https://orcid.org/0000-0002-3175-3967
Abstract
Detection of serum-specific immunoglobulin G (sIgG) to Aspergillus fumigatus traditionally relied on precipitin assays, which lack standardization and have poor analytical sensitivity. Automated quantitative immunoassays are now more widely used alternatives. A challenge, however, is determining reference interval (RI) cutoffs indicative of disease presence.
Sera from 152 local healthy donors were tested for Aspergillus fumigatus sIgG using the ImmunoCAP assay to calculate a nonparametric RI cutoff. Results from 178 patient samples cotested by the precipitin and ImmunoCAP assays were analyzed using receiver operator characteristic (ROC) curve to determine an optimal sIgG concentration for precipitin positivity. Clinical information available for 46 patients tested by the ImmunoCAP assay was also used to estimate an optimal sIgG cutoff for pulmonary aspergillosis diagnosis.
Specific-IgG concentration at 81.5 mcg/mL corresponded to the 97.5th percentile of tested healthy donors. The ROC-driven optimal IgG cutoff for precipitin positivity was at 40.4 mcg/mL with 67.8% sensitivity [95% confidence interval (CI): 54.4% to 79.4%%] and 72.3% specificity (95% CI: 63.3% to 80.1%). Using clinical diagnoses, an IgG concentration at 64.7 mcg/mL had optimal sensitivity (77.8%; 95% CI: 61.9% to 88.3%) and specificity (66.7%, 95% CI 39.1% to 86.2%) for pulmonary aspergillosis.
Our healthy donor-driven RI cutoff was higher than estimated optimal sIgG values based on precipitin positivity and disease presence. As fungal sIgG levels can be impacted by local environmental exposures, and given the limited size of our clinical dataset, adopting an assay cutoff based on precipitin results (40.4 mcg/mL) can be more objective.
Quantitative immunoassays for Aspergillus fumigatus sIgG have improved analytical and clinical diagnostic performances compared to traditional precipitin assays. As sIgG presence can indicate normal immune response to environmental exposure or underlying aspergillosis disease, determining cutoffs for assay positivity is challenging. This study highlights the variability in estimated sIgG cutoffs depending on the study method: 81.5 mcg/mL from local healthy donors and 40.4 and 64.7 mcg/mL based on optimal sIgG concentrations for precipitin positivity and pulmonary aspergillosis diagnoses, respectively. Understanding how clinical laboratories establish their positivity cutoffs and emphasizing the nondiagnostic role of sIgG in aspergillosis can be helpful in the interpretation of results.
INTRODUCTION
Aspergillus fumigatus is an ubiquitous saprophytic fungus implicated in various forms of pulmonary disease (1). Aspergillosis is a generic term used to describe almost all opportunistic Aspergillus infections typically acquired by inhalation of the airborne conidia in the environment (1). Depending on host immune response and underlying lung pathology, aspergillosis can range from an inflammatory hypersensitivity reaction in predisposed asthma and cystic fibrosis patients to chronic pulmonary aspergillosis (CPA) phenotypes in immunocompetent or mildly immunocompromised patients. Structural lung damage, as occurs in tuberculosis patients, is a strong predisposing factor for CPA (2, 3). CPA can progress to fatal necrotizing invasive aspergillosis in immunocompromised subjects (2–5). Acute invasive disease can also occur in immunosuppressed patients as in chemotherapy-induced neutropenia or post bone marrow transplant states (5). Although the prevalence of CPA in the United States is relatively low (<1 in 100 000), the disease has high morbidity and mortality rates (6). CPA is also likely underdiagnosed in the United States (6). The poor sensitivity and specificity of Aspergillus cultures from sputum and bronchial lavage specimens have led to reliance on the detection of antibody responses to Aspergillus along with clinical and radiologic features for allergic bronchopulmonary aspergillosis (ABPA) and CPA diagnoses (4, 7–9).
Detection of IgG antibodies has historically relied on qualitative gel immunoprecipitation techniques that use crude extracts of cultured Aspergillus conidia and hyphae (10). This methodology has limitations with respect to assay sensitivity, specificity, interpretation subjectivity, and complexity. Multiple studies have compared the performances of precipitin assays to modern automated immunoassay formats used to measure Aspergillus fumigatus serum-specific IgG (sIgG) concentrations (11–13). A common challenge of such quantitative immunoassays is to determine the appropriate reference interval (RI) cutoffs given the presence of varying antibody titers in disease-free subjects induced by environmental exposures. Using precipitin assays as the reference method, a subsequent challenge was the discrepancies between the qualitative precipitin results and the measured Aspergillus-specific IgG concentrations given the different antigen preparations used in both assays.
In this study, our aim was to determine Aspergillus fumigatus sIgG RI or positivity cutoffs to distinguish between disease and healthy subjects using the Phadia ImmunoCAP immunoassay. We performed a RI study using sera collected from local healthy donors. Results of specimens cotested by the precipitin and ImmunoCAP assays were also used to estimate an optimal sIgG cutoff based on precipitin positivity.
MATERIALS AND METHODS
Residual Clinical Samples and Data Retrieval
Results of a total of 178 samples submitted to our reference laboratory were retrieved in accordance with institutional guidelines (University of Utah IRB studies 7275, 82990, and 00159361). These samples are comprised of (a) 57 patient samples tested for Aspergillus fumigatus antibodies using the precipitin assay (residual sera from these samples were used for additional testing using the ImmunoCAP™ assay) and (b) results of additional 121 unique clinical samples cotested for Aspergillus fumigatus-specific IgG using ImmunoCAP and precipitin assays (ordered by treating providers). Clinical information was available for only 2 of the 178 patients.
To correlate ImmunoCAP measurements with clinical diagnoses, Aspergillus fumigatus sIgG results of an additional 44 patients with available clinical information were retrieved.
Immunoprecipitation Assay
To detect the presence of Aspergillus fumigatus-specific antibodies in patient sera, an Ouchterlony double immunodiffusion was performed. Serum was loaded into a center well, and different Aspergillus fumigatus antigenic extracts were added to surrounding wells. The formation of a visible precipitin line between the sample well and the opposing antigenic extracts is a positive reaction and indicates antigen-specific antibody presence. Aspergillus fumigatus extracts from 4 different strains were obtained from Stallergenes Greer. Extracts #1, #2, and #3 were prepared with Aspergillus fumigatus strains 1–102, ATCC 1022, and 507-PSIFD. Broth from cultures (grown for 28–32 days) was sterile filtered, dialyzed, vacuum desiccated, and reconstituted in sterile saline. Extract #6 was a cellular extract and was prepared using Aspergillus fumigatus strain Reed 6–666.
Quantitative Aspergillus fumigatus sIgG Immunoassay
Testing was performed using the Aspergillus fumigatus ImmunoCAP (m3) with IgG-specific conjugate on a Phadia 250 analyzer (ThermoFisher Scientific).
Aspergillus fumigatus sIgG RI Study
For the sIgG RI study, sera were collected from 152 self-proclaimed healthy individuals. Exclusion criteria included prescription medications, with the exception of birth control or hormone replacement therapy, and no autoimmune disease or current infections. Sera were tested for sIgG to Aspergillus fumigatus using the ImmunoCAP assay. Data were analyzed in Prism software (GraphPad Software). Outlier exclusion was performed using the Prism robust regression and outlier removal function (14).
RESULTS
Determination of Aspergillus fumigatus sIgG RI Cutoff Using the ImmunoCAP Assay
Sera from 152 local heathy donors were tested for Aspergillus fumigatus sIgG using the ImmunoCAP assay. Of the results obtained, 13 samples were statistically excluded as outliers using the robust regression and outlier removal method (14). Measured sIgG concentration was >70 mcg/mL in 23 samples. Twelve of the 23 samples with enough residual volumes were tested for precipitin formation using the 4 Aspergillus strain extracts (1, 2, 3, and 6) and were negative. The nonparametric RI cutoff at the 97.5th percentile corresponded to the sIgG level of 81.5 mcg/mL (Fig. 1A).
(A) Aspergillus fumigatus sIgG levels (mcg/mL) in healthy donors measured by the ImmunoCAP method. Dotted lines represent 2.5th (lower) and 97.5th (upper) percentiles. Triangle symbols are excluded outliers (13 donors). (B) Bar chart of sIgG levels in patients with or without Aspergillus-related pulmonary disease. Error bars indicate mean and SD. (C) ROC of sIgG levels based on presence or absence of Aspergillus fumigatus related pulmonary disease. (D) Line graph of specificity (open circles) and sensitivity (closed circles) calculated at different sIgG cutoffs from the ROC analysis. Vertical line is at 64.7 mcg/mL cutoff of highest sensitivity and specificity for Aspergillus fumigatus-related disease.
Clinical information was available for 46 unique patients tested for Aspergillus fumigatus sIgG by the ImmunoCAP method. Eleven of 46 patients had clinical findings or diagnoses of Aspergillosis-related lung disease: ABPA (7 patients), aspergilloma (2 patients), and CPA (2 patients) (Fig. 1B). Using receiver operating characteristic (ROC) analyses of sIgG levels based on presence or absence of Aspergillus fumigatus-related pulmonary disease, a cutoff of 64.7 mcg/mL resulted in the highest estimated sensitivity and specificity for disease presence: 77.8%; 95% confidence interval (CI): 61.9% to 88.3% and 66.7%; 95% CI: 39.1% to 86.2%, respectively (Fig. 1C and 1D). The 81.5 mcg/mL cutoff determined by our healthy donor study was of higher estimated specificity and lower sensitivity (Fig. 1D).
ImmunoCAP sIgG Cutoff Estimation Based on the Precipitin Assay
To correlate the precipitin qualitative results (i.e., positive or negative) to the ImmunoCAP sIgG assay and to estimate a cutoff for assay positivity, residual sera from 57 patients tested by the precipitin method (27 tested for the 4 strains and 30 tested for strains #1 and #6) were tested in the study using the ImmunoCAP assay. All precipitin-negative samples were tested using the 4 Aspergillus fumigatus strain extracts. Of these 57 sera, 38 were positive for one or more of the tested extracts (Fig. 2A). Seventeen of these 38 precipitin-positive samples had sIgG higher than the healthy donor RI cutoff, 81.5 mcg/mL (range 81.8 to >200 mcg/mL). Twenty-one precipitin-positive samples were lower than 81.5 mcg/mL (Table 1; range 8.4–76.7 mcg/mL; Fig. 2A). Only 1 of 19 precipitin-negative samples was of sIgG concentration >81.5 mcg/mL (range 4.7-129 mcg/mL).
Aspergillus fumigatus sIgG levels in precipitin positive vs negative sera. (A) Dot plot of 57 residual precipitin sera tested by the ImmunoCAP method. (B) Dot plot of 121 clinical patient samples cotested by the precipitin and ImmunoCAP assays. Y axes are sIgG levels (mcg/mL). Horizontal lines indicate median values of each group. Horizontal dotted lines at 81.5 and 40.4 mcg/dL indicate 97.5th RI value obtained from the healthy donor study and the optimal cutoff based on precipitin assay, respectively. P values are of unpaired Student t-test. (C) ROC of sIgG levels (178 samples) in precipitin positive vs negative samples. (D) Line graph of specificity and sensitivity calculated at different sIgG cutoffs from the ROC analysis. Vertical line is at 40.4 mcg/mL cutoff of highest sensitivity and specificity for positivity by the precipitin assay.
Numbers of precipitin positive and negative samples above and below estimated sIgG cutoff RI cutoff from healthy donor study.
| Precipitin positive | Precipitin negative | |||
|---|---|---|---|---|
| RI cutoff (mcg/mL) | <81.5 | ≥81.5 | <81.5 | ≥81.5 |
| Residual specimens | 21 | 17 | 18 | 1 |
| Cotested specimens | 12 | 9 | 91 | 9 |
| Total | 59 | 119 | ||
| Precipitin positive | Precipitin negative | |||
|---|---|---|---|---|
| RI cutoff (mcg/mL) | <81.5 | ≥81.5 | <81.5 | ≥81.5 |
| Residual specimens | 21 | 17 | 18 | 1 |
| Cotested specimens | 12 | 9 | 91 | 9 |
| Total | 59 | 119 | ||
Numbers of precipitin positive and negative samples above and below estimated sIgG cutoff RI cutoff from healthy donor study.
| Precipitin positive | Precipitin negative | |||
|---|---|---|---|---|
| RI cutoff (mcg/mL) | <81.5 | ≥81.5 | <81.5 | ≥81.5 |
| Residual specimens | 21 | 17 | 18 | 1 |
| Cotested specimens | 12 | 9 | 91 | 9 |
| Total | 59 | 119 | ||
| Precipitin positive | Precipitin negative | |||
|---|---|---|---|---|
| RI cutoff (mcg/mL) | <81.5 | ≥81.5 | <81.5 | ≥81.5 |
| Residual specimens | 21 | 17 | 18 | 1 |
| Cotested specimens | 12 | 9 | 91 | 9 |
| Total | 59 | 119 | ||
Results of an additional set of clinical samples (121) that were cotested for precipitin and ImmunoCAP Aspergillus fumigatus sIgG (ordered by treating providers) were retrieved for analysis. sIgG levels were significantly higher in precipitin-positive (21 samples) than -negative (100 samples) (P < 0.0001; unpaired Student's t-test; Fig. 2B). Nine of the 21 precipitin-positive (43%) and 9 of 100 precipitin-negative (9%) samples had higher sIgG than the determined healthy donor RI cutoff (81.5 mcg/mL).
Using ROC analyses of all available precipitin-positive (59 samples) and -negative (119 samples) results to determine an optimal cutoff, sIgG concentration of 40.4 mcg/mL achieved the highest sensitivity (67.8%; 95% Cl: 54.4% to 79.4%) and specificity (72.3%; 95% CI: 63.3% to 80.1%) for precipitin positivity (Fig. 2C and 2D; Table 2). Applying the cutoff obtained from the healthy donors study (81.5 mcg/mL), specificity was higher (91.6%; 95% CI: 85.1% to 95.9%), but sensitivity was lower (44.1%; 95% CI: 31.2% to 57.6%) for precipitin positivity (Table 2).
Analytical performance of determined ImmunoCAP assay cutoffs applied to 178 precipitin samples, including 59 positive and 119 negative results.
| ImmunoCAP cutoff (mcg/mL) | % Sensitivity (95% confidence interval) | % Specificity (95% confidence interval) | % PPV | % NPV |
|---|---|---|---|---|
| 40.4 | 67.8 (54.4–79.4) | 72.3 (63.3–80.1) | 54.8 | 81.9 |
| 64.7 | 50.9 (37.5–64.1) | 86.6 (79.1–92.1) | 65.2 | 78.0 |
| 81.5 | 44.1 (31.2–57.6) | 91.6 (85.1–95.9) | 72.2 | 76.8 |
| ImmunoCAP cutoff (mcg/mL) | % Sensitivity (95% confidence interval) | % Specificity (95% confidence interval) | % PPV | % NPV |
|---|---|---|---|---|
| 40.4 | 67.8 (54.4–79.4) | 72.3 (63.3–80.1) | 54.8 | 81.9 |
| 64.7 | 50.9 (37.5–64.1) | 86.6 (79.1–92.1) | 65.2 | 78.0 |
| 81.5 | 44.1 (31.2–57.6) | 91.6 (85.1–95.9) | 72.2 | 76.8 |
Abbreviations: NPV, negative predictive value; PPV, positive predictive value.
Analytical performance of determined ImmunoCAP assay cutoffs applied to 178 precipitin samples, including 59 positive and 119 negative results.
| ImmunoCAP cutoff (mcg/mL) | % Sensitivity (95% confidence interval) | % Specificity (95% confidence interval) | % PPV | % NPV |
|---|---|---|---|---|
| 40.4 | 67.8 (54.4–79.4) | 72.3 (63.3–80.1) | 54.8 | 81.9 |
| 64.7 | 50.9 (37.5–64.1) | 86.6 (79.1–92.1) | 65.2 | 78.0 |
| 81.5 | 44.1 (31.2–57.6) | 91.6 (85.1–95.9) | 72.2 | 76.8 |
| ImmunoCAP cutoff (mcg/mL) | % Sensitivity (95% confidence interval) | % Specificity (95% confidence interval) | % PPV | % NPV |
|---|---|---|---|---|
| 40.4 | 67.8 (54.4–79.4) | 72.3 (63.3–80.1) | 54.8 | 81.9 |
| 64.7 | 50.9 (37.5–64.1) | 86.6 (79.1–92.1) | 65.2 | 78.0 |
| 81.5 | 44.1 (31.2–57.6) | 91.6 (85.1–95.9) | 72.2 | 76.8 |
Abbreviations: NPV, negative predictive value; PPV, positive predictive value.
DISCUSSION
Inhalation of environmental fungal conidia is typically nonpathogenic and does not lead to inflammatory innate immune reactions in immunocompetent and nonatopic hosts. This immunologic inertness of dormant conidia has been attributed to a surface rodlet hydrophobic layer that shields many immunogenic antigens from macrophages and neutrophils (15). Although germinating conidia can be cleared by bronchial epithelium and interacting immune cells, airway and tissue colonization can still occur in immune-competent hosts (16). The formation of a specific antibody response to Aspergillus species occurs in healthy individuals, and whether this represents a subclinical disease state or age-dependent acquired immunity is not clear (17).
As precipitin-based assays are cumbersome, lack sensitivity, and use antigen extract preparations that lack standardization, clinical labs have transitioned to other more objective and practical immunoassays (18). The ImmunoCAP assay is widely used by clinical laboratories and has been demonstrated to have better analytical and diagnostic performance (11, 12, 19). A common challenge for quantitative assays is to determine a cutoff for assay positivity, given the overlap in antibody levels between healthy individuals with and without colonization. Cutoff determination can be based on either RI studies of healthy donors or, more commonly, statistical analyses (e.g., ROC) to obtain decision limits that achieve optimal sensitivity and specificity in differentiating between disease and control subjects (20). Using nonparametric RI methods, cutoffs based on healthy donor cohorts were at 66.5 (21), 68.7 (22), and up to 136 mcg/mL (23) by the ImmunoCAP assay. Using ROC analyses, 50 mcg/mL was an optimal cutoff established by multiple studies of CPA (24, 25), including a study of biopsy-proven aspergillosis cases (26). Lower (40.5 mcg/mL) optimal cutoffs were also found by other studies (27). For ABPA, 26.9, 78, and 90 mcg/mL were calculated optimal ROC cutoffs for diagnoses (28–30). Variability in sIgG cutoff concentrations can be attributed in part to differences in fungal environmental exposure, ethnicity, and geographic region (31). In the recently revised European ABPA diagnostic criteria, it is recommended to establish population-specific cutoffs using lateral flow or enzyme immunoassays (32). In our local healthy donor RI study, despite excluding 13 results as outliers, the estimated cutoff was relatively higher (81.5 mcg/mL) than reported in the literature. Interestingly, when we compared antibody titers of clinical samples submitted to our laboratory from different US states and after excluding high-level outliers, titers were highest in our region (state of Utah; Supplemental Fig. 1). Although this comparison can be biased by test volumes, differences in tested patient populations, and test utilization practices of ordering clinicians, it could also point to environmental factors leading to higher respiratory exposure to conidia. A recent study showed that positivity rates of fungal-specific IgE serologies (including Aspergillus fumigatus) are higher in the mountainous plains (33). Seasonal temperature variations are also known to impact conidia levels in air (34). It was shown that short-term exposure of conidia to polluted urban air enhances their allergenicity as indicated by the increased expression of the major allergen component (Asp f1) and increased binding to IgE in pooled patient sera (35). Lastly, according to air quality monitoring data from the US Environmental Protection Agency (2000 to 2022 period), PM10 (particulate matter 10 micrometers or smaller in diameter) mean counts in air samples show the highest increases in the southwestern, northwestern, and northern Rockies US regions (36). Higher levels of Aspergillus fumigatus sIgG in our local normal donors may therefore reflect regional environmental trends.
In our dataset, precipitin-positive samples had overall significantly higher sIgG levels as measured by the ImmunoCAP method, which is consistent with prior studies (11, 17, 19, 25, 37). However, there is considerable overlap in sIgG levels between positive and negative precipitin samples. In addition to host and environmental factors, this overlap can also be partially attributed to precipitin reactions formed by immunoglobulin-M/A or antibodies to other cross-reactive fungal allergens (38, 39). In a study that simultaneously compared IgG, IgA, and IgM, IgG achieved the highest sensitivity and specificity for CPA diagnosis (40). Expert recommendations also state that Aspergillus fumigatus-specific IgG is considered the most sensitive serologic test for CPA diagnoses (4). Serologic and microbiologic (e.g., sputum culture or nucleic acid amplification testing) evidence alone cannot differentiate Aspergillus fumigatus sensitization from infection (4, 41, 42). In CPA, the quantitative results of immunoassays have led to multiple studies investigating the utility of trending IgG titers as markers of disease progression and response to antifungal treatment (17, 30, 42, 43). In ABPA, Aspergillus fumigatus-specific-IgE was found to be more useful for disease follow-up, which is feasible given the hypersensitivity pathophysiology of the disease (30).
There is no consensus on the sIgG titer to adopt as a cutoff indicative of disease presence. In our study, ROC-driven cutoffs were 40.4 and 64.7 mcg/mL based on precipitin positivity and clinical diagnosis, respectively. The reference interval cutoff using healthy donor sera was higher at 81.5 mcg/mL. For the ImmunoCAP assay, we propose adopting the 40.4 mcg/mL cutoff based on ROC analysis of precipitin assay results for the following reasons and limitations of the study: (a) the RI interval cutoff derived from local healthy donor study is relatively higher than values reported in the literature. Applying the 40.4 mcg/mL cutoff to local normal donors, 47 of 152 (31%) would be false positives, which can be due to local or regional environmental factors. As our laboratory receives specimens from across the country, a local cutoff may not be applicable to other regions. (b) The intention of the study is to determine assay quantitative cutoff as an alternative to the qualitative precipitin assay. A relatively large number of tested clinical samples were included in ROC analysis based on precipitin results. This cutoff also achieves higher sensitivity (67.8%) and reasonable specificity (72.3%) for precipitin positivity compared to the other 2 cutoffs (Table 2). (c) The precipitin ROC-driven cutoff also achieves reasonable clinical sensitivity and specificity for pulmonary aspergillosis based on our limited clinical dataset as well as values reported in the literature.
Limitations to using the 40.4 mcg/mL cutoff are mainly that it is obtained from comparison to precipitin assays, which are poorly standardized, in addition to the lack of sufficient supporting clinical data. Despite the availability of immunoassays, precipitin tests are still ordered more frequently by clinicians. As guidelines are updated to include immunoassay measurement, more clinical data will eventually be available, and the diagnostic performance of assay cutoffs can be better assessed.
In summary, this study highlights the challenges of determining Aspergillus fugmigatus sIgG cutoffs for assay positivity analogous to the qualitative precipitin assays. For clearer interpretation of sIgG levels, clinical laboratories may include result notes indicating how assay cutoffs were decided and emphasizing interlaboratory differences related to study design and local and regional environmental factors.
Supplemental Material
Supplemental material is available at The Journal of Applied Laboratory Medicine online.
Nonstandard Abbreviations: CPA, chronic pulmonary aspergillosis; ABPA, allergic bronchopulmonary aspergillosis; sIgG, serum-specific immunoglobulin G; RI, reference interval; ROC, receiver operating characteristic.
Author Contributions:The corresponding author takes full responsibility that all authors on this publication have met the following required criteria of eligibility for authorship: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved. Nobody who qualifies for authorship has been omitted from the list.
Buck Lozier (Data curation-Equal, Formal analysis-Equal, Methodology-Equal), Tom Martins (Data curation-Equal, Formal analysis-Equal), Patricia Slev (Writing—review & editing-Equal), and Abdulrahman Saadalla (Conceptualization-Lead, Data curation-Equal, Formal analysis-Equal, Methodology-Equal, Project administration-Lead, Writing—original draft-Lead)
Authors' Disclosures or Potential Conflicts of Interest:No authors declared any potential conflicts of interest.
Role of Sponsor: No sponsor was declared.
