Prevalence and predictors of Aspergillus seropositivity and chronic pulmonary aspergillosis in an urban tertiary hospital in Sierra Leone: A cross-sectional study

Background In the World Health Organization Global Tuberculosis (TB) Report 2022, 37% of pulmonary TB patients were clinically diagnosed and thus many people were treated for TB without evidence of the disease. Probably the most common TB misdiagnosis is chronic pulmonary aspergillosis (CPA). In this study, we aimed to assess the prevalence and predictors of Aspergillus seropositivity and CPA in patients with chronic respiratory symptoms in an urban tertiary hospital in Sierra Leone. Methodology/principal findings We used a cross-sectional study design to recruit adults (≥18 years) from the Chest Clinic of Connaught Hospital, Freetown between November 2021 and July 2022. Aspergillus antibody was detected using LDBio Aspergillus IgM/IgG. Logistic regression was performed to assess the independent predictors of Aspergillus seropositivity and CPA. Of the 197 patients with chronic respiratory symptoms, 147 (74.6%) were male. Mean age was 47.1 ± 16.4 years. More than half (104, 52.8%) had been diagnosed with TB in the past, while 53 (26.9%) were on TB treatment at the time of recruitment. Fifty-two (26.4%) patients were HIV positive, 41 (20.8%) were seropositive for Aspergillus and 23 (11.6%) had CPA, 2 (3.8%) with current TB and 18 (17.3%) with past TB. Common radiologic abnormalities reported were localized fibrotic changes 62 (31.5%), consolidation 54 (27.4%), infiltrates 46 (23.4%), hilar adenopathy 40 (20.3%) and pleural effusion 35 (17.85) and thickening 23 (11.7%). Common symptoms were weight loss 144 (73.1%), cough 135 (68.5%), fever 117 (59.4%) and dyspnea 90 (45.7%). Current or past TB infection {aOR 3.52, 95% CI (1.46, 8.97); p = 0.005} was an independent predictor of Aspergillus seropositivity and CPA. Conclusions/significance We report a high prevalence of Aspergillus antibody seropositivity and CPA, underscoring the need to integrate the prevention and management of pulmonary fungal infections with TB services and asthma care in order to reduce unnecessary morbidity and mortality.


Conclusions/significance
We report a high prevalence of Aspergillus antibody seropositivity and CPA, underscoring the need to integrate the prevention and management of pulmonary fungal infections with TB services and asthma care in order to reduce unnecessary morbidity and mortality.

Author summary
Chronic pulmonary aspergillosis (CPA) is a common cause of chronic lung disease. It mimics tuberculosis (TB), and can occur during or after TB treatment, mainly in patients with lung cavities. Since nearly 40% of TB cases worldwide are undiagnosed microbiologically, CPA may be the most common cause of symptoms in patients treated for TB without a microbiological diagnosis. Understanding the burden of CPA using the Aspergillus antibody test is an important initial step in addressing this persistent and chronic neglected disease in low-resource settings and presents an opportunity for healthcare workers to acquire the skills needed to reduce unnecessary CPA-related mortality. This study assessed the prevalence of Aspergillus seropositivity and CPA and found that 20.8% of patients were positive for Aspergillus antibodies and 11.6% had CPA. Common symptoms were cough, weight loss, difficulty breathing and fever and TB was an independent predictor of Aspergillus seropositivity and CPA.

Background
The prevention and control of fungal diseases like chronic pulmonary aspergillosis (CPA) remains a major challenge, especially in resource-poor countries [1,2]. Estimates of global post-TB prevalence of CPA based on 2005 TB data suggest that approximately 1.2 million people are affected [3]. This fact demonstrates the trajectory of the increasing burden of CPA and supports recent estimates of 3 million global cases of CPA [4].
Aspergillus can coexist with or complicate tuberculosis (TB) and other chronic lung diseases and thus may pose a significant challenge to their management [3]. In the World Health Organization (WHO) Global TB Report 2022, approximately 37% of pulmonary TB patients were clinically diagnosed (i.e., were smear/Xpert negative or not done) [5]. Thus, many people were treated for TB without evidence of the disease. Probably the most common TB misdiagnosis is CPA, which causes considerable morbidity and unnecessary mortality. CPA can mimic TB and can occur during or after TB treatment, mainly in those with lung cavities [4].
CPA is a global health problem that primarily affects low-and middle-income countries. A study in Iran found that 6% of people had CPA at the time of their first TB episode within 12 months of diagnosis [6]. A recent study in Indonesia found that 13% of patients developed CPA after 6 months of TB treatment [7]. Recent data from India and Vietnam found that more than 50% of TB patients with new symptoms had CPA [8,9].
In sub-Saharan Africa, there are no comprehensive estimates of the number of misdiagnosed TB when the diagnosis is actually CPA. A prospective longitudinal study in northern Uganda found that 6.5% of people with cavities each year developed CPA 2-7 years after TB, compared with 0.2% of people without cavities [10]. A study from Nigeria found that 8.7% of the smear/ Xpert-negative population had CPA [11]. In Ghana, 9.7% of patients with presumed TB had CPA [12]. We recently estimated that 6000 people in Sierra Leone suffer from CPA [13].
Sierra Leone is one of the 30 countries with the highest TB burden in the world, with 24,000 cases (298 TB cases per 100,000 population) in 2021 [3]. In our recent studies, nearly 50% of patients treated for TB at the national referral hospital in Sierra Leone were clinically diagnosed [14,15]. Despite this and the fact that CPA is probably the main cause of their symptoms, there is limited data on the burden of this disease among patients with chronic respiratory symptoms in Sierra Leone owing to the limited human resource and diagnostic capacity for the detection, management and prevention of this fungal infections. The study aimed to assess prevalence and predictors of Aspergillus seropositivity and CPA in patients with chronic respiratory symptoms in an urban tertiary hospital in Sierra Leone in order to determine the local needs for training of healthcare workers on the detection and management of CPA, with the overall aim of improving health outcomes of people with CPA in Sierra Leone.

Ethics statement
Ethics approval was obtained from the Sierra Leone Ethics and Scientific Review Committee (SLESRC) of the Ministry of Health and Sanitation, Government of Sierra Leone in accordance with the relevant guidelines and regulations and declaration of Helsinki. Approval to conduct this study was granted by SLESRC, dated 16 th June 2021. Written informed consent was obtained from individual participants before enrolling in the study.

Study design
We used a cross-sectional study design to collect data from adult patients aged 18 years or older from the Chest Clinic of Connaught Hospital, Freetown, Sierra Leone.

Study setting
Sierra Leone is divided into five geographical regions, including the Western Area. The Western Area is the most populous region of Sierra Leone and includes Freetown (the capital city). According to the 2015 population census, Sierra Leone has a total population of 7 million, of which 22% (1.5 million) live in the Western Area [16].
Of the 25 public hospitals in Sierra Leone, 10 provide tertiary care. The study was conducted at Connaught Hospital, Sierra Leone's main tertiary hospital in Western Area with a capacity of 300 beds. Connaught Hospital's Chest Clinic has the largest number of TB patients in the country.

Sampling method
Though a minimal sample size of 118 patients was obtained with Fisher's formula using prevalence of 8.7% of CPA in Nigeria [11], with a coefficient interval of 95% and a margin of error of 0.05, a total of 197 patients were recruited.
Adults (�18 years) with respiratory symptoms were recruited sequentially from the first week of November 2021 to reach a sample size of 197 patients (approximately 30 weeks). Demographic and clinical information were collected from patients using a standardized data collection form and cross-checked with clinical records. Symptoms and chest radiograph findings were recorded using a diagnostic algorithm published in 2018 [17]. The presumed pneumonia or bronchitis cases were inferred from antibiotic use in this patient cohort.

Laboratory procedure
Blood samples were aseptically collected into EDTA test tubes and immediately centrifuged at 4000 revolutions per minutes for 5 minutes to generate plasma. Separated plasma were then analyzed using Aspergillus IgM/IgG (LDBIO Diagnostics, Lyon, France) point-of-care immunochromatographic (ICT) lateral flow assay [18]. The LDBio Aspergillus IgM/IgG ICT test has a sensitivity and specificity of 88.9% and 96.3%, respectively. Similar to other point-of-care tests, it has been added to the 2021 WHO Model of Essential Diagnostic List (EDL) [18,19].
TB was diagnosed using the Xpert Mycobacterium tuberculosis (MTB)/RIF assay (Cepheid) with compatible clinical features and chest radiograph findings [20]. HIV test was performed using Determine and SD Bioline HIV-1/2 3.0 test kits (Standard Diagnostics Inc) in the National HIV Testing Algorithm [21]. CD4 cell count was determined using the Alere Pima™ Analyzer (Abbott), a point-of-care testing platform validated in resource-limited settings [22].

Definitions of Aspergillus seropositivity and CPA
Patients with chronic respiratory symptoms such as weight loss, cough, and/or hemoptysis persisting for more than 3 months and positive for LDBIO Aspergillus IgG were considered to have Aspergillus seropositivity regardless of their radiological findings.
The Global Action For Fungal Infections (GAFFI) and its international experts established a case definition for CPA in 2018 to support research and clinical and public health practices in low-and middle-income countries [23]. The GAFFI panel defined CPA as an illness lasting more than three months with the presence of all the typical symptoms (weight loss, persistent cough, and/or hemoptysis) and chest radiograph showing progressive cavitary infiltrates and/ or a fungal ball and/or peri-cavitary fibrosis or infiltrates or pleural thickening in patients with a positive Aspergillus IgG serology [23]. Chest radiographs were read and reported by Radiologists and further reviewed by the research team using guidance provided by Leading the International Fungal Education (LIFE) Worldwide (available at: https://www.youtube.com/watch? v=zuYoLW-n_2w).
We applied the GAFFI case definition to categorized confirmed CPA and the remaining patients with only two of the key features as probable CPA. In keeping with standard practice, patients with probable and confirmed CPA were given oral itraconazole 400 mg per day for a minimum of 6 months.

Data management and analysis
The data was collected using the Epicollect data platform, pooled into an Excel sheet and clean, coded and exported to SPSS 21.0 (Armonk, NY: IBM Corp) for analysis. The data were summarized using frequencies and measures of central tendency. Normally distributed numerical variables were summarized using mean and standard deviation (SD). Categorical variables were summarized with frequencies and proportions. Demographic, clinical and radiographic characteristics were compared between patients who are seropositive for Aspergillus or had CPA using chi-square or Fisher's exact test, as appropriate. Logistic regression analysis was used to identify potential predictors of Aspergillus seropositivity and CPA. Only

Sociodemographic characteristics and associated disease
Of the 197 patients with chronic respiratory symptoms, 147 (74.6%) were male. The mean age is 47.1 ± 16.4 years. Many patients were married 112 (56.9%) and had secondary education as their highest education level 85 (43.1%). The mean BMI was 17.4 ± 3.7 kg/m 2 . More than half (104, 52.8%) had been diagnosed with TB in the past, while 53 (26.9%) were on anti-TB treatment at the time of recruitment. Fifty-two (26.4%) patients were HIV positive, 129 (65.5%) had recent or current pneumonia or bronchitis, and 8 (4.1%) each had chronic obstructive pulmonary disease (COPD) or asthma. This method is prone to error but not likely to differ radically from the true number of pneumonia cases given the usual pattern of clinical practice. Pneumonia/bronchitis was not verified by radiology or microbiology testing. Forty-one (20.8%) patients were positive for Aspergillus IgG/IgM antibody (Tables 1 and 2).

Radiologic findings
Chest radiographs were available for all 197 patients with chronic respiratory symptoms, many of whom showed compatible features with CPA. Fig 1A-1C provide examples of typical x-ray changes in patients with chronic respiratory symptoms. Xray changes reported were bilateral in 100 (50.8%) patients.

Symptoms
Respiratory symptoms of the 197 patients are shown in Table 2.  Table 2). Amongst the 53 patients with current TB, 3 (4.1%) and 2 (3.8%) were Aspergillus seropositive or had CPA respectively ( Table 2). Of those with past TB (n = 104), 31 (29.8%) had a positive Aspergillus antibody result and 18 (17.3%) had CPA. The 92 patients without TB had completed TB treatment and were therefore not eligible for a repeat TB test by national standards (n = 8), recruited at a time when there were disruptions in the supply of GeneXpert cartridges (n = 72) or unable to produce sputum for Xpert test (n = 12).  (Table 4). Likewise, asthma, COPD and pneumonia/bronchitis were not linked to CPA. The presence of past or current TB infection {aOR 3.71, 95% CI (1.23, 11.24); p = 0.008} was the only independent predictor of CPA on multivariable analysis (Table 5).

Discussion
We analyzed the prevalence and predictors of Aspergillus seropositivity and CPA in adult patients with chronic respiratory symptoms in a national referral hospital in Sierra Leone. Amongst the 197 patients with chronic respiratory symptoms, 20.8% had a positive serum Aspergillus antibody assay, higher than the serological evidence of the 19.5% reported among TB patients with persistent pulmonary symptoms in Kenya [24]. Aspergillus IgG in combination with compatible clinical symptoms and radiologic findings has become the preferred approach for diagnosing CPA [25]. Using the case definition developed by GAFFI [23], we reported a CPA prevalence of 11.6%, split between those being treated for TB (3.8%) and those with a past history of TB (17.3%). A lower prevalence of CPA was reported in Ghanaian patients with suspected TB (9.7%) and Nigerian patients with smear-negative TB (8.7%) [11,12].
The higher prevalence of CPA in our study could be attributed to the additional culture method applied in the detection of CPA in the Nigerian and Ghanian studies. Evidence is accumulating that false-positive results of Aspergillus IgG exist, because a positive Aspergillus antibody is not specific to CPA, and can represent several other Aspergillus-related disorders, including Aspergillus rhinosinusitis, allergic bronchopulmonary aspergillosis, Aspergillus bronchitis and subacute invasive aspergillosis. We partly address this by categorizing the remaining patients with only two of three key features as probable CPA [23]. Regardless of the category, the high burden of Aspergillus seropositivity reported in our study reflect on our earlier call to improve on the diagnostic capacity of fungal infections in Africa and underscores the need to initiate routine screening among patients with chronic respiratory symptoms and provide timely and appropriate anti-fungal treatment to those with CPA [26,27]. Most of the CPA patients in this study experienced one or more chronic respiratory symptoms. Common symptoms were weight loss (65.9%), cough (61.0%) and dyspnea (43.5%), but none was associated with Aspergillus seropositivity or CPA. Nonetheless, hemoptysis, which is one of the most dramatic and cardinal symptoms of CPA as described by Sapienza was present in only 22% of the patients [28]. The lack of distinctive clinical features clearly indicates that early detection of CPA in patients with chronic respiratory symptoms requires detection of Aspergillus antibody and compatible radiologic findings for diagnosis [29,30]. Because fever is inherently an uncommon symptom of CPA, the high proportion of febrile patients with Aspergillus seropositivity reported in this study warrants a thorough investigation of concurrent pneumonia (or bronchitis), which is reported in 65.5% of patients in this study or subacute invasive aspergillosis whenever an acute presentation occurs in patients with CPA [31,32].
In our study, 13.8% of Aspergillus antibody-positive patients were HIV-infected and 9.8% had dual CPA/HIV infections, which is different from the 7% reported from India [33]. Although these findings may reflect the high burden of HIV in this setting, people living with HIV were less likely to have seropositive Aspergillus antibody on multivariate analysis [34,35]. The reduced ability of HIV-infected individuals to produce antibodies or induce inflammatory changes due to impaired immune responses may explain the protective effect of HIV against Aspergillus seropositivity [36]. Notably, many patients with a positive Aspergillus antibody and CPA had either been diagnosed with TB in the past or were receiving anti-TB treatment at the time of recruitment. TB was an independent predictor of Aspergillus seropositivity and CPA in our univariate analysis and remained significant even in multivariable analysis. CPA is known to coexist with other comorbidities, especially TB. More importantly, nearly 50% of patients  [14]. CPA may be the underlying disease in these patients, emphasizing the need to integrate management and prevention of pulmonary fungal infections, including CPA into TB care. Asthma has a well-established association with Aspergillus infection [37,38] and in our study, it predicts Aspergillus seropositivity in multivariable analysis. However, we did not confirm the presence of fungal asthma in any of the patients due to the limited diagnostic capacity locally [36,37]. Like asthma, COPD has a link with Aspergillus infection but was not predictive of seropositivity in our study [39]. The radiographic hallmarks of CPA in patients with chronic lung disease are cavities with or without intracavity fungal ball formation, pleural thickening, and pulmonary fibrosis [40]. We found no patient with a visible fungal ball on chest X-ray, but we didn't have resource to do a computed axial tomography (CT) scan, which would probably have been more sensitive. In the present study, hilar lymphadenopathy and consolidation was less frequent in CPA, whereas extensive pulmonary fibrosis peri-cavitary fibrosis, pleural thickening and pleural effusion were all more likely to be reported in patients with Aspergillus seropositivity and thus could provide valuable information in the diagnosis of CPA. Since CT, which provides better lung details, was not used to diagnose CPA, we may have underestimated the true prevalence of CPA in this study.
Our study used a relatively small sample size and was conducted in one center, so the findings cannot be easily generalized to the Sierra Leonean population with chronic respiratory symptoms. Despite these inherent limitations, we provide the first evidence on Aspergillus seropositivity and CPA in Sierra Leone to highlight the need to train healthcare workers and intensify the diagnosis, management and prevention of CPA in a high TB burden country.

Conclusion
We report a high prevalence of Aspergillus antibody seropositivity and CPA. A common independent predictor of Aspergillus seropositivity and CPA was present or past TB infection. These findings underscore the need to integrate the prevention and control of pulmonary fungal infections with TB services and asthma care in order to reduce unnecessary morbidity and mortality.
Supporting information S1 Data. SPSS data set for aspergillus seropositivity and chronic pulmonary aspergillosis. (SAV)