Treatment outcome definitions in chronic pulmonary aspergillosis: a CPAnet consensus statement

To the Editor:

Chronic pulmonary aspergillosis (CPA) is an uncommon but dreaded complication of many respiratory diseases occurring in non- or mildly immunocompromised patients [1]. CPA affects approximately three million people worldwide, an estimation which is undoubtedly affected by under-reporting of this neglected disease [2]. The morbidity and mortality is high, with 5-year survival rates of 15–60% depending on comorbidities and age [3, 4]. The diagnosis of CPA is typically established on a combination of clinical, radiological and microbiological criteria all present for at least 3 months and the absence of an alternative diagnosis [1]. The most common form of CPA is chronic cavitary pulmonary aspergillosis (CCPA), besides other disease entities, including chronic fibrosing pulmonary aspergillosis (CFPA), Aspergillus nodule(s) and single aspergilloma. Subacute invasive pulmonary aspergillosis (SAIA), also called chronic necrotising pulmonary aspergillosis, usually occurs in moderately immunocompromised patients and comprises a more rapidly progressive clinical course (<3 months) [1].

Treatment decisions in CPA largely depend on the pulmonary and general symptoms, and any pulmonary function loss or radiographic progression, especially in patients with CCPA. Treatment options include mostly oral triazoles with fungicidal activity against Aspergillus spp., such as itraconazole, voriconazole, posaconazole and isavuconazole, of which the first two are the best documented agents in the context of CPA. Treatment duration is recommended for at least 6 months and follow-up on or off therapy is performed every 3 to 6 months, including clinical monitoring, Aspergillus serology and/or microbiology, chest radiographs and periodic computed tomography (CT) scans [1, 5, 6].

Current guidelines on diagnosis and management of CPA, established by the European Society for Clinical Microbiology and Infectious Diseases (ESCMID) in collaboration with the European Respiratory Society (ERS) and the European Confederation of Medical Mycology (ECMM), as well as those developed by the Infectious Diseases Society of America (IDSA), provide diagnostic criteria for CPA but no clear treatment outcome definitions [1, 2, 6]. This lack of standardisation of end-points parallels the limited number of available randomised controlled trials (RCTs) on antifungal treatment of CPA [7, 8], which, beside some larger retrospective and non-randomised prospective studies [9–11], leads to a scarcity of high-quality data on CPA treatment. Therefore, establishing consensus on treatment outcome definitions was considered as one of the top four research priorities of the Chronic Pulmonary Aspergillosis network (CPAnet), an international research collaboration established in 2017 and funded as a Clinical Research Collaboration (CRC) by the ERS in 2020 [2].

Following two CPAnet meetings linked to the ERS Congresses in Paris 2018 and Madrid 2019, a CPAnet initiative to develop a consensus statement on treatment outcome definitions for CPA was established. Well-recognised CPA experts consisting of pulmonologists, infectious disease specialists, microbiologists and radiologists, as well as a CPA patient advisory group supported by the European Lung Foundation, were invited to participate in the development of this consensus statement [2]. The methodological approach used has proven useful for other rare pulmonary infections, such as multidrug-resistant tuberculosis (MDR-TB) and non-tuberculous mycobacterial pulmonary disease (NTM-PD) [12, 13]. The process involved the following 4 steps.

Step 1: Forming a CPA expert panel.

Step 2: The coordinating authors (I. Page and H.J.F. Salzer) proposed 15 CPA treatment outcome categories and drafted one statement for each category based on their expertise and after review of the available literature. Based on the drafted statements, all co-authors were asked to provide alternative statements. All voting steps were managed by the coordinating authors using a prepared Microsoft Word file (Microsoft, Washington, USA) attached to e-mail. The coordinating authors counted all votes independently.

Step 3: Co-authors were asked to select one preferred statement among the original and the alternative statements. All co-authors were blinded to the votes. The statement that received the most votes within each outcome category was selected for inclusion in the manuscript.

Step 4: Finally, co-authors were asked to indicate their agreement, disagreement or whether they preferred to abstain from a decision.

To evaluate the clinical response the Respiratory Symptom Score (RSS) should be used, which is based on six items including cough, sputum production, dyspnoea, haemoptoic sputum, chest tightness and nocturnal awakening. For each symptom a simple visual analogical scale (VAS) (10-cm line) can be used, comparable to the commonly established VAS to measure pain in patients. Stability is defined by a score variation between −25 and +25% (equivalent to ±2.5 cm on the VAS), while improvement and deterioration is defined by a decrease or increase in the score greater than 25% (table 1) [10].

29 CPA experts from nine countries in the European region (Austria, Belgium, Denmark, France, Germany, Moldova, Serbia, Spain and UK), one country from the Eastern Mediterranean region (Pakistan), one country from the South-East Asian region (India), two countries from the Western Pacific region (Japan, South Korea), one country from the African region (Nigeria), and one country from the Americas (USA) contributed to the development of this consensus statement; 28 completed all 4 steps. A median of seven alternative statements (minimum 1; maximum 17) for a category was proposed by the co-authors during step 2. Table 1 summarises the final definitions, including the results of voting in step 3 and the support level achieved in step 4.

By proposing these consensus definitions on outcome parameters for the treatment of patients with CPA, we aim to provide guidance for the design of future clinical trials in CPA, which encompass a highly unmet need [2]. Particularly, RCTs comparing different antifungal treatment regimens are lacking, partly due to the infrequent occurrence in single centres and disease heterogeneity of CPA. The consensus statement is an essential framework for the evaluation of CPA treatment outcome measures, which should ideally be harmonised throughout clinical trial design.

The overall agreement level within the expert panel was quite high for most questions, with ultimate agreement ranging from 78–100%. The value of Aspergillus-specific serology in treatment response assessment was the most important matter of debate. Indeed, since most available Aspergillus immunoglobulin G detection kits were originally designed to detect A. fumigatus, these assays might have limitations in areas where non-fumigatus strains are epidemiologically important [14]. The wording of the question did not include quantitative changes on a linear or semi-logarithmic scale, limiting this conclusion. Treatment response was generally assessed on three levels: 1) clinical response assessed by the RSS comprised of six cardinal symptoms in CPA (cough, dyspnoea, sputum, haemoptysis, chest pain and nocturnal awakening) [10]; 2) microbiological response defined by negative fungal cultures; and 3) radiological response through assessment of cavity wall thickness and pleural thickness [15]. In one previous study on cavitary CPA disease, regression or remission of the latter two radiological criteria during concomitant 6 months of antifungal treatment was associated with clinical improvement. Loss of fungus ball on radiological imaging (without surgery) was also associated with favourable treatment response, while increase of cavity size or assessment of number and size of nodules and tree-in-bud patterns were less concordant with clinical progression [15].

Similar to the consensus statement previously published on NTM-PD, the definition of clinical cure in CPA is complicated by any remaining symptoms relating to underlying non-CPA lung disease, and therefore based on an improvement of at least 25% in the RSS [10, 13]. In European cohorts, COPD is the leading comorbidity in patients with CPA, while a history of tuberculosis is generally predominant in endemic countries [1, 3, 4]. The definition of cure in CPA is further hampered by the fact that there is no standard of care concerning treatment duration in CPA, except for the minimum duration of 6 months recommended in the current European and American guidelines for CCPA [1, 6]. CPA also has some underlying genetic risk factors, and cure may indeed be elusive, with long-term remission being a more appropriate status.

Microbiological cure in CPA may be difficult to definitively demonstrate, since culture of Aspergillus species lacks both sensitivity and specificity, and is unable to distinguish colonisation from infection in the absence of radiological findings [14]. Presence of Aspergillus species on culture from bronchoscopic specimens is more likely in infection, while sputum is usually not recommended, because of high risk of contamination/colonisation [1]. Alternative assays (antigen detection or quantitative PCR) to establish microbiological evidence for Aspergillus are insufficiently validated on broncho-pulmonary samples in the context of CPA [14]. Moreover, performing a bronchoscopy to obtain deep respiratory tract samples is seldomly justifiable in an improving patient, and less available in resource-limited settings. Serum biomarkers (galactomannan, β-D-glucan, DNA) are only occasionally detected in SAIA. Finally, while serology (specific antibody detection) is a key tool in establishing the diagnosis of CPA, its utility in follow-up remains to be demonstrated [16]. Thus, the panel agreed on including culture as one of the criteria for cure, provided that no Aspergillus was grown for respiratory samples during 2 years after completion of treatment, and with at least two different negative cultures per year from a respiratory sample (e.g. sputum, bronchial secretion, bronchoalveolar lavage).

Another limitation is that radiological changes require careful evaluation and CT image quality parameters can differ (e.g. slice thickness, pixel size, dose levels). Consensus on radiological criteria is very much concentrated on a single study [15]. Further studies are needed defining the most relevant CT imaging variables for assessing treatment response.

The methodology of this stepwise process has the intrinsic limitation of being mainly expert-opinion-based, yet has been usefully applied for MDR-TB and NTM-PD therapy outcome [12, 13]. As the available evidence on CPA treatment outcomes is even scarcer, performing a systematic review of literature on CPA treatment is intrinsically challenging [2]. In conclusion, this consensus statement on treatment outcome definitions in CPA is an important first step in evolving towards more qualitative and prospective data, with the highest priority for the development of state-of-the-art randomised clinical trials investigating treatment of CPA.

Shareable PDF