A Systematic Review of Chronic Pulmonary Aspergillosis Among Patients Treated for Pulmonary Tuberculosis

Abstract

Background

Tuberculosis (TB) is a major global health concern, with long-term complications persisting even after successful treatment. Chronic pulmonary aspergillosis (CPA) is a progressive fungal disease that frequently develops in TB survivors, contributing to post-TB lung disease. The true burden of CPA among patients with TB remains unclear due to diagnostic challenges and limited data. We aimed to estimate the prevalence of CPA among patients with prior or concurrent TB.

Methods

We conducted a systematic search in PubMed, Cochrane Library, Web of Science, and Science Direct through 10 January 2025. Eligible cohort and cross-sectional studies reported CPA prevalence in patients diagnosed with TB based on clinical symptoms, radiographic abnormalities, and microbiological evidence. Three reviewers screened 1575 unique studies, assessed 118 full texts, and included 22 studies (2884 patients). We conducted a meta-analysis using a random-effects model to estimate pooled CPA prevalence, with subgroup and meta-regression analyses exploring factors influencing CPA burden.

Results

CPA prevalence varied by timing of assessment and symptom status. Among all patients with TB, CPA prevalence was 9% (95% confidence interval [CI]: 6%–12%) during treatment and 13% (95% CI: 6%–27%) posttreatment. Among patients with persistent respiratory symptoms, CPA prevalence was 20% during treatment and 48% (95% CI: 36%–61%) posttreatment. Meta-regression identified symptom status and timing of CPA assessment as significant predictors of CPA prevalence.

Conclusions

The high CPA burden among TB survivors, particularly those with persistent symptoms, underscores the need for routine CPA screening in TB programs. Early detection and targeted interventions could reduce respiratory complications and improve patient outcomes.

chronic pulmonary aspergillosis, post-tuberculosis lung disease, prevalence, systematic review, tuberculosis

Tuberculosis (TB) is a leading cause of death from an infectious disease globally [1]. The World Health Organization (WHO) estimated that 10.6 million people worldwide developed active TB in 2021 and 1.6 million died from the disease [2]. Although TB treatment has an average success rate of 86% among patients started on first-line regimens [2], complications of TB can lead to significant mortality and disability even after successful treatment. Previously treated patients are 3 times more likely to die within 5–10 years than the general population even when they do not have recurrent TB [3, 4]. Importantly, pulmonary TB (PTB) often leads to significant lung impairment, and patients treated for PTB are at risk for long-term complications, including chronic obstructive pulmonary disease, restrictive lung disease, and bronchiectasis [5], syndromes usually referred to as post-TB lung disease (PTLD). Patients with a history of PTB constituted a large proportion of the adult population with a prevalence of 8.4% in several high-burden communities [6]; thus, a large portion of the population is at risk for PTLD.

After PTB treatment, a significant proportion of patients have residual lung cavities [7, 8]. These lingering spaces in the lungs provide an opportunity for fungal colonization, which can lead to the expansion of such cavities and subsequent surrounding lung damage [9]. Chronic pulmonary aspergillosis (CPA) is a progressive fungal disease affecting an estimated 3 million people globally, especially those with a current or prior underlying lung disease [10]. CPA typically manifests with symptoms such as chronic cough, dyspnea, and weight loss, which closely resemble those of TB [11]. This similarity often complicates the diagnosis of CPA, particularly in areas with limited access to diagnostic tools. Consequently, CPA is frequently underdiagnosed or misdiagnosed as new or relapsed TB [12, 13]. These diagnostic challenges can delay the initiation of appropriate treatment, potentially resulting in preventable deaths. CPA has a case fatality rate of 15% within 1 year of diagnosis and 30%–50% mortality after 5 years, depending on CPA subtype and underlying lung disease [14, 15]. This highlights the need for data to understand the impact of fungal diseases such as aspergillosis on PTLD and the extent of this contribution to inform diagnostics, treatment, and public health efforts.

The current global burden of CPA among patients treated for TB is unknown. In this systematic review and meta-analysis, we investigated the contribution of CPA to TB-related lung disease by reviewing and summarizing the existing evidence regarding the burden of CPA among patients with prior or active TB.

METHODS

This systematic review followed the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [16]. The PRISMA checklist is provided in Supplementary Table 1.

Search Strategy and Selection Criteria

We conducted a systematic search of published literature across 4 databases: PubMed, Cochrane Library, Web of Science, and Science Direct. The search was conducted through 10 January 2025, using the search strategy shown in Supplementary Table 2. We included cohort and cross-sectional studies that reported the prevalence of CPA in patients who had been diagnosed or treated for TB. Following current guidelines from the Infectious Diseases Society of America and Global Action for Fungal Infections [17, 18], we considered studies eligible if the diagnosis of CPA in patients with TB included all the following: (i) compatible clinical symptoms of >3 months’ duration; (ii) radiographic abnormalities including cavitation, pleural thickening, infiltrates, or fungal ball; and (iii) a positive Aspergillus-specific immunoglobulin G (IgG) assay or other microbiological evidence of Aspergillus infection. We excluded review studies, preprints, case reports, and case series. We included letters to the editor and conference presentations if relevant details were extractable.

Data Extraction

Three co-authors (S. K. O., A. M., and M. B.) screened the titles and abstracts of retrieved studies and assessed full texts for eligibility using the Covidence platform. In the event of a disagreement, the authors discussed until consensus was reached. The reviewers used a piloted form to capture data on study titles, year of publication, study design, country(s) of study, sample size, years of the study, age and sex of patients, CPA prevalence, patient recruitment, timing of CPA assessment, and proportion of the population with known risk factors for tuberculosis, including human immunodeficiency virus (HIV), diabetes mellitus (DM), and smoking history. We emailed study authors if the full-text article was not available or if relevant patient information could not be extracted from the text.

Statistical Analyses

We carried out a meta-analysis of the burden of CPA in patients with TB using a random-effects model to account for the variability of effect sizes, generating a pooled burden and 95% confidence interval (CI) [19]. We used inverse-variance weighting to account for differences in sample size among the included studies. We applied maximum-likelihood estimation to determine the between-study variance τ² and the Hartung-Knapp method for generating robust CIs for random-effects estimates [20]. We carried out subgroup analyses stratifying on the timing of CPA assessment (during or after TB treatment), sampling population (all patients with TB or only those with persistent pulmonary symptoms), and WHO study region [21]. We estimated heterogeneity among studies with the I² statistic [22]. We summarized the outcomes of the meta-analysis with forest plots and assessed potential publication bias using a funnel plot [19].

We used mixed-effects meta-regression in R to explore the effect of the following possible sources of heterogeneity: WHO region, timing of CPA assessment, whether participants were recruited based on persistent symptoms, age category, proportion of male participants, prevalence of HIV infection, and prevalence of diabetes. We excluded HIV and DM from the multivariable meta-regression since few studies reported these variables. Within each risk factor, we assigned the category with the lowest prevalence in the univariate model as the reference group and estimated the prevalence ratio (PR) as the prevalence compared to this reference. A categorical variable was used to describe the main subgroups of patients by symptom status and timing of CPA assessment. We applied the Hartung-Knapp method for generating conservative CIs and performed a permutation test with 1000 iterations to assess the robustness of our model results [23].

RESULTS

Our initial search yielded 2534 results; 1575 records remained after removing duplicates (Figure 1). After title and abstract screening, we identified 118 studies for which we conducted a full-text eligibility review. Ninety-six full texts were excluded for reasons detailed in Supplementary Table 3, leaving 22 studies that met all eligibility criteria and were included in our meta-analysis.

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart of study screening and selection. Additional details on the systematic search and full-text exclusions are available in the Supplementary material. Abbreviations: CPA, chronic pulmonary aspergillosis; TB, tuberculosis.

Characteristics of Included Studies

Table 1 describes the characteristics of the 22 included studies and their participants [24–45]. All studies were published between 2015 and 2025, with the majority (n = 16 [73%]) published since January 2022. Participant data for the studies were collected between 2012 and 2024. Twelve studies (55%) were cross-sectional, 7 (32%) were prospective cohort studies, and 3 (14%) were retrospective cohort studies. The number of patients with past or present TB evaluated for CPA in each study ranged from 37 to 508. Reported DM prevalence ranged from 0 to 40%, and most studies either excluded HIV-positive patients or did not report HIV prevalence. The majority of studies were conducted in Southeast Asia and Africa (Supplementary Figure 1).

Table 1.

Characteristics of Included Studies

Study Author	Publication Year	Country	Study Design	TB Sample Sizea	Timing of CPA Assessmentb	Sampled Populationc	Male Sex, %	Patient Age, yd	Diabetes, %	History of Smoking, %e	HIV Positive, %
Chirumamilla et al [24]	2024	India	Cross-sectional	111	Posttreatment	Symptomatic	61.3	42.6 ± 15.7	9.9	…	0.0
Davies et al [25]	2024	Nigeria	Prospective cohort	141	Posttreatment	Mixed	53.0	40 (30–52)*	…	…	17.0
Hedayati et al [26]	2015	Iran	Cross-sectional	124	Mixed	Mixed	53.2	48*	10.5	…	0.0
Jha et al [27]	2024	India	Prospective cohort	255	During treatment	TB treated	60.0	31.5 ± 13.9	11.8	40.0	0.0
Kim et al [28]	2022	South Korea	Retrospective cohort	345	Posttreatment	TB treated	66.7	55 (37–69)*	20.9	56.5	0.0
Klinting et al [29]	2022	Denmark	Retrospective cohort	37	During treatment	TB treated	67.6	42.0	5.4	…	0.0
Lakhtakia et al [30]	2022	India	Cross-sectional	60	Posttreatment	Symptomatic	60.0	47.9 ± 12.9	…	51.7	…
Lakoh et al [31]	2023	Sierra Leone	Cross-sectional	104	Posttreatment	Symptomatic	…	…	…	…	…
Mei et al [32]	2023	China	Cross-sectional	140	Posttreatment	Symptomatic	71.4	58.4 ± 2.0	21.4	…	0.0
Namusobya et al [33]	2022	Uganda	Cross-sectional	162	During treatment	Symptomatic	60.0	30 (25–40)*	…	8.0	29.6
Nguyen et al [34]	2021	Vietnam	Retrospective cohort	68	Posttreatment	Symptomatic	…	…	…	…	…
Ocansey et al [35]	2023	Ghana	Prospective cohort	41	During treatment	TB treated	78.0	40.2	…	…	24.3
Oladele et al [36]	2017	Nigeria	Cross-sectional	208	During treatment	TB treated	40.4	39.8 ± 12.3	…	…	73.6
Oladele et al [37]	2022	Nigeria	Prospective cohort	193	During treatment	TB treated	66.8	38 (30–46)	0.0	…	0.0
Page et al [38]	2019	Uganda	Prospective cohort	285	Posttreatment	TB treated	64.9	42.0	…	10.6	47.5
Ray et al [39]	2022	India	Cross-sectional	116	Posttreatment	Symptomatic	…	…	…	…	…
Sehgal et al [40]	2025	India	Prospective cohort	508	Posttreatment	Mixed	71.4	40 (30–55)*	13.4	…	0.0
Setianingrum et al [41]	2022	Indonesia	Prospective cohort	216	During treatment	TB treated	47.0	39.8	18.0	…	…
Singla et al [42]	2021	India	Cross-sectional	100	Posttreatment	Symptomatic	96.0	42.2 ± 11.8	6.0	35.0	0.0
Soeroso et al [43]	2024	Indonesia	Cross-sectional	50	During treatment	TB treated	72.0	48*	40.0	30.0	4.0
Toychiev et al [44]	2022	Uzbekistan	Cross-sectional	200	During treatment	TB treated	27.5	43 (30–56)*	0.0	7.0	0.0
Volpe-Chaves et al [45]	2022	Brazil	Cross-sectional	193	Mixed	TB treated	78.8	49.0 ± 16.9	14.5	53.4	…

Abbreviations: CPA, chronic pulmonary aspergillosis; HIV, human immunodeficiency virus; IQR, interquartile range; TB, tuberculosis.

^aTotal number of patients with past or present TB who were evaluated for CPA. All percentages and age statistics refer to this population of patients specifically.

^bWhether CPA was assessed during or after TB treatment. Hedayati et al and Volpe-Chaves et al assessed patients with a history of TB as well as patients undergoing treatment for a first episode of TB disease.

^cWhether all patients treated for TB were recruited, or only patients with persistent respiratory symptoms following the initiation of TB treatment.

^dAge is reported as either mean ± standard deviation or median (IQR). Medians and IQRs are indicated by an asterisk (*).

^ePrevalence of ever smokers.

The types of patients recruited varied across studies. Eleven studies evaluated past or present patients with TB regardless of their symptoms [27–29, 35–38, 41, 43–45] while 6 studies [24, 30, 32–34, 42] evaluated only patients with TB who reported persistent respiratory symptoms ≥2 months after the initiation of TB treatment. Three studies enrolled all patients with respiratory symptoms but then reported CPA prevalence among those with past or present TB [26, 31, 39]. Two studies employed a combination of recruitment methods, with some participants recalled from records of past TB treatment and others presenting to TB clinics or hospitals with respiratory symptoms [25, 40]. Posttreatment patient cohorts often included cases of recurrent or relapsed TB; only 6 studies explicitly excluded active TB in posttreatment patients [26, 30, 34, 40, 42, 45], while others included active TB [25, 31, 32, 35, 38, 39] or did not assess patients for TB [24, 28, 37].

Figure 2 summarizes the timing of CPA assessments relative to TB treatment. Four studies assessed CPA at TB treatment baseline and at end of TB therapy, and 2 of these made an additional assessment at 12 months. In other studies that assessed posttreatment patients, the time between TB treatment completion and CPA assessment varied from less than a year to >13 years.

Figure 2.

Graphical overview of chronic pulmonary aspergillosis (CPA) assessment timing relative to tuberculosis (TB) treatment baseline. *Studies include patients enrolled based on persistent respiratory symptoms after the start of TB treatment. Namusobya et al (2022) enrolled currently treated patients with TB who had persistent respiratory symptoms despite 2 months of anti-TB treatment. Oladele et al (2022) made a fourth assessment of CPA at 3 months posttreatment that was not included in our analysis.

Burden of CPA

We divided studies with heterogenous patient populations into 4 cohorts based on timing of CPA assessment (ie, during or after treatment), and whether the patient population included all treated patients with TB regardless of symptoms or included only patients with persistent symptoms after ≥2 months after the initiation of treatment (Figure 3). Among studies that enrolled all patients with TB, CPA prevalence was 9% (95% CI: 6%–12%) in studies that evaluated patients during TB treatment and 13% (95% CI: 6%–27%) in those that evaluated patients after treatment. Among patients with persistent respiratory symptoms, CPA prevalence was 20% in those evaluated during treatment and 48% (95% CI: 36%–61%) in those evaluated after treatment completion.

Figure 3.

Forest plot of chronic pulmonary aspergillosis (CPA) burden in patients with tuberculosis (TB). Study cohorts are stratified by timing of CPA assessment and TB patient population. Repeated follow-ups are listed separately and labeled with the timing of CPA assessment relative to TB treatment baseline. Only 1 study enrolled patients who had persistent respiratory symptoms during TB treatment, so no pooled prevalence estimate was calculated for this category. Abbreviations: CI: confidence interval; TB, tuberculosis.

Meta-regression

The univariate meta-regression identified CPA assessment after treatment and enrollment of symptomatic patients as statistically significant predictors of CPA burden in the included studies. In contrast, the association of age category, sex, DM prevalence, HIV prevalence, and WHO region with CPA was not statistically significant at the P = .05 level (Supplementary Table 4). In the multivariable regression (Table 2), the association between CPA prevalence and study population remained statistically significant. After we adjusted for age category, sex, and WHO region, patients with persistent symptoms after treatment had 1.61 times the prevalence of CPA compared to those assessed during treatment (95% CI: 1.40–1.86). Studies that enrolled patients with persistent respiratory symptoms during TB treatment reported higher CPA prevalence compared to those that enrolled general TB treatment cohorts (PR, 1.14 [95% CI: .92–1.42]), though this difference did not reach the statistical significance. Studies that assessed patients after TB treatment, regardless of symptom status, did not report significantly higher CPA prevalence than those that assessed patients during treatment (PR, 1.08 [95% CI: .97–1.21]). Age category of participants, sex, and WHO region of the study were not associated with CPA burden in the multivariable model.

Table 2.

Mixed-Effects Meta-regression for Chronic Pulmonary Aspergillosis Risk Factors in Patients With Tuberculosis

Variable	Adjusted PR	(HK Robust 95% CI)	HK Robust P Value	Permutation Test P Value
Age category (y)
30–40	ref
>40–49	0.98	(.86–1.10)	.675	.678
50+	1.16	(.80–1.69)	.396	.381
Sex (%)
Male	1.00	(1.00–1.00)	.997	.999
WHO Region
European Region	ref
African Region	0.97	(.81–1.16)	.729	.748
Region of the Americas	0.98	(.75–1.24)	.762	.721
South-East Asia Region	1.03	(.85–1.25)	.736	.747
Western Pacific Region	0.76	(.51–1.12)	.154	.167
CPA assessment timing and population
During TB treatment, all	ref
After TB treatment, all	1.08	(.97–1.21)	.151	.170
During TB treatment, persistent symptoms	1.14	(.92–1.42)	.205	.155
After TB treatment, persistent symptoms	1.61	(1.40–1.86)	<.001	.001

The Hartung–Knapp method for random-effects meta-analysis was applied to generate robust CIs. Age category reflects the mean or median age of the patient cohort. A permutation test with 1000 iterations was performed. Human immunodeficiency virus and diabetes mellitus variables were not included in the multivariable regression due to sparse data.

Abbreviations: CI: confidence interval; CPA, chronic pulmonary aspergillosis; HK, Hartung–Knapp; PR, prevalence ratio; ref, reference group; TB, tuberculosis; WHO, World Health Organization.

Publication Bias

The funnel plot of all included studies is highly asymmetrical (Egger test P = .0019) (Supplementary Figure 2), but this heterogeneity is largely explained by study differences in the timing of assessment and enrollment criteria.

DISCUSSION

In this meta-analysis of the burden of CPA among patients with current or previous PTB, we found significant differences in prevalence based on symptom status and timing of evaluation. The global prevalence of CPA diagnosis among 2884 patients with past or present TB was 9%. Among 1194 patients with persistent respiratory symptoms after TB treatment, CPA was observed in 48%. The consistently high prevalence of CPA—over 9%—across all groups indicates that routine evaluation for CPA at the end of TB treatment could help mitigate long-term complications, especially if evaluation leads to effective therapy. Furthermore, the strikingly high prevalence in patients with residual lung symptoms on completion of TB treatment strongly suggests that CPA assessment should be integrated into standard TB care for this group.

Other than the presence of persistent respiratory symptoms after treatment, our meta-regression identified few risk factors for CPA. We did not find a significant increase in CPA after treatment in the overall TB treatment group due to the small number of studies available, but Volpe-Chaves et al [45] and Oladele et al [37] both reported increases in CPA prevalence with increasing time from TB diagnosis. We found no evidence for variation in CPA prevalence by world region after controlling for other factors. Although studies with higher proportions of patients with diabetes had slightly elevated prevalences of CPA, this finding was not statistically significant. Instead, the variability in the results of these studies was largely explained by timing and symptoms and the interaction between these 2 variables.

Most of our studies excluded HIV-positive patients or did not report data on HIV infection. However, it is interesting to observe that among the few studies that enrolled HIV-positive patients, the prevalence of CPA was lower in the HIV-positive participants compared to patients without HIV. Namusobya et al [33] reported CPA among 17% of HIV-positive patients compared to 21% of HIV-negative patients, Page et al [38] reported CPA in 2.9% of patients with HIV compared to 6.7% without HIV, and Oladele et al [36] reported CPA in 6.5% of patients with HIV compared to 14.5% without HIV. This may be attributable to a lower prevalence of TB cavitary lesions in patients with HIV [46], but further study of CPA development in HIV-positive patients with TB is warranted.

Our work extends the only other meta-analysis of Aspergillus prevalence in patients with TB of which we are aware. Hosseini et al pooled data from 17 studies from Africa and Asia that reported on the prevalence of different clinical presentations and species of aspergillosis in patients with TB, without restricting studies to those with a formal definition of CPA [47]. The reported aspergillosis prevalence among the included studies ranged from 5% to 40% with a pooled prevalence of 15.4%. Data were not provided on either the timing of assessment in relation to TB treatment or on the presence of persistent symptoms in this population. Notably, because of differences in the study selection criteria, there was no overlap between the studies included in the Hosseini et al review and our meta-analysis.

These estimates of Aspergillus prevalence from cohorts of patients with TB are somewhat higher than modeled estimates reported by Denning et al in a 2011 study that developed a deterministic model based on the expected proportion of patients with TB with cavitary lesions and the observed rate of development of aspergillosis in patients with PTB with cavitary lesions [7]. This model estimated that 372 000 of the 7.7 million incident TB cases in 2007 (4.8%) had developed CPA. In a later study using similar methodology, Denning et al estimated that 10% of patients with PTB in India presented with CPA in the first year after PTB diagnosis [48].

Our study has a number of limitations. First, many of the studies we included assessed CPA either at the time of TB treatment or within 6 months after the completion of treatment, whereas it is probable that incident CPA continues to develop in former TB patients with persistent cavities for ≥5 years after successful treatment of PTB. Conversely, while most studies show that CPA prevalence increases over time from diagnosis of TB, some have reported reversion of Aspergillus IgG over the course of treatment, which may introduce another source of error. As noted by Denning et al, population-level CPA prevalence will be influenced not only by the later development of aspergillosis but also by the death rate in those with the disease. Second, we did not restrict studies to those which confirmed TB diagnosis microbiologically, so it remains possible that some patients with CPA in these studies were misdiagnosed with TB and instead had aspergillosis alone that masqueraded as TB. False-positive diagnoses of TB, in addition to causing stress on both patients and the healthcare system [49], hinder the study of CPA. Although we classified patients with CPA according to the recommended criteria, we note that neither the sensitivity nor specificity of Aspergillus antibody is 100%, and test performance can vary with the species of Aspergillus involved. Few of the included studies collected data on relevant patient characteristics such as smoking or lifetime TB history that would be expected to have an impact on CPA prevalence. In particular, since CPA is most likely to occur in patients with lung damage from PTB, those with a history of repeated TB disease would be expected to be at especially high risk.

Finally, it is important to note that while we discuss CPA as a single diagnosis, it is actually a heterogenous disease. The overlapping presentations of CPA include single aspergilloma, Aspergillus nodules, chronic cavitary pulmonary aspergillosis, chronic fibrosing pulmonary aspergillosis, and subacute invasive pulmonary aspergillosis [50]. Chronic cavitary pulmonary aspergillosis is the most common subtype of CPA, with more complicated treatment and poorer survival than single aspergilloma or nodules. The recent meta-analysis by Sengupta et al [15] estimates that mortality from CPA is highest in the first year after diagnosis, suggesting that early detection and appropriate management by subtype are likely to be critical for reducing CPA deaths globally.

These limitations highlight the need to carefully design future studies of CPA among PTB survivors to estimate the contribution of CPA to PTLD and to identify those who could benefit from antifungal treatment.

CONCLUSIONS

This systematic review and meta-analysis highlights the substantial burden of CPA among TB survivors, particularly those with persistent respiratory symptoms after treatment. Given the high prevalence of CPA in this group, integrating CPA screening into routine TB care could improve early diagnosis and treatment, potentially reducing the long-term respiratory complications of PTLD. Future research should focus on refining diagnostic strategies and evaluating targeted interventions to improve outcomes for TB survivors at risk of CPA.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Note

Financial support. M. B. M. and A. E. M. were supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health (grant number U19 AI142793-03). S. K. O. was supported on contract 200–2016-91779 with the Centers for Disease Control and Prevention (CDC). Other authors were not compensated for this work.

Author notes

A. E. M. and S. K. O. contributed equally to this work.

Potential conflicts of interest. The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.