunction. Detailed exclusion criteria, along with study withdrawal criteria and classification criteria for PA are provided in Additional file Appendix 2.

Endpoint

The primary endpoint was Aspergillus infection recurrence, required combination of following three characteristics (meeting any single criterion alone did not qualify). (1) Clinical deterioration: progressive respiratory or systemic symptoms, including worsen cough, increased sputum production, recurrent or new-onset hemoptysis. (2) Objective evidence of Aspergillus infection (≥ 1 of the following criteria): a). radiological progression on chest imaging, such as enlarging pulmonary consolidation, development of cavitary lesions; b). serological evidence, including positive serum galactomannan (GM) test, elevated Aspergillus-specific IgG (sIgG) level; c). microbiological confirmation: positive culture for Aspergillus species from lower respiratory specimens. (3) Exclusion of alternative diseases: bacterial pulmonary infection, active tuberculosis, or non-tuberculous mycobacterial infection.

The safety endpoints were side effects of VRZ such as drug-induced liver injury (DILI), kidney associated side effects and cardiac arrhythmia. DILI was defined as Alanine aminotransferase (ALT) > 5 upper limit of normal (ULN), or ALT > 3 ULN together with total bilirubin (TBil) > 2 ULN [20]. Kidney associated side effects included direct toxicity (serum creatinine [Scr] or blood urea increasing to 1.5 times baseline) and indirect toxicity (hypokalemia and hyponatremia) [21]. Cardiac arrhythmia was characterized as QTc interval prolongation: ΔQTc > 60 ms from baseline or absolute QTc > 500 ms attributed to VRZ administration [22].

Missing data

Despite complete telephone follow-up, some examinations (e.g., spirometry and ECG) were partially missing in part of patients due to real factors (e.g., coronavirus disease 2019 [COVID-19] pandemic-related access barriers). Therefore, partial data were unavailable. We decided to omit these data. Final analyses included only available case data without imputation.

Statistical analysis

We employed an intention-to-treat (ITT) approach, analyzing all randomly assigned participants who completed at least one follow-up assessment. Statistical analysis was conducted with SPSS 29 (IBM Corp.). Graphical presentations were created using GraphPad Prism 10. Continuous variables were expressed as mean ± standard deviation (SD) or median (interquartile range, IQR), categorical variables were presented as numbers (percentages, %). Normal distribution was assessed by means of Shapiro-Wilk tests. Normally distributed continuous variables were compared using independent Student’s t-test (unpaired) or paired t-test (matched). If non-normally distributed, Mann-Whitney U test (unpaired) or Wilcoxon signed-rank test (matched) was performed. As for categorical variable comparisons, we used χ² test or Fisher’s exact test (unpaired) and McNemar’s test (paired, for baseline vs. follow-up comparisons). Logistic regression was performed to identify factors associated with clinical features. Odds ratio (OR) with corresponding 95% confidence interval (CI) were presented. A two-tailed p < 0.05 was considered statistically significant.

Results

Characteristics at baseline

The final cohort comprised 50 immunocompetent patients (mean age: 47.9 ± 14.6 years, male predominance 62.0%, Fig. 3A) with histologically confirmed PA. The median follow-up duration was 27 months (IQR: 17.8–39.3, counted from the date of surgery to last telephone follow-up), with 56.0% (n = 28) of participants completing the scheduled 2-year follow-up and only 10.0% (n = 5) followed for less than one year. 74.0% (n = 37) attended at least one hospital visit for comprehensive evaluation, 26.0% (n = 13) relied solely on telephone follow-up, with no study discontinuations due to adverse events (Fig. 2C).

Fig. 3.

Baseline characteristics and symptom changes. A Distributions of age at baseline. B Distribution of pulmonary aspergilloma. C Symptom profile at baseline and last telephone follow-up. Con, control group; VRZ, voriconazole

64.0% (n = 32) were complex PA and 36.0% (n = 18) were simple PA, with predominant upper lobe involvement (71.9%, Fig. 3B). The most common clinical manifestations included hemoptysis (68.0%) and cough (68.0%), while white sputum was observed in 20.0% of cases. Predominant underlying pulmonary diseases were bronchiectasis (48.0%) and tuberculosis (30.0%). Comparative analysis demonstrated significant immunological and pulmonary function differences between disease subtypes: simple PA patients exhibited higher CD8⁺% (24.40 ± 10.99% vs. 17.78 ± 5.32%, p = 0.016), superior pulmonary function (FVC% predicted 93.8 ± 22.6% vs. 80.0 ± 17.8%, p = 0.011; FEV1% predicted 92.9 ± 25.6% vs. 80.5 ± 20.6%, p = 0.035) and reduced hemoptysis risk (OR 0.28, 95% CI 0.08–0.975, p = 0.045).

Randomization yielded two treatment groups (VRZ n = 36, control n = 14) with comparable baseline characteristics, as presented in Table 1. The VRZ and control groups showed similar distributions of PA complexity (69.4% vs. 50.0% complex PA, p = 0.325) and aspergilloma size (median 2.5 cm vs. 2.3 cm, p = 0.795). Inflammatory markers including C-reactive protein (CRP), procalcitonin (PCT), erythrocyte sedimentation rate (ESR) and serum amyloid A (SAA) were well-balanced between groups (all p > 0.05). Though the VRZ group demonstrated significantly higher baseline interleukin-17 (IL-17) levels (6.37 ± 8.25 pg/mL vs. 1.28 ± 1.19 pg/mL, p = 0.041), other cytokines including IL-1β, IL-2, IL-6, interferon-α (IFN-α), IFN-γ and tumor necrosis factor-α (TNF-α) showed no significant differences (all p > 0.05, see Additional file Appendix 3).

Table 1.

Baseline demographics, clinical features and surgery procedures

Variables	All (n = 50)	VRZ group (n = 36)	Control group (n = 14)	p-value
Age, mean (SD), y.o.	47.9 (14.6)	49.5 (14.9)	43.7 (13.1)	0.210
Ethnicity, n (%)
Chinese Han	50 (100.0%)	36(100.0%)	14(100.0%)	1.000
Sex, n (%)				0.750
Male	31 (62.0%)	23 (63.9%)	8 (57.1%)
Female	19 (38.0%)	13 (36.1%)	6 (42.9%)
BMI, mean (SD)	22.3 (3.5)	22.0 (3.1)	23.1 (4.4)	0.341
Complex PA, n (%)	32 (64.0%)	25 (69.4%)	7 (50.0%)	0.325
Median sizea (IQR), cm	2.5 (1.9, 3.6)	2.5 (1.8, 3.6)	2.3 (1.9, 3.6)	0.795
Symptoms, n (%)
White sputum	10 (20.0%)	8 (22.2%)	2 (14.3%)	0.704
Cough	34 (68.0%)	25 (69.4%)	9 (64.3%)	0.746
Hemoptysis	34 (68.0%)	25 (69.4%)	9 (64.3%)	0.746
History of BAE, n (%)				0.215
Yes	28 (56.0%)	18 (50.0%)	10 (71.4%)
No	22 (44.0%)	18 (50.0%)	4 (28.6%)
Underlying lung disease, n (%)
Bronchiectasis	24 (48.0%)	20 (55.5%)	4 (28.6%)	0.119
Tuberculosis	15 (30.0%)	15 (41.7%)	0	0.004*
Surgical procedure, n (%)				0.167
VATS	44 (88.0%)	30 (83.3%)	14 (100.0%)
Thoracotomy	6 (12.0%)	6 (16.7%)	0

BAE Bronchial artery embolization, BMI Body mass index, IQR Interquartile range, PA Pulmonary aspergilloma, SD Standard deviation, VATS Video-assisted thoracoscopic surgery

^aSize of PA refers to maximal diameter of fungal ball measured on axial chest CT images in the lung window setting. ^*p<0.05

Surgical approaches and postoperative complications

The majority of patients (88.0%, n = 44) underwent video-assisted thoracoscopic surgery (VATS), while thoracotomy was performed in the remaining 12.0% (n = 6). Lobectomy constituted the most common surgical approach (78.0%, n = 39), followed by sublobectomy (12.0%, n = 6) and bilobectomy (8.0%, n = 4). All patients received complete resection of aspergillomas, and spillage of fungal materials into adjacent tissue was carefully avoided by surgeons during surgery. No intraoperative or postoperative 30-day death was observed. Postoperative complications occurred in 32.0% (n = 16) of patients, with no significant difference observed between the VRZ and control group (p > 0.05). The most frequently reported complications included chronic cough (12.0%, n = 6), empyema (6.0%, n = 3) and hydrothorax (4.0%, n = 2). Other complications were similarly distributed across both treatment arms, with detailed outcomes summarized in Additional file Appendix 3.

Aspergillus infection recurrence

During the follow-up period, no confirmed cases of Aspergillus infection recurrence were observed in either the VRZ or control group. Notably, one patient in the control group exhibited a positive serum GM test at the 1-year postoperative evaluation, without corresponding radiographic abnormalities or clinical symptoms suggestive of aspergillosis recurrence. This false-positive result was potentially attributable to concurrent piperacillin-tazobactam (PTZ) therapy, as PTZ may contain GM cross-reactive epitopes derived from its Penicillium-derived production process, a documented cause of GM assay interference [23]. Additionally, two patients (one per treatment group) were diagnosed as allergic bronchopulmonary aspergillosis (ABPA), evidenced by recurrent respiratory symptoms and elevated Aspergillus-specific IgE (69.6 kUA/L in the VRZ group) or total IgE (2110 IU/mL in the control group), with absence of microbiological (negative cultures and GM tests) or radiographic evidence of active infection. As ABPA represents a hypersensitivity response rather than invasive infection, these cases weren’t regarded as Aspergillus infection recurrence [24].

Comparison of clinical features

Symptoms

Key symptoms suggestive of Aspergillus infection recurrence—including cough, hemoptysis and white sputum—were rigorously monitored throughout the mean 28-month study period. The cumulative incidence of intermittent cough was comparable between groups (VRZ: 13.9% [5/36] vs. control: 21.4% [3/14], p = 0.670). Fisher’s exact test revealed the improvement of cough was similar in two groups (p = 0.649). Clinical documentation also confirmed balanced amelioration in cough severity, including reduced frequency and diminished impact on quality of life. During follow-up, hemoptysis recurrence occurred in only one patient (2.8%) in the VRZ group, while no cases were observed in the control group (Fig. 3C). Comparative analysis revealed no significant difference in hemoptysis reduction between groups (p = 1.000). The prevalence of white sputum production remained stable throughout the study period, with comparable proportions observed between the VRZ group (22.2%, 8/36) and control group (21.4%, 3/14) at the final follow-up (p = 1.000). Complete symptom profiles are detailed in Additional file Appendix 4. These findings indicate that postoperative VRZ prophylaxis did not significantly alter the risk of refractory cough and hemorrhage after surgery.

Immune function

T cell-mediated immunity, particularly through CD4⁺ Type 1 T helper (Th1) cell and Th17 cell responses, is essential for controlling invasive aspergillosis by orchestrating macrophage activation and neutrophil recruitment to infected tissues. CD4⁺ T cell percentage (CD4⁺%), CD8⁺ T percentage (CD8⁺%) and CD4⁺/CD8⁺ ratio (CD4⁺/CD8⁺) were measured as markers of cell-mediated immunity. These parameters were available for 38 patients at baseline and 31 patients at the final follow-up. Comparisons between the VRZ group and the control group revealed balanced baseline levels of CD4⁺% (p = 0.883), CD8⁺% (p = 0.432) and CD4⁺/CD8⁺ (p = 0.204, see Additional file Appendix 5). Similarly, no significant differences were observed in CD4⁺% (33.93 ± 7.31% vs. 36.48 ± 5.93%, p = 0.331), CD8⁺% (19.10 ± 5.03% vs. 19.47 ± 4.65%, p = 0.870) and CD4⁺/CD8⁺ (1.96 ± 0.84 vs. 1.94 ± 0.34, p = 0.535) between the VRZ group and the control group at the final hospital follow-up. Paired analysis of the VRZ group demonstrated stable cellular immunity parameters throughout the treatment period, with no significant changes observed in CD4⁺% (p = 0.883), CD8⁺% (p = 0.057), or CD4⁺/CD8⁺ (p = 0.340) between baseline and final follow-up (Fig. 4A).

Fig. 4.

Immune function and spirometry results at baseline and last follow-up. A Cellular immune function. B Humoral immune function. C Ventilatory function parameters. D Ventilatory function (FVC% pred and FEV1/FVC% pred) and diffusing capacity (DLCO% pred). Con, control group; DLCO, diffusing capacity of the lung for carbon monoxide, measured by single breath method; FEV1, forced expiratory volume in 1 s; FEV1% pred means the percentage of FEV1 in the predicted value, the same below; FEV1/FVC% pred means the percentage of FEV1/FVC% in the predicted value; FVC, forced vital capacity; VRZ, voriconazole group. ns: p > 0.05; *: p < 0.05; ***: p < 0.001

B cell-mediated humoral responses generate antifungal antibodies, their direct fungicidal capacity is limited compared to cellular immunity. Levels of IgG, IgA and IgM were measured to assessed humoral immune activity. Baseline comparison between groups showed no significant differences in these parameters (Additional file Appendix 5). At last follow-up, IgA level in VRZ group was significantly higher than that in control group (2.93 ± 1.42 g/L vs. 1.95 ± 0.68 g/L, p = 0.040). In contrast, IgG and IgM levels remained comparable between groups (Fig. 4B). Longitudinal analysis revealed VRZ treatment induced a clinically significant 2.22 g/L increase in mean IgG level (p < 0.001), while both IgA (+ 0.75 g/L, p = 0.089) and IgM (+ 0.38 g/L, p = 0.053) showed modest but statistically non-significant elevations. Though baseline sIgG levels were unavailable for the control group, following measurements demonstrated comparable median sIgG levels between the VRZ (79.16 AU/mL, IQR: 68.30–168.90, n = 15) and control group (77.34 AU/mL, IQR: 34.59–94.32, n = 6; p = 0.302). In summary, VRZ therapy appeared to promote systemic humoral immune recovery, as evidenced by elevated immunoglobulin levels. Comparable sIgG levels between groups suggested surgical clearance of Aspergillus antigens limited further Aspergillus-specific antibody production mediated by VRZ.

Pulmonary function

Spirometry was performed to evaluate postoperative recovery and assess the therapeutic effects of antifungal treatment. Baseline testing was completed in 49 patients, but only five underwent follow-up evaluations. Initial spirometry revealed comparable lung function parameters between the VRZ group and the control group, including forced vital capacity (FVC, p = 0.179) and forced expiratory volume in one second (FEV1, p = 0.094) and FEV1/FVC% (p = 0.108, Additional file Appendix 6). FVC% pred (p = 0.167), FEV1/FVC% pred (p = 0.332) and DLCO% pred (p = 0.656) at baseline were also balanced. These parameters remained similar at final follow-up, though limited by substantial attrition (Fig. 4C and D). At baseline, ventilatory dysfunction (VD) was prevalent (50.0%, n = 25), comprising restrictive (p = 0.744 between groups) and obstructive (p = 0.211 between groups) patterns. Diffusion impairment affected 16.0% (n = 8, p = 0.656 between groups). The small follow-up cohort precluded analysis of longitudinal changes in these functional abnormalities, highlighting a critical need for enhancement in postoperative monitoring to evaluate respiratory recovery after intervention.

Safety and tolerability

Though VRZ is highly recommended for aspergillosis treatment and prophylaxis, its clinical utility is constrained by some frequent adverse effects, particularly hepatotoxicity, renal impairment and cardiac complications [25]. In this study, we monitored hepatic function, renal function and QTc interval following VRZ use to evaluate its safety.

Hepatotoxicity

Hepatic impairment is a common and dose-dependent side effect of VRZ. Comprehensive monitoring revealed no instances of jaundice or other clinical manifestations suggestive of hepatotoxicity within six months of starting antifungal treatment. Overall, no cases met the diagnostic criteria for DILI.

At baseline, liver function test results were available in all participants. ALT, aspartate aminotransferase (AST) and TBil levels were balanced between groups (see Additional file Appendix 7). After 3-month VRZ treatment, evaluations (25 VRZ vs. 11 control patients) demonstrated no significant intergroup differences in ALT (21.4 ± 10.3 U/L vs. 21.7 ± 9.2 U/L, p = 0.593), AST (26.5 ± 7.8 U/L vs. 23.9 ± 9.0 U/L, p = 0.570) and TBil (11.9 ± 4.4 µmol/L vs. 10.7 ± 3.6 µmol/L, p = 0.362). Longitudinal analysis in the VRZ group showed a significant 5.6 U/L increase in mean AST level (p = 0.023), while ALT (p = 0.400) and TBil (p = 0.272) levels remained stable (Fig. 5A).

Fig. 5.

Side effect profiles of VRZ therapy. A Hepatotoxicity. B Cardiac safety. C Renal safety profile. D Electrolyte levels. Con, control group; ALT, alanine aminotransferase; AST, aspartate aminotransferase; Scr, serum creatinine; TBil, Total bilirubin; VRZ, voriconazole group. ns: p > 0.05, *: p < 0.05

The prevalence of liver function abnormalities elevated in the VRZ group, with aminotransferase abnormality (> 40 U/L) increasing from 5.6 to 12.0% (n = 2/36 to n = 3/25, p = 0.392) and TBil abnormality (> 17.1 µmol/L) rose from 13.9 to 16.0% (n = 5/36 to n = 4/25, p = 0.675). In comparison, the control group showed complete resolution of aminotransferase abnormality (14.3% [2/14] to 0% [0/11]; p = 0.487) and non-significant increase in hyperbilirubinemia (7.1% [1/14] to 9.1% [1/11]; p = 0.821). Baseline liver abnormalities likely resulted from surgical stress (e.g., ischemia, inflammation). It seems that VRZ modestly increased hepatic insufficiency because untreated patients demonstrated spontaneous normalization of liver enzymes.

Cardiac arrhythmia

A known cardiac effect of VRZ is QTc interval prolongation on ECG—potentially triggering torsades de pointes [26]. Baseline QTc intervals were balanced between groups (31 VRZ vs. 11 control, 404.4 ± 32.7 ms vs. 387.7 ± 15.6 ms, p = 0.070, see Additional file Appendix 7). Post-treatment ECG evaluations were available in a limited subset (6 VRZ vs. 3 control), revealing comparable QTc intervals (406.2 ± 29.0 ms vs. 397.8 ± 20.6 ms, p = 0.668). Longitudinal analysis revealed no significant alteration in mean QTc interval duration following VRZ therapy (p = 0.999), as was shown in Fig. 5B. The VRZ group showed a mean QTc shortening of 20.5 ± 16.3 ms, suggesting an absence of clinically meaningful prolongation.

Kidney associated side effects

The correlation between intravenous VRZ administration and renal function deterioration has been demonstrated by several studies [27]. Nevertheless, the nephrotoxic potential of oral voriconazole may be clinically underrecognized. We evaluated renal function and electrolyte balance via measuring Scr, blood urea, potassium and sodium levels. Baseline assessment revealed comparable renal and electrolyte profiles between groups (all p > 0.05, see Additional file Appendix 7). Follow-up analysis at the completion of VRZ treatment demonstrated maintained equilibrium between the VRZ and the control group, with Scr (69.0 ± 15.6 µmol/L vs. 64.8 ± 17.1 µmol/L, p = 0.524), urea (5.3 ± 1.6 mmol/L vs. 5.6 ± 1.5 mmol/L, p = 0.695), potassium (4.0 ± 0.5 mmol/L vs. 4.0 ± 0.2 mmol/L, p = 0.961) and sodium levels (139.5 ± 2.7 mmol/L vs. 140.2 ± 4.0 mmol/L, p = 0.726) showing no significant differences. Longitudinal analysis within the VRZ group confirmed the stability of these parameters, with paired comparisons between baseline and treatment cessation showing no statistically significant variations (all p > 0.05, Fig. 5C and D).

At baseline, hypokalemia presented in 8.3% (3/36) of VRZ-treated patients compared to 21.4% (3/14) of controls (p = 0.331), suggesting no significant initial imbalance between groups. Baseline hypokalemia may result from chronic malnutrition and acute postoperative anorexia. Following the 3-month treatment course, whereas all control patients achieved normal potassium levels, 21.4% (3/14) of VRZ-treated patients remained hypokalemia (p = 0.535 vs. controls). Notably, the single case of baseline hyponatremia (2.8%) in the VRZ group resolved completely with treatment. To sum up, VRZ therapy was associated with higher risk of hypokalemia.

Discussion

The optimal treatment strategy after surgical intervention for PA remains extensively controversial. Some studies advised that postoperative antifungals are unnecessary [15], while others suggested protective effects for the recurrence of Aspergillus infection [12, 13], recent studies are listed in Table 2. In addition, few previous studies are designed prospectively and randomly, suggesting the paucity of high-quality evidence regarding this issue [17]. To the best of our knowledge, this is the first prospective, controlled, cohort study to evaluate the effects of postoperative antifungal therapy on recurrence of Aspergillus infection in patients with PA. In this trial, we innovatively assessed the quantitative alterations after antifungal treatment from various perspectives including symptoms, host immunity status and pulmonary function. Our results confirm that postoperative antifungal treatment had no apparent benefit for the prophylaxis of Aspergillus infection recurrence, which was similar to the conclusion of series retrospective study [15]. Moreover, we formulated a postoperative follow up protocol which may serve as a transferable framework for other tertiary centers managing surgically treated PA cases [28].

Table 2.

Summary of studies on postoperative AF pharmacotherapy for preventing Aspergillus infection recurrence in PA patients

Reference	Surgical procedure	Antifungal regimen	Sample size	Results	Conclusions	Study design	Shortcomings
Gebitekin et al. [16]	Single stage cavernostomy and myoplasty	Perioperative (2w prior to surgery and 3 m postoperative), oral ICZ 200 mg twice daily.	7	Six were alive and had a mean hemoptysis free duration of 57.2 m.	Perioperative AF agents were safe and reliable.	Prospective	No control group, small sample size
Sagan et al. [15]	Lobectomy, pneumonectomy and so on	Perioperative (30d before and 30 d after operation), iv. and oral AmB-D, ICZ and FCZ.	72	Postoperative morbidity and 10 -year cumulative survival had no significant alteration.	No benefit from adjuvant AF agents.	Retrospective	Not prospective study. AF regimens were inconsistent in participants.
Farid et al. [13]	Multiple procedures	If spillage of fungi occurred, perioperative (2w prior to surgery and 2 m postoperative, or longer if partial lesions retained), iv. VRZ.	30	Eight patients (26%) had recurrence of disease, most having prior CCPA (75%).	Postoperative AF therapy was recommended if intraoperative fungal spillage occurred or complete lesion resection could not be achieved	Retrospective	Not prospective study
Setianingrum et al. [12]	Multiple procedures	Preoperative, perioperative, or postoperative, drugs unknown.	61	AF therapy before surgery or both before and after surgery were protective for relapse.	Postoperative AF agents didn’t reduce relapse.	Retrospective	Not prospective study

AF Antifungal, AmB-D Amphotericin B deoxycholate, CCPA Chronic cavitary pulmonary aspergilloma, d day, FCZ Fluconazole; ICZ, itraconazole, iv. intravenous, m month, PA Pulmonary aspergilloma, w week, VRZ Voriconazole

In patients with PA, surgery is highly recommended for suitable patients to eradicate lesions [29]. Published studies reported perioperative mortality rates of 0.9–4.4% and complication rates of 23.6–40.3% for PA resection, which strongly influenced by patient selection and proper perioperative management. For example, case with complex PA easily experience a worse outcome than simple type. In previous studies, usual causes of death include recurrent massive hemoptysis and pneumonia after surgery. Relapse rate for aspergillosis after surgery varies in different studies, such as 5%, 26%, 41% [11–13]. Intraoperative spillage of the fungal material and retention of involved lung maybe the principal cause. In our trial, no one died in perioperative duration. Particularly, no one experienced recurrence postoperatively.

The discrepancy between other findings and our trial may result from evolving surgical techniques and medical equipment over the past decade. The development and widely application of VATS lobectomy in selected patients with aspergillomas might help to reduce the possibility of recurrence [29]. On the one hand, VATS is associated with relatively lower risk of Aspergillus spillage and surgical complication, as well as shorter hospital stay, compared with open thoracotomy approach [30, 31]. On the other hand, lobectomy is more likely to achieve complete removal of lesion than sublobectomy (including segmentectomy and wedge resection). In this study, complete resection of aspergillomas with no detectable fungal effusion was accomplished in all participants. Additionally, with the enhancement of health consciousness in public after the COVID-19 pandemic, patients with PA could be detected in earlier stage than before and achieve satisfactory operation effect. Furthermore, standardized treatment for pulmonary tuberculosis enables early disease control, thereby preventing the formation of cavity and subsequent Aspergillus ball in tuberculous focus. Additionally, we did not include patients who showed evidence of immunosuppression, which maybe another reason for our relatively good outcome.

In this cohort, key symptoms including cough and hemoptysis achieved similar amelioration between the VRZ group and control group during follow-up. This finding indicates that surgical intervention itself may play a dominant role in preventing symptomatic recurrence, with postoperative antifungal therapy providing limited additional benefit. Additionally, persistent white sputum suggests that sputum characteristics may reflect underlying comorbidities such as COPD rather than active infection, and VRZ prophylaxis did not significantly alter mucoid secretion patterns after surgery. The stability of this symptom across both groups implies that white sputum alone may have limited value as a standalone marker for Aspergillus recurrence in postoperative monitoring.

The host cell-mediated immune capacity was evaluated through assessment of CD4⁺%, CD8⁺% and CD4⁺/CD8⁺. These parameters collectively reflect CD4⁺ T cell-driven cytokine activation of phagocytes and CD8⁺ T cell-directed cytotoxic elimination of infected host cells, which are critical for effective clearance of fungal pathogens. Following recognition and phagocytosis towards Aspergillus by host, immune cascades are triggered, with distinct effector mechanisms targeting different fungal morphotypes: macrophages primarily mediate conidial killing, while neutrophils are crucial for hyphal destruction [32]. This coordinated antifungal response depends critically on cytokine networks orchestrated by various T cell subsets. For instance, TNF-α, granulocyte-macrophage colony-stimulating factor (GM-CSF) and IFN-γ generated by Th1 cell and CD8⁺ T cell effectively stimulate macrophage, enhancing its phagocytosis capacity and reactive oxygen species (ROS) releasing [33]. Th17 cell is characterized by the production of IL-17 and IL-22, inducing neutrophil chemotaxis towards infection sites and stimulating airway epithelial cells produce antimicrobial peptides [34]. CD8⁺ T cell, also known as cytotoxic T lymphocyte (CTL), exhibits direct fungicidal activity via releasing perforin and granzyme. Therefore, reduction in CD4⁺%, CD8⁺% and CD4⁺/CD8⁺ ratio suggests immune suppression and is associated with higher risk for opportunistic infection [35]. We found patients with complex type of PA had lower proportion of CD8⁺ T cell than patients with simple type in the trial, in line with prior findings, suggesting CD8⁺ T cell is a promising tool in predicting severity of PA [36]. Our analysis revealed that levels of CD4⁺%, CD8⁺% and CD4⁺/CD8⁺ in VRZ group had little change after antifungals intervention, which indicates postoperative antifungal therapy had little influence on cellular immunity function. Similarity in T cell-mediated immune activity between the VRZ group and control group was in accordant with the parallel recurrence rate of Aspergillus infection in two groups, which implies promotion T cell responses probably in favor of Aspergillus elimination and remission of pulmonary aspergillosis. With advances in vaccinology, eliciting stronger responses by new vaccines may be a promising approach to overcome relapse of aspergillosis [37].

The role of humoral immunity in fungal defense is complicated [38]. IgG, IgA and IgM protect host against fungus by enhancing complement activation, opsonizing phagocytosis and antibody-dependent cellular cytotoxicity [39]. Notably, Af-binding IgG is essential to neutrophil-mediated phagocytosis in vivo [40]. Secretory IgA (SIgA) is the major class of immunoglobin in respiratory mucosal immune system, mediating pathogen neutralization and elimination [41]. However, thick cell wall of A. fumigatus prevents the formation of the membrane attack complex (MAC). Masking of pathogen-associated molecular pattern (PAMP, like β-glucan) in Aspergillus surface also prevent antibodies binding [38]. In this trial, IgG, IgA and IgM levels elevated in different degrees after VRZ treatment, suggesting selective B-cell activation. The marked IgG enhancement potentially reflects improved opsonizing capacity against Aspergillus, while the borderline increases in IgA and IgM may represent enhancing mucosal immunity or early-phase antibody production. These responses imply the immunomodulatory potential of VRZ in the postoperative setting. According to Simitsopoulou et al., VRZ was demonstrated to up-regulate IFN-γ and TNF-α expression in human monocytic cells exposed to Aspergillus hyphal fragment [42]. These Th1 pro-inflammatory cytokines probably enhance T follicular helper cell (Tfh cell) functions, thereby promoting the differentiation of B cells into plasma cells and the secretion of IgG. Further investigations are needed to confirm these immunological effects and their clinical correlations. If VRZ is proven to increase serum immunoglobulin, distinguish this pharmacologic effect from true Aspergillus infection recurrence requires comprehensive evaluation incorporating both clinical manifestations and radiological evidence. Isolated total immunoglobulin elevation carries lower diagnostic value for recurrence compared to concurrent clinical deterioration and progressive radiological abnormalities.

Standing as the first-line recommended antifungal agent for aspergillosis, VRZ is primarily metabolized by CYP2C19 enzyme in liver [43]. Due to a narrow therapeutic window and side-effect profile, therapeutic drug monitoring is recommended to prevent side effects of VRZ. Well-known adverse effects of VRZ include visual disturbances (19.0%), hypokalemia (13.0%), hepatotoxicity (12.4%), nausea (5.4%), rash (5.3%), hyponatremia (7.9%), cardiac arrhythmia and so on [44]. Particularly, intravenous VRZ demonstrates dose-dependent kidney injury effect due to the incorporation of cyclodextrin [45]. VRZ tablets are theoretically associated with lower renal risk due to their cyclodextrin-free composition. Yet, clinical data specifically evaluating the renal effects of oral VRZ remain limited. Our study provides novel evidence that VRZ tablet administration was associated with elevating AST level rather than bilirubin, indicates that VRZ tablets may be inclined to cause hepatocellular injury [46]. Even though no renal deficiency happened after VRZ intervention, higher incidence of hypokalemia was observed in the VRZ group. Apart from triggering adverse events, VRZ also leads to considerable cost in China. The usage of original brand has a significant higher cost over the generic drugs [47]. Stop unessential postoperative VRZ treatment not only avoids underlying side effects, undesirable drug interactions and the emergence of drug resistance, but also relieves financial challenges for patients.

There are several limitations in our trial. First, spirometry tests and ECG were available in a small number of patients during follow-up. Since nearly all the participants didn’t report severe discomfort, therefore, only small part of them were willing to conduct these examinations in follow-up visit. Second, final sample size was relatively small in our study, which resulted from strict inclusion criteria in the study. Unequal group allocation (36:14) results from our use of simple randomization without stratification, though post-hoc analysis confirmed baseline characteristic balance. Third, only oral VRZ treatment was evaluated in our trial. Other antifungal compounds weren’t tested due to VRZ is the first-line drug of aspergillosis and routinely available in clinical practice. Larger populations and stricter adherence to the follow-up protocol are thus required for future research. Despite these drawbacks, our study offers significant support for the development of clinical guidelines for aspergillosis.

Conclusions

Our trial reveals that postoperative antifungal therapy appears not to reduce Aspergillus infection recurrence but increase drug-related side effects. We believe that, for immunocompetent patients with PA, postoperative antifungal medication isn’t warranted when complete removal of pathological lung is implemented and no spillover of Aspergillus materials occurs during operation. Nonetheless, the decision must be made in consideration of the well perioperative nutrition status and successful recovery of patients.

Abbreviations

ABPA

Allergic bronchopulmonary aspergillosis

AF

Antifungal

ALT

Alanine aminotransferase

AmB-D

Amphotericin B deoxycholate

AST

Aspartate aminotransferase

BAE

Bronchial artery embolization

BMI

Body mass index

CCPA

Chronic cavitary pulmonary aspergilloma

CI

Confidence interval

Con

Control group

COPD

Chronic obstructive pulmonary disease

COVID-19

Coronavirus disease 2019

CPA

Chronic pulmonary aspergillosis

CRP

C-reactive protein

CT

Computed tomography

CTL

Cytotoxic T lymphocyte

DILI

Drug-induced liver injury

DLCO

Diffusing capacity of the lung for carbon monoxide

ECG

Electrocardiograph

ESR

Erythrocyte sedimentation rate

FEV1

Forced expiratory volume in 1 second

FCZ

Fluconazole

FVC

Forced vital capacity

GM

Galactomannan

GM-CSF

Granulocyte-macrophage colony-stimulating factor

ICZ

Itraconazole

IFN-α

Interferon-α

IL-17

Interleukin-17

IQR

Interquartile range

ITT

Intention-to-treat

MAC

Membrane attack complex

OR

Odds ratio

PA

Pulmonary aspergilloma

PAMP

Pathogen-associated molecular pattern

PCT

Procalcitonin

PTZ

Piperacillin-tazobactam

ROS

Reactive oxygen species

SAA

Serum amyloid A

Scr

Serum creatinine

SD

Standard deviation

sIgG

Aspergillus-specific IgG

SIgA

Secretory IgA

TB

Tuberculosis

TBil

Total bilirubin

Th1

Type 1 T helper

Th17

Type 17 T helper

TNF-α

Tumor necrosis factor-α

ULN

Upper limit of normal

VATS

Video-assisted thoracoscopic surgery

VRZ

Voriconazole

Authors’ contributions

S.L. designed and funded the research, S.L. and YZ recruited the participants, treated and followed up with the patients. J.X., H.C., J.Z., H.W. and D.Z. provided patients cases. J.Y. performed data collection, statistics analysis, drafted the manuscript and provided the Figures. C.L. and J.L. participated in the data analysis. J.X. and S.L. revised the manuscript. X.C. and B.W. assisted in data collection. All authors have read and approved the final version.

Funding

This trial was supported by grants from National Natural Science Foundation of China (No. 82170051), Shanghai Municipal Health Commission (No. 202140257), Clinical Research foundation of Shanghai Pulmonary Hospital (No. FKLY20025) and Tongji University (No. 2021YXAL08).

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author [SL] on reasonable request.

Declarations

Ethics approval and consent to participate

This trial was approved by the Ethics Committee of Shanghai Pulmonary Hospital, Tongji University (number: K20-261). The trial adhered to the Declaration of Helsinki and Good Clinical Practice (GCP) guidelines. All participants provided written informed consents.

Consent for publication

This study contains original, unpublished data and is not being submitted for publication elsewhere. Written informed consent was obtained from the patients for publication. A copy of the written consent is available for review.

Competing interests

The authors declare no competing interests.