Abstract
Background and Objective:
Serum periostin and carcinoembryonic antigen (CEA) are markers of type 2 inflammation. However, their role in diagnosing and monitoring treatment responses in allergic bronchopulmonary aspergillosis (ABPA) remains uncertain. The objective of the study was to assess the diagnostic performance of serum CEA and periostin in distinguishing ABPA from asthma. We also evaluate their usefulness in monitoring treatment responses.
Methods:
We enrolled consecutive subjects with ABPA (cases) and asthmatic patients without ABPA (controls). Serum periostin and CEA levels were measured at baseline and again 2 months after oral prednisolone. We constructed receiver operating characteristic (ROC) curves and determined sensitivity and specificity using the optimal cut-off derived from Youden’s index.
Results:
We enrolled 112 and 108 subjects with ABPA and asthma (median age: 34 years) respectively. At baseline, the median serum periostin (22.03 vs 16.36 ng/mL; P < 0.001) and CEA levels (4.80 vs 2.35 ng/mL; P < 0.001) were significantly higher in the ABPA group than in the controls. CEA (AUROC = 0.77) showed better diagnostic accuracy than serum periostin (AUROC = 0.64) in differentiating asthma from ABPA. After 2 months of treatment, median serum CEA (4.8 ng/mL vs 3.7 ng/mL) and periostin levels (22.03 vs 17.74 ng/mL) declined significantly. However, 34% and 30% of subjects exhibited increased periostin and CEA levels following treatment.
Conclusion:
While serum CEA and periostin exhibit modest diagnostic performance in differentiating ABPA from asthma, they are suboptimal for monitoring treatment responses. Further studies are required to validate our findings.
KEY WORDS: ABPA, allergy, bronchiectasis, eosinophilic inflammation, Th2 cytokines
INTRODUCTION
Allergic bronchopulmonary aspergillosis (ABPA) is an allergic disorder marked by an exaggerated type-2 immune response against the ubiquitous fungus Aspergillus, typically complicating the course of patients with asthma or cystic fibrosis.[1] Ineffective fungal clearance, skewed type-2 immune responses, and eosinophil accumulation result in the characteristic clinical, radiological (bronchiectasis, high attenuation mucus), and immunological features seen in ABPA.[2] Patients with ABPA experience recurrent exacerbations, necessitating periodic monitoring. India has a high prevalence of ABPA-complicating asthma, probably attributable to environmental or hitherto unknown genetic factors,[3,4] and effective management strategies are required.
Currently, ABPA patients are monitored every 8–12 weeks following treatment with a combination of symptom assessment, chest radiography, and serum total IgE.[5] However, as more targeted therapies emerge, there is a growing need for more reliable markers of disease activity and treatment response. Potential markers include sputum eosinophil counts, fraction of exhaled nitric oxide (FeNO), and type-2 cytokines (IL-4, IL-5, and IL-13).[6]
Serum periostin, a marker of type 2 inflammation produced in response to IL-4 and IL-13, is elevated in ABPA. Serum periostin may be particularly relevant in individuals with low IgE levels (<500 IU/mL), such as those in Japanese cohorts or elderly ABPA patients.[7,8] Additionally, carcinoembryonic antigen (CEA), initially recognised as a tumour marker, is also associated with eosinophilia in ABPA. Only a few studies have evaluated serum CEA and periostin levels in ABPA, and more evidence is required.[7,9]
Our study objectives were to evaluate (1) the diagnostic performance of serum periostin and CEA in differentiating asthma patients with and without ABPA and (2) whether serum periostin or CEA levels help assess treatment response in ABPA.
METHODS
Study design and setting
We conducted a prospective observational study from July 2021 to December 2022 at our Institute’s Chest Clinic. We adopted a case-control design for the first objective, comparing asthmatic subjects with and without ABPA. For the second objective, we followed the ABPA cohort after therapy. The institute’s ethics committee approved the study protocol (NK/7631/MD-625, 14 September 2021), and all participants provided written informed consent.
Study participants
We enrolled participants with newly diagnosed acute ABPA and asthmatic subjects without ABPA. We diagnosed asthma in subjects not meeting the criteria for ABPA but fulfilling at least two of the following: (1) recurrent episodes of breathlessness, cough, or chest tightness; (2) wheeze on chest auscultation; (3) obstructive defect on spirometry, defined as ratio of forced expiratory volume in the first second (FEV1) to forced vital capacity (FVC) below the lower limit of normal, along with bronchodilator responsiveness (increase in FEV1 or FVC by more than 12% and 200 mL after 400 mcg of salbutamol inhalation).
We diagnosed ABPA according to the revised International Society for Human and Medical Mycology (ISHAM) working group criteria.[5] Patients met both of the following: serum A. fumigatus-IgE ≥0.35 kUA/L and serum total IgE ≥500 IU/mL, and any two of the following: (i) serum A. fumigatus-IgG ≥27 mgA/L; (ii) blood eosinophil count (BEC) ≥500 cells/µL; (iii) thoracic imaging showing bronchiectasis, high-attenuation mucus (HAM), or fleeting opacities.
We excluded subjects with any of the following: cystic fibrosis, treatment with antifungal triazoles or systemic glucocorticoids for more than 3 weeks in the last 6 months, pregnancy, concomitant malignancies, immunosuppressive states, or chronic kidney, heart, and liver diseases, or allergic bronchopulmonary mycoses (due to fungi other than Aspergillus spp.).
Management of ABPA and response assessment at two months
We treated acute ABPA with oral prednisolone (0.5 mg/kg/day for 4 weeks, tapered and completed over 4 months) and assessed patients 2 months after therapy with clinical review, serum total IgE, and chest radiograph. Treatment response was defined as a major improvement in symptoms (at least 50%) and chest radiograph, and a ≥20% decline in serum total IgE from pre-treatment levels.[10,11]
Data collection and study procedure
We recorded relevant clinical details, chest computed tomography (CT) findings, spirometry, and immunological parameters of the enrolled participants. Spirometry was performed using the ultrasonic flow sensing spirometer (Easy on-PC; Version V02, NDD, Medizintechnik AG, Zurich, Switzerland) per the national guidelines.[12] The following parameters were recorded: FEV1, FVC, FEV1/FVC ratio, and bronchodilator responsiveness. We measured A. fumigatus-IgE, A. fumigatus-IgG, and total IgE in the serum by using the commercially available fluorescent enzyme immunoassay (Phadia 250, Thermofisher Scientific, Uppsala, Sweden).
We obtained blood samples for the index tests (serum CEA and periostin levels) from all the participants at enrolment and after 2 months of therapy (only in the ABPA cohort). Serum samples were refrigerated at −80°C until analysis, and the index tests were conducted by investigators (PK, AP) blinded to clinical and treatment details. Serum CEA level (ng/mL) was estimated using an immunoenzymometric assay. Periostin levels were estimated using a solid-phase sandwich enzyme-linked immunosorbent assay (ELISA) with a human periostin-2 Duoset ELISA kit (cat. no.: DY3548B, R and D Systems) at 1:10 dilution. Results were analysed using a 4-parameter logistics (4-PL) curve with MARS data analysis software (Version 3.20 R2) provided by SPECTROstar® Nano.
Statistical analysis
We analysed data using R (version 4.3.1) and Rstudio (tidyverse and gtsummary packages).[13,14] Continuous data are presented as medians with interquartile range, while categorical data are represented as frequencies and proportions. We compared the differences between categorical and continuous variables using the Chi-square test and Mann-Whitney U test. We constructed the receiver operating characteristic (ROC) curve and estimated the area under the ROC (AUROC) for serum periostin and CEA to differentiate asthma with and without ABPA. We identified the optimal cut-off values for CEA and periostin in differentiating ABPA from asthma using Youden’s index (J statistic = sensitivity + specificity –1) and reported the sensitivity, specificity, and negative and positive predictive values at the identified cut-off. In the ABPA cohort, we compared the baseline and 2-month values to assess the usefulness of these index tests in assessing treatment response. We calculated the change in CEA and periostin as a percentage of the difference between baseline and 2-month values divided by the baseline value. Additionally, we explored the utility of serum CEA and periostin in specific ABPA subgroups (BEC ≥500 cells/µL, BEC ≥1000 cells/µL, HAM, or bronchiectasis).
RESULTS
We enrolled 112 subjects with ABPA and 108 people with asthma. The baseline characteristics of the participants are summarised in Table 1. Compared to asthmatic subjects, those with ABPA had a significantly longer duration of asthma and a higher incidence of haemoptysis, expectoration of brownish-black mucus plugs, and inappropriate prior antituberculosis therapy. Most ABPA subjects (86%) had bronchiectasis, and one-third had HAM. Blood eosinophil count and other immunological markers (A. fumigatus-IgE and IgG, and serum total IgE) were significantly higher in ABPA than in asthmatic subjects [Table 1]. Aspergillus sensitisation was seen in 128/220 enrolled subjects (including those with [n = 112] and without ABPA [n = 16]). Aspergillus-sensitised asthmatics without ABPA had significantly lower A. fumigatus-IgG, serum total IgE, eosinophil counts, and baseline CEA levels than those with ABPA [Table S1].
Table 1.
Baseline characteristics of asthmatic subjects with and without ABPA
| Characteristic1 | Overall, n=2202 | ABPA, n=1122 | Asthma, n=1082 | P 3 | ||||
|---|---|---|---|---|---|---|---|---|
| Age, years | 34 (24–45) | 34 (25–43) | 34 (23–46) | 0.98 | ||||
| Female sex | 117/220 (53%) | 54/112 (48%) | 63/108 (58%) | 0.13 | ||||
| Duration of asthma, years | 7 (4–15) | 12 (4–20) | 6 (3–10) | <0.001 | ||||
| Haemoptysis | 33/220 (15%) | 31/112 (28%) | 2/108 (1.9%) | <0.001 | ||||
| Brownish black mucus plug | 37/220 (17%) | 29/112 (26%) | 8/108 (7.4%) | <0.001 | ||||
| ATT in the past | 30/220 (14%) | 27/112 (24%) | 3/108 (2.8%) | <0.001 | ||||
| FEV1, litres | 2.06 (1.33–2.65) | 1.99 (1.27–2.59) | 2.28 (1.85–2.96) | 0.13 | ||||
| FVC, litres | 2.67 (2.26–3.64) | 2.64 (2.16–3.60) | 2.74 (2.60–3.65) | 0.16 | ||||
| FEV1/FVC | 70 (57–80) | 65 (56–79) | 76 (67–81) | 0.11 | ||||
| Chest radiograph | 53/213 (25%) | 50/106 (47%) | 3/107 (2.8%) | <0.001 | ||||
| Bronchiectasis on CT | 94/116 (81%) | 94/109 (86%) | 0/7 (0%) | <0.001 | ||||
| High attenuation mucus | 31/94 (33%) | 31/94 (33%) | 0/0 (NA%) | >0.99 | ||||
| Blood eosinophil count, cells/µL | 436 (189–805) | 700 (420–1,115) | 252 (158–479) | <0.001 | ||||
| Serum A. fumigatus-IgE, kUA/L | 1.28 (0.03–22.23) | 22.00 (9.42–42.90) | 0.03 (0.00–0.14) | <0.001 | ||||
| Serum A. fumigatus-IgG, mgA/L | 82.50 (27.13–163.25) | 111.50 (44.05–176.25) | 17.50 (5.67–22.03) | <0.001 | ||||
| Serum total IgE, IU/mL | 3,586 (1,685–7,821) | 4,655 (2,418–8,670) | 285 (132–1,547) | <0.001 | ||||
| Serum CEA, ng/mL | 3.46 (2.01–6.60) | 4.80 (3.33–10.39) | 2.35 (1.44–3.47) | <0.001 | ||||
| Serum periostin, ng/mL | 20.34 (11.80–29.36) | 22.03 (14.96–35.34) | 16.36 (8.84–25.09) | <0.001 |
1ABPA - allergic bronchopulmonary aspergillosis; CEA - Carcinoembryonic antigen; CT - computed tomography; FEV1 - Forced expiratory volume in one second; FVC - forced vital capacity. 2Median (IQR); n/N (%). 3Wilcoxon rank-sum test; Pearson’s Chi-squared test; Fisher’s exact test
Table S1.
Baseline characteristics of Aspergillus sensitized asthmatic subjects with and without ABPA
| Characteristic1 | Overall N=1282 | ABPA N=1122 | Aspergillus sensitized asthma N=162 | P 3 | ||||
|---|---|---|---|---|---|---|---|---|
| Age, years | 34 (25 – 46) | 34 (25 – 43) | 37 (25 – 54) | 0.55 | ||||
| Female sex | 62/128 (48%) | 54/112 (48%) | 8/16 (50%) | 0.89 | ||||
| Duration of asthma, years | 12 (4 – 20) | 12 (4 – 20) | 8 (5 – 14) | 0.34 | ||||
| Hemoptysis | 32/128 (25%) | 31/112 (28%) | 1/16 (6.3%) | 0.071 | ||||
| Expectoration of brownish black mucus plugs | 32/128 (25%) | 29/112 (26%) | 3/16 (19%) | 0.76 | ||||
| History of inappropriate ATT in the past | 28/128 (22%) | 27/112 (24%) | 1/16 (6.3%) | 0.19 | ||||
| FEV1, L | 2.01 (1.29 – 2.66) | 1.99 (1.24 – 2.60) | 2.94 (1.68 – 3.17) | 0.20 | ||||
| FVC, L | 2.66 (2.18 – 3.67) | 2.64 (2.15 – 3.64) | 3.76 (2.72 – 4.13) | 0.10 | ||||
| FEV1/FVC | 66 (56 – 78) | 65 (55 – 79) | 75 (66 – 77) | 0.61 | ||||
| Abnormal chest radiograph | 50/122 (41%) | 50/106 (47%) | 0/16 (0%) | <0.001 | ||||
| Bronchiectasis on CT | 94/113 (83%) | 94/109 (86%) | 0/4 (0%) | <0.001 | ||||
| High attenuation mucus | 31/94 (33%) | 31/94 (33%) | 0 | >0.99 | ||||
| Blood eosinophil count, cells/µL | 607 (346 – 1,030) | 700 (420 – 1,117) | 411 (267 – 599) | 0.026 | ||||
| A. fumigatus-IgE, kUA/L | 19.80 (6.29 – 36.90) | 22.00 (9.28 – 43.30) | 0.82 (0.47 – 1.28) | <0.001 | ||||
| A. fumigatus-IgG, mgA/L | 95.70 (31.50 – 174.00) | 111.50 (44.00 – 176.50) | 6.12 (3.15 – 19.50) | <0.001 | ||||
| Serum total IgE, IU/mL | 4,168 (2,024 – 8,262) | 4,655 (2,414 – 8,782) | 212 (125 – 1,216) | <0.001 | ||||
| Serum CEA, ng/mL | 4.60 (2.73 – 8.38) | 4.80 (3.32 – 10.44) | 2.32 (1.82 – 4.16) | <0.001 | ||||
| Serum periostin, ng/mL | 21.26 (14.00 – 33.80) | 22.03 (14.73 – 35.58) | 17.35 (10.20 – 22.94) | 0.058 |
1ABPA - allergic bronchopulmonary aspergillosis; CEA - Carcinoembryonic antigen; CT - computed tomography; FEV1 - Forced expiratory volume in one second; FVC – forced vital capacity. 2Median (IQR); n/N (%). 3Wilcoxon rank sum test; Pearson’s Chi-squared test; Fisher’s exact test
Serum periostin and CEA levels are elevated in ABPA but have modest diagnostic performance in differentiating from asthma
The median serum periostin (22.03 vs 16.36 ng/mL; P < 0.001) and CEA (4.80 vs 2.35 ng/mL; P < 0.001) levels were significantly higher in the ABPA group compared to the asthma group. Serum CEA (AUROC = 0.77) outperformed periostin (AUROC = 0.64) in differentiating asthma from ABPA [Figure 1]. However, the diagnostic performance of A. fumigatus-IgE and IgG, and serum total IgE was superior to these index tests [Figure 1]. The optimal cut-off identified by Youden’s index was 16.99 ng/mL for serum periostin, corresponding to a sensitivity and specificity of 69.6% and 53.3%, respectively. The optimal cut-off point for serum CEA was 3.31 ng/mL, yielding a sensitivity and specificity of 75.9% and 73.8%, respectively. The correlation between serum CEA, periostin, or total IgE with BEC in ABPA subjects was weak [Figure S1 (1.1MB, tif) ].
Figure 1.
Receiver operator characteristic (ROC) curves for various investigations to differentiate asthma from ABPA. Each investigation is marked with a different colour, and the corresponding AUROC is provided (Af-IgE: A. fumigatus specific IgE, Af-IgG: A. fumigatus specific IgG, BEC: blood eosinophil count, and CEA: carcinoembryonic antigen)
Serum CEA may be better than periostin in assessing treatment response in ABPA
All study participants were treated with glucocorticoids and demonstrated symptomatic and radiologic improvement. The median serum CEA and periostin levels significantly declined after treatment [Table 2 and Figure 2]. However, there was an increase in the CEA and periostin levels following treatment in 30/100 (30%) and 34/99 (34%) subjects with ABPA, respectively [Table S2]. While all these immunological markers were imperfect, Figure 2 shows that the 2-month decline was relatively more consistent with serum CEA (especially when baseline levels were elevated) than serum periostin levels. Further, subjects experiencing a >20% decline in CEA levels had significantly higher immunological markers (BEC, serum total IgE, and A. fumigatus-IgG) and more frequent bronchiectasis than those with ≤20% decline (or increase in levels) following treatment [Table S3]. Contrarily, subjects with a periostin decline of >20% did not have a significant difference in immunological severity or bronchiectasis than those without a 20% decline [Table S4].
Table 2.
Immunological parameters at baseline and two months in patients with ABPA
| Characteristic1 | Baseline n=1122 | Follow-up n=1122 | P 3 | |||
|---|---|---|---|---|---|---|
| Serum total IgE, IU/mL | 4,655 (2,414–8,782) | 3,216 (1,620–6,311)4 | 0.003 | |||
| Serum periostin, ng/mL | 22.03 (16.16–34.66) | 17.74 (13.25–24.93)5 | 0.013 | |||
| Serum CEA, ng/mL | 4.8 (3.3–10.4) | 3.7 (2.6–5.3)6 | <0.001 |
1ABPA - allergic bronchopulmonary aspergillosis CEA - Carcinoembryonic antigen. 2Median (IQR). 3Wilcoxon rank-sum test. Data during follow-up was not available for 413, 59, and 612 of the 112 participants
Figure 2.
Line diagram showing the relation of baseline and post-treatment values (at two months) of serum carcinoembryonic antigen and periostin in ABPA patients
Table S2.
Change in serum carcinoembryonic antigen (CEA) and periostin levels following treatment
| Characteristics1 | ABPA2 | |
|---|---|---|
| % change in CEA post treatment | ||
| Increased | 30/100 (30%) | |
| ≤10% decline | 36/100 (36%) | |
| ≤20% decline | 40/100 (40%) | |
| ≤30% decline | 50/100 (50%) | |
| % change in periostin post treatment | ||
| Increased | 34/99 (34%) | |
| ≤10% decline | 42/99 (42%) | |
| ≤20% decline | 55/99 (56%) | |
| ≤30% decline | 61/99 (62%) |
1Data not available for CEA in 12 and periostin in 13 participants with ABPA. 2n/N (%)
Table S3.
ABPA subjects with a decline in CEA (>20%) after treatment vs those without a decline or an increase in CEA levels
| Characteristic1 | Overall N=1002 | ≤20% N=402 | >20% N=602 | P 3 | ||||
|---|---|---|---|---|---|---|---|---|
| Age, years | 34 (25 – 45) | 34 (26 – 46) | 34 (24 – 43) | 0.58 | ||||
| Female sex | 45/100 (45%) | 22/40 (55%) | 23/60 (38%) | 0.10 | ||||
| Duration of asthma, years | 12 (4 – 20) | 13 (3 – 20) | 12 (5 – 20) | 0.71 | ||||
| Hemoptysis | 29/100 (29%) | 9/40 (23%) | 20/60 (33%) | 0.24 | ||||
| Brownish black mucus plug | 25/100 (25%) | 8/40 (20%) | 17/60 (28%) | 0.35 | ||||
| FEV1, L | 2.06 (1.33 – 2.65) | 1.71 (1.22 – 2.69) | 2.17 (1.53 – 2.60) | 0.30 | ||||
| FVC, L | 2.67 (2.18 – 3.64) | 2.46 (1.97 – 3.38) | 2.89 (2.28 – 3.79) | 0.12 | ||||
| FEV1/FVC | 66 (57 – 78) | 68 (60 – 82) | 66 (56 – 77) | 0.41 | ||||
| Abnormal chest radiograph | 47/94 (50%) | 15/38 (39%) | 32/56 (57%) | 0.093 | ||||
| Bronchiectasis on CT | 85/97 (88%) | 28/37 (76%) | 57/60 (95%) | 0.009 | ||||
| High attenuation mucus | 29/85 (34%) | 7/28 (25%) | 22/57 (39%) | 0.21 | ||||
| Blood eosinophil count, cells/µL | 700 (421 – 1,075) | 595 (240 – 800) | 850 (500 – 1,240) | 0.005 | ||||
| A. fumigatus- IgE, kUA/L | 22.00 (9.28 – 44.20) | 15.25 (5.73 – 45.65) | 27.15 (18.45 – 39.65) | 0.066 | ||||
| A. fumigatus- IgG, mgA/L | 111.50 (44.45 – 179.00) | 67.50 (31.80 – 128.00) | 150.00 (76.00 – 197.00) | 0.002 | ||||
| Serum total IgE, IU/mL | 4,655 (2,414 – 8,341) | 2,677 (1,695 – 5,756) | 5,951 (3,438 – 9,073) | <0.001 | ||||
| Baseline serum CEA, ng/mL | 5.44 (3.32 – 10.59) | 3.53 (2.37 – 4.52) | 8.24 (5.44 – 12.94) | <0.001 | ||||
| Baseline serum periostin, ng/mL | 21.22 (14.18 – 33.48) | 20.06 (13.46 – 36.65) | 21.64 (15.86 – 33.06) | 0.75 | ||||
| Decline in periostin by >20% | 44/99 (44%) | 14/40 (35%) | 30/59 (51%) | 0.12 |
1ABPA - allergic bronchopulmonary aspergillosis; CEA - Carcinoembryonic antigen; CT - computed tomography; FEV1 - Forced expiratory volume in one second; FVC – forced vital capacity. 2n/N (%); Median (IQR). 3Fisher’s exact test; Pearson’s Chi-squared test; Wilcoxon rank sum test
Table S4.
ABPA subjects with a decline in periostin (>20%) after treatment vs those without a decline (or an increase)
| Characteristic1 | Overall N=992 | ≤20% N=552 | >20% N=442 | P 3 | ||||
|---|---|---|---|---|---|---|---|---|
| Age, years | 34 (25 – 45) | 31 (25 – 42) | 35 (25 – 46) | 0.35 | ||||
| Female sex | 45/99 (45%) | 29/55 (53%) | 16/44 (36%) | 0.10 | ||||
| Duration of asthma, years | 12 (4 – 20) | 11 (4 – 20) | 12 (4 – 20) | 0.91 | ||||
| Hemoptysis | 29/99 (29%) | 17/55 (31%) | 12/44 (27%) | 0.69 | ||||
| Brownish black mucus plug | 25/99 (25%) | 16/55 (29%) | 9/44 (20%) | 0.33 | ||||
| ATT in the past | 24/99 (24%) | 12/55 (22%) | 12/44 (27%) | 0.53 | ||||
| FEV1, L | 2.06 (1.33 – 2.65) | 2.08 (1.35 – 2.79) | 2.06 (1.32 – 2.45) | 0.65 | ||||
| FVC, L | 2.67 (2.18 – 3.64) | 2.69 (2.21 – 3.64) | 2.67 (2.15 – 3.70) | 0.87 | ||||
| FEV1/FVC | 66 (57 – 78) | 72 (60 – 81) | 65 (52 – 74) | 0.12 | ||||
| Abnormal chest radiograph | 47/93 (51%) | 29/52 (56%) | 18/41 (44%) | 0.26 | ||||
| Bronchiectasis on CT | 85/96 (89%) | 49/53 (92%) | 36/43 (84%) | 0.21 | ||||
| High attenuation mucus | 29/85 (34%) | 16/49 (33%) | 13/36 (36%) | 0.74 | ||||
| Blood eosinophil count, cells/µL | 700 (420 – 1,090) | 600 (300 – 1,000) | 790 (495 – 1,200) | 0.14 | ||||
| A. fumigatus-IgE, kUA/L | 21.90 (9.00 – 44.10) | 22.60 (9.56 – 45.80) | 21.90 (7.62 – 36.90) | 0.53 | ||||
| A. fumigatus-IgG, mgA/L | 113.00 (44.10 – 181.00) | 122.00 (64.20 – 171.00) | 96.00 (33.20 – 187.00) | 0.83 | ||||
| Serum total IgE, IU/mL | 4,382 (2,406 – 8,420) | 5,470 (2,406 – 9,139) | 4,340 (2,398 – 8,028) | 0.52 | ||||
| Baseline serum CEA, ng/mL | 5.40 (3.31 – 10.63) | 4.83 (2.83 – 10.70) | 5.69 (3.44 – 10.24) | 0.78 | ||||
| Decline in CEA by>20% | 59/99 (60%) | 29/55 (53%) | 30/44 (68%) | 0.12 | ||||
| Baseline serum periostin, ng/mL | 21.24 (14.09 – 33.53) | 17.54 (11.92 – 23.19) | 30.56 (20.84 – 41.67) | <0.001 |
1ABPA - allergic bronchopulmonary aspergillosis; CEA - Carcinoembryonic antigen; CT - computed tomography; FEV1 - Forced expiratory volume in one second; FVC – forced vital capacity. 2n/N (%); Median (IQR). 3Fisher’s exact test; Pearson’s Chi-squared test; Wilcoxon rank sum test
Subgroup analysis: CEA decline greater in bronchiectasis and high BEC
The percentage change in CEA levels was significantly greater in subjects with bronchiectasis than in those without [Table 3]. The percentage change in serum periostin or total IgE levels was not significantly different in bronchiectasis [Table 3] or in those with HAM [Table 4]. In subjects with BEC ≥ 1000 cells/µL (n = 37), baseline serum CEA levels and the percentage decline after treatment were significantly greater than in those with BEC <1000 cells/µL [Table S5]. However, there were no statistically significant differences in serum periostin, CEA, or total IgE levels in subjects with BEC ≥500 vs <500 cells/µL at enrolment [Table S6].
Table 3.
Immunological parameters at baseline and two months in ABPA patients with and without bronchiectasis
| Characteristic1 | Overall n=1092 | Bronchiectasis n=942 | No bronchiectasis n=152 | P 3 | ||||
|---|---|---|---|---|---|---|---|---|
| Serum total IgE, IU/mL | 4,382 (2,422–8,558) | 5,519 (2,642–9,168) | 2,374 (1,825–4,353) | 0.018 | ||||
| Serum total IgE at 2 months, IU/mL | 3,202 (1,646–6,385) | 3,713 (1,942–6,668) | 1,687 (995–3,605) | 0.068 | ||||
| Percentage change in serum total IgE | 32 (-5–54) | 32 (-5–55) | 30 (-6–51) | 0.78 | ||||
| Serum CEA at baseline, ng/mL | 5.03 (3.41–10.54) | 6.13 (3.82–11.40) | 2.75 (1.94–3.51) | <0.001 | ||||
| Serum CEA at 2 months, ng/mL | 3.64 (2.63–5.22) | 4.12 (2.68–5.68) | 2.77 (2.31–3.68) | 0.10 | ||||
| Percent change in CEA | 34 (-12–63) | 41 (2–66) | -22 (-38–21) | 0.012 | ||||
| Serum periostin at baseline, ng/mL | 21.99 (15.18–36.05) | 22.03 (16.16–34.66) | 17.68 (14.09–41.78) | 0.94 | ||||
| Serum periostin at 2 months, ng/mL | 17.65 (12.51–25.05) | 17.74 (13.25–24.93) | 14.35 (10.05–27.34) | 0.47 | ||||
| Percent change in periostin | 16 (-41–49) | 15 (-39–48) | 34 (-91–65) | 0.50 |
1ABPA - allergic bronchopulmonary aspergillosis; CEA - Carcinoembryonic antigen. 2Median (IQR). 3Wilcoxon rank-sum test
Table 4.
Immunological parameters at baseline and 2 months in ABPA patients with and without HAM
| Characteristic1 | HAM n=312 | No HAM n=632 | P 3 | |||
|---|---|---|---|---|---|---|
| Serum total IgE, IU/mL | 7,186 (3,446–9,006) | 4,382 (2,355–9,255) | 0.20 | |||
| Serum total IgE at 2 months, IU/mL | 4,280 (2,055–7,183) | 3,634 (1,906–6,458) | 0.37 | |||
| Percent change in serum total IgE | 40 (12–55) | 19 (-8–56) | 0.14 | |||
| Serum CEA at baseline, ng/mL | 7.66 (4.55–13.02) | 4.83 (3.46–10.54) | 0.037 | |||
| Serum CEA at 2 months, ng/mL | 3.64 (2.07–6.02) | 4.14 (2.86–5.40) | 0.84 | |||
| Percent change in CEA | 53 (21–74) | 32 (1–60) | 0.13 | |||
| Serum periostin at baseline, ng/mL | 25.10 (16.99–34.66) | 21.24 (13.86–35.10) | 0.41 | |||
| Serum periostin at 2 months, ng/mL | 19.37 (16.03–26.77) | 16.86 (11.41–24.09) | 0.16 | |||
| Percent change in periostin | 16 (-39–50) | 15 (-32–44) | 0.98 |
1ABPA - allergic bronchopulmonary aspergillosis; CEA - Carcinoembryonic antigen; HAM - high attenuation mucus. 2Median (IQR). 3Wilcoxon rank-sum test
Table S5.
Immunological parameters at baseline and two months in ABPA patients based on blood eosinophil counts at enrollment
| Characteristics1 | Eosinophil count ≥1000 (n=37)2 | Eosinophil count <1000 (n=75)2 | P 3 | |||
|---|---|---|---|---|---|---|
| Serum total IgE, IU/mL | 7,186 (3,562, 13,061) | 4,131 (2,230, 7,695) | 0.004 | |||
| Serum total IgE at 2 months, IU/mL | 4,120 (2,144, 8,330) | 3,137 (1,258, 5,621) | 0.080 | |||
| Percent change in serum total IgE | 30 (-4, 53) | 34 (-4, 58) | 0.9 | |||
| Serum CEA at baseline, ng/mL | 7.83 (3.86, 11.85) | 4.49 (3.15, 7.66) | 0.018 | |||
| Serum CEA at 2 months, ng/mL | 3.47 (2.17, 4.85) | 3.73 (2.75, 5.48) | 0.4 | |||
| Percent change in CEA | 58 (34, 75) | 15 (-25, 51) | <0.001 | |||
| Serum periostin at baseline, ng/mL | 23.01 (13.47, 36.05) | 21.99 (15.18, 35.10) | 0.8 | |||
| Serum periostin at 2 months, ng/mL | 16.54 (12.06, 21.34) | 17.74 (11.87, 25.85) | 0.3 | |||
| Percent change in periostin | 22 (-8, 52) | 14 (-44, 48) | - |
1ABPA - allergic bronchopulmonary aspergillosis; CEA - Carcinoembryonic antigen. 2Median (Q1, Q3). 3Wilcoxon rank sum test
Table S6.
Immunological parameters at baseline and two months in ABPA patients based on blood eosinophil count
| Characteristics1 | Eosinophil count ≥500 cells/µL (n=76)2 | Eosinophil count <500 cells/µL (n=36)2 | P 3 | |||
|---|---|---|---|---|---|---|
| Serum total IgE, IU/mL | 4,997 (2,398 – 9,212) | 3,901 (2,430 – 7,737) | 0.24 | |||
| Serum total IgE at 2 months, IU/mL | 3,643 (1,949 – 6,999) | 3,128 (988 – 4,622) | 0.15 | |||
| Percent change in serum total IgE | 28 (-9 – 54) | 35 (9 – 59) | 0.34 | |||
| Serum CEA at baseline, ng/mL | 5.94 (2.86 – 11.51) | 4.60 (3.50 – 7.41) | 0.40 | |||
| Serum CEA at 2 months, ng/mL | 3.64 (2.62 – 5.28) | 3.96 (2.82 – 5.22) | 0.76 | |||
| Percent change in CEA | 38 (-2 – 66) | 12 (-36 – 55) | 0.072 | |||
| Serum periostin at baseline, ng/mL | 22.80 (16.30 – 36.38) | 20.47 (12.84 – 34.36) | 0.36 | |||
| Serum periostin at 2 months, ng/mL | 17.58 (12.16 – 24.46) | 17.67 (10.95 – 26.77) | 0.71 | |||
| Percent change in periostin | 19 (-25 – 51) | 12 (-76 – 48) | - |
1ABPA - allergic bronchopulmonary aspergillosis; CEA - Carcinoembryonic antigen. 2Median (IQR). 3Wilcoxon rank sum test
DISCUSSION
We found elevated serum CEA and periostin levels in patients with ABPA compared to those with asthma. However, the diagnostic performance of serum periostin and CEA in differentiating asthma from ABPA was modest. While both biomarkers showed a significant decline after 2 months of treatment, their values increased in almost a third of the patients. To our knowledge, this is the largest study evaluating these biomarkers in ABPA.
Serum periostin has been linked to airway eosinophilia in asthma, with studies suggesting that periostin may be a more stable marker of eosinophilic inflammation than peripheral blood eosinophils, even after glucocorticoid therapy.[7,15] Periostin gene expression in the respiratory epithelium is upregulated by the type 2 cytokines IL-4 and IL-13, making it a marker of type 2 inflammation.[15,16] Asthmatic subjects with a high periostin level have been shown to respond better to treatment in the lebrikizumab trial (anti-IL-13 antibody).[17] They also have a higher exacerbation rates upon withdrawal of inhaled corticosteroids.[18] However, despite a positive association between serum periostin and asthma severity or airflow limitation in several studies, the evidence remains inconsistent.[19,20,21,22,23]
ABPA is a phenotype of severe asthma, where eosinophils are central to pathogenesis. Therefore, we anticipated that serum periostin could be a reliable marker to differentiate ABPA from asthma. Previous studies investigating periostin in ABPA are limited.[7,9] A study from Japan reported significantly higher median serum periostin levels in ABPA (107 ng/mL) compared to severe asthma (72 ng/mL), with an AUROC of 0.75 for distinguishing ABPA from Aspergillus-sensitised asthma. Notably, in the subgroup of patients receiving oral glucocorticoids, the study found serum periostin better in differentiating ABPA (n = 20) from Aspergillus-sensitised severe asthma (n = 12) than serum total IgE or BEC.[7] Another study found an AUROC of 0.81 for serum periostin to differentiate ABPA (n = 20; median value: 85.5 ng/mL) from SAFS (n = 11; median level: 50.9 ng/mL). In addition, the baseline serum periostin correlated with IL-5 levels and serum total IgE. However, when 12 patients of ABPA were followed up longitudinally, there was no significant decline in serum periostin levels following treatment.[9] In contrast to the previous data, our study found a weak correlation between CEA and periostin or BEC.
In our cohort, while median serum periostin levels decreased with treatment, a substantial proportion of patients experienced increased levels after 2 months of therapy. Notably, our study participants’ median serum periostin levels in asthma and ABPA were lower than those reported in Japanese and Chinese studies. Ethnic differences may explain these discrepancies, as previous research on asthma patients in India also reported lower periostin levels (mean: 45.2 ng/mL).[24] Other factors, including the age of the patients, assay technique, and the presence of alternative periostin sources, could have contributed to these variations across studies.
Serum CEA, although traditionally associated with malignancies such as colorectal, breast, and ovarian cancers, has been identified as elevated in ABPA, likely secreted from eosinophils.[25,26] In a recent study from China, immunohistochemistry showed CEA co-localised with eosinophils in lung tissue from ABPA patients.[25] Further, there was a correlation between serum CEA and BEC. These studies also reported an improvement in CEA levels following ABPA treatment.[26] Even in the current study, CEA outperformed periostin; nevertheless, it still was an imperfect marker.
Our study has a few limitations. It was conducted at a single centre, and we did not measure other type 2 cytokines. Many participants were enrolled during the COVID-19 pandemic, leading to incomplete follow-up data on several variables, including spirometry, asthma control, and BEC. Consequently, we could not evaluate the correlation between eosinophil count decline or asthma control and serum CEA or periostin levels after treatment. Additionally, all ABPA patients were treated with oral prednisolone; thus, the results may differ with alternative treatments, such as itraconazole or biologics. Lastly, we were unable to determine whether patients with persistently elevated periostin or CEA levels after 2 months of treatment had a higher risk of future exacerbations.
In conclusion, serum CEA and periostin are elevated in ABPA but show modest diagnostic performance in differentiating it from asthma. Future studies should investigate their utility as markers of treatment response, particularly in patients receiving Th2-directed biologic therapy.
Ethics committee approval
NK/7631/MD-625, 14-September-2021, Postgraduate Institute of Medical Education and Research, Chandigarh, India.
Author contribution statement
Conceptualisation: VM, RA,
Data curation: RS, PK
Formal Analysis: VM
Funding Acquisition: RA, AP
Investigation: RS, VM, PK, SD, ISS, KTP, SMR, AP, ANA, RA
Methodology: RS, VM, PK, SMR, MG, ANA, RA
Project administration: VM, SMR, ANA, RA
Resources: VM, RA, AP
Software: RA
Validation: RS, VM, SMR, AP, RA
Visualisation: VM, RA
Writing Original draft: RS, VM
Writing, review, and editing: RS, VM, PK, AP, SD, ISS, KTP, SMR, ANA, RA.
Data availability statement:
Data available from the corresponding author on reasonable request.
Conflicts of interest
There are no conflicts of interest.
Correlogram showing a weak correlation (r, marked within the colored squares) between CEA, periostin, serum total IgE, and blood eosinophil count (BEC). Green indicates positive, and blue indicates negative correlation
Funding Statement
Nil.
REFERENCES
- 1.Agarwal R, Muthu V, Sehgal IS. Clinical manifestation and treatment of allergic bronchopulmonary aspergillosis. Semin Respir Crit Care Med. 2024;45:114–27. doi: 10.1055/s-0043-1776912. [DOI] [PubMed] [Google Scholar]
- 2.Agarwal R, Muthu V, Sehgal IS. Relationship between aspergillus and asthma. Allergol Int. 2023;72:507–20. doi: 10.1016/j.alit.2023.08.004. [DOI] [PubMed] [Google Scholar]
- 3.Agarwal R, Devi D, Gupta D, Chakrabarti A. A questionnaire-based study on the role of environmental factors in allergic bronchopulmonary aspergillosis. Lung India. 2014;31:232–6. doi: 10.4103/0970-2113.135762. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Agarwal R, Muthu V, Sehgal IS, Dhooria S, Prasad KT, Soundappan K, et al. Prevalence of aspergillus sensitization and allergic bronchopulmonary aspergillosis in bronchial asthma: A systematic review of Indian studies. Lung India. 2023;40:527–36. doi: 10.4103/lungindia.lungindia_69_23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Agarwal R, Sehgal IS, Muthu V, Denning DW, Chakrabarti A, Soundappan K, et al. Revised ISHAM-ABPA working group clinical practice guidelines for diagnosing, classifying and treating allergic bronchopulmonary aspergillosis/mycoses. Eur Respir J. 2024;63:2400061. doi: 10.1183/13993003.00061-2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Venegas Garrido C, Mukherjee M, Svenningsen S, Nair P. Eosinophil-mucus interplay in severe asthma: Implications for treatment with biologicals. Allergol Int. 2024;73:351–61. doi: 10.1016/j.alit.2024.03.001. [DOI] [PubMed] [Google Scholar]
- 7.Tanaka J, Hebisawa A, Oguma T, Tomomatsu K, Suzuki J, Shimizu H, et al. Evaluating serum periostin levels in allergic bronchopulmonary aspergillosis. Allergy. 2020;75:974–7. doi: 10.1111/all.14114. [DOI] [PubMed] [Google Scholar]
- 8.Muthu V, Sehgal IS, Prasad KT, Dhooria S, Garg M, Aggarwal AN, et al. Epidemiology and outcomes of allergic bronchopulmonary aspergillosis in the elderly. Mycoses. 2022;65:71–8. doi: 10.1111/myc.13388. [DOI] [PubMed] [Google Scholar]
- 9.Wu G, Meng X, Zheng P, Zhang XD, Li L, Hu H, et al. Elevated serum levels of periostin in patients with allergic bronchopulmonary aspergillosis. Mycoses. 2019;62:780–9. doi: 10.1111/myc.12957. [DOI] [PubMed] [Google Scholar]
- 10.Agarwal R, Sehgal IS, Muthu V, Dhooria S, Prasad KT, Aggarwal AN, et al. Long-term follow-up of allergic bronchopulmonary aspergillosis treated with glucocorticoids: A study of 182 subjects. Mycoses. 2023;66:953–9. doi: 10.1111/myc.13640. [DOI] [PubMed] [Google Scholar]
- 11.Agarwal R, Sehgal IS, Muthu V. The 2024 International Society for Human and Animal Mycology (ISHAM)-ABPA working group guidelines: Transforming diagnosis and management of ABPA. Lung India. 2025;42:79–83. doi: 10.4103/lungindia.lungindia_28_25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Aggarwal AN, Agarwal R, Dhooria S, Prasad KT, Sehgal IS, Muthu V, et al. Joint Indian Chest Society-National College of Chest Physicians (India) guidelines for spirometry. Lung India. 2019;36:S1–35. doi: 10.4103/lungindia.lungindia_300_18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Sjoberg DD, Whiting K, Curry M, Lavery JA, Larmarange J. Reproducible summary tables with the gtsummary Package. R Journal. 2021;13:570–94. [Google Scholar]
- 14.Wickham H, Averick M, Bryan J, Chang W, McGowan L, François R, et al. Welcome to the Tidyverse. J Open Source Softw. 2019;4:1686. [Google Scholar]
- 15.Jia G, Erickson RW, Choy DF, Mosesova S, Wu LC, Solberg OD, et al. Periostin is a systemic biomarker of eosinophilic airway inflammation in asthmatic patients. J Allergy Clin Immunol. 2012;130:647–54. doi: 10.1016/j.jaci.2012.06.025. e10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Baldo DC, Romaldini JG, Pizzichini MMM, Cançado JED, Dellavance A, Stirbulov R. Periostin as an important biomarker of inflammatory phenotype T2 in Brazilian asthma patients. J Bras Pneumol. 2023;49:e20220040. doi: 10.36416/1806-3756/e20220040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hanania NA, Noonan M, Corren J, Korenblat P, Zheng Y, Fischer SK, et al. Lebrikizumab in moderate-to-severe asthma: Pooled data from two randomised placebo-controlled studies. Thora×. 2015;70:748–56. doi: 10.1136/thoraxjnl-2014-206719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kato G, Takahashi K, Izuhara K, Komiya K, Kimura S, Hayashi S. Markers that can reflect asthmatic activity before and after reduction of inhaled corticosteroids: A pilot study. Biomark Insights. 2013;8:97–105. doi: 10.4137/BMI.S12537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kanemitsu Y, Matsumoto H, Izuhara K, Tohda Y, Kita H, Horiguchi T, et al. Increased periostin associates with greater airflow limitation in patients receiving inhaled corticosteroids. J Allergy Clin Immunol. 2013;132:305–12. doi: 10.1016/j.jaci.2013.04.050. e3. [DOI] [PubMed] [Google Scholar]
- 20.Licari A, Brambilla I, Sacchi L, Marseglia G, Ciprandi G. Periostin, type 2 biomarker, is not associated with asthma control grade in asthmatic allergic children. Respir Med. 2019;151:118–20. doi: 10.1016/j.rmed.2019.04.010. [DOI] [PubMed] [Google Scholar]
- 21.Matsusaka M, Kabata H, Fukunaga K, Suzuki Y, Masaki K, Mochimaru T, et al. Phenotype of asthma related with high serum periostin levels. Allergol Int. 2015;64:175–80. doi: 10.1016/j.alit.2014.07.003. [DOI] [PubMed] [Google Scholar]
- 22.Simpson JL, Yang IA, Upham JW, Reynolds PN, Hodge S, James AL, et al. Periostin levels and eosinophilic inflammation in poorly-controlled asthma. BMC Pulm Med. 2016;16:67. doi: 10.1186/s12890-016-0230-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Al-Shaikhly T, Murphy RC, Lai Y, Frevert CW, Debley JS, Ziegler SF, et al. Sputum periostin is a biomarker of type 2 inflammation but not airway dysfunction in asthma. Respirology. 2023;28:491–4. doi: 10.1111/resp.14491. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Emprm V, Rajanandh MG, Nageswari AD. Periostin-A novel systemic biomarker for eosinophilic airway inflammation: A case control study. J Clin Diagn Res. 2016;10:Oc01–4. doi: 10.7860/JCDR/2016/14553.7166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Yang Y, Gao Q, Jin Y, Qi M, Lu G, Hequan Li. Eosinophils may serve as CEA-secreting cells for allergic bronchopulmonary aspergillosis (ABPA) patients. Sci Rep. 2021;11:4025. doi: 10.1038/s41598-021-83470-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Noguchi T, Yamamoto K, Moriyama G, Saito Y, Kyoyama H, Mikami S, et al. Evaluation of serum levels of carcinoembryonic antigen in allergic bronchopulmonary aspergillosis. J Nippon Med Sch. 2013;80:404–9. doi: 10.1272/jnms.80.404. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Correlogram showing a weak correlation (r, marked within the colored squares) between CEA, periostin, serum total IgE, and blood eosinophil count (BEC). Green indicates positive, and blue indicates negative correlation
Data Availability Statement
Data available from the corresponding author on reasonable request.