Indication - Aspergillosis
Aspergillus antigen test
First added in 2020
Aid to diagnosis
To aid in the diagnosis of invasive aspergillosis in immunocompromised patients
Serum, Broncho-alveolar lavage (BAL)
ICD11 code: 1F20.Z
Summary of evidence evaluation
There is substantial evidence of the accuracy of both serum and BAL GM ELISA tests for diagnosing IA in patients receiving intensive chemotherapy or stem cell transplants with high likelihood of neutropenia. Choice of threshold is critical to balance risks of missing cases against false positives. Evidence is from both using the test for diagnosis in those suspected of having infection (for BAL and serum) and using it for surveillance for infection in those at high risk (serum). There has been little evaluation of the benefits of using GM ELISA tests to improve patient care and outcomes, or of using the test in formal screening or monitoring programmes. Guideline panels have judged the accuracy and evidence as adequate to recommend using the test across several different patient groups. There are few published studies of the accuracy of the newer LFA versions of the antigen tests. Available evidence suggests similar performance to the ELISA tests in BAL samples, but estimates of accuracy remain uncertain. The full evidence review for this test category is available online at: https://www.who.int/medical_devices/diagnostics/selection_in-vitro/selection_in-vitro-meetings/new-prod-categories_3
Summary of SAGE IVD deliberations
IA is a WHO priority and a very severe disease. Its incidence may not be high, but it is an almost uniformly fatal disease in LMICs, where the mortality rate is mainly affected by the timing and initiation of antifungal therapy. It is a difficult disease to diagnose, and Aspergillus antigen testing is imperfect. For example, it has a highly variable sensitivity in serum testing in LFA test formats and the antigen detected (GM) can sometimes cross-react with other species. In all cases, Aspergillus antigen tests should only be used when accompanied by a strong clinical suspicion of IA. The cost per test seems prohibitive, but it likely comes at the benefit of better or more selective use of voriconazole, which is on the EML specifically to treat IA. The submission contains convincing evidence on the public health relevance of the conditions and the usefulness, performance and accessibility of the test. And clinical practice guidelines around the world already include Aspergillus antigen testing in the diagnostic workup for IA. The test is easy to perform compared with other methods, although availability in LMICs is doubtful.
SAGE IVD recommendation
SAGE IVD recommended the inclusion of the Aspergillus antigen test category in the third EDL: • as a disease-specific IVD for clinical laboratories (EDL 3; Section II.b; Aspergillosis); • using an RDT or immunoassay format; • to aid in the diagnosis of invasive aspergillosis in immunocompromised patients. In considering the test formats available, SAGE IVD noted that while both LFA and ELISA devices should be listed in the EDL, the use of LFA devices should be limited to BAL specimens until more data can be obtained on the accuracy of these devices using serum. SAGE IVD further recommended comparing LFA devices with PCR for EDL 4 as an alternative, because many tertiary institutions already have PCR technologies available. The group also noted that in the case of this IVD, it is particularly important that the EDL is clearly linked to relevant guidelines so that users of the test are aware of the potential cross-reactivity issues associated with GM assays.
Details of submission from 2020
Disease condition and impact on patients IA is almost always fatal unless diagnosed and treated promptly. It is silent in its initial clinical manifestations, and microbiological cultures are very insensitive. Unless clinical suspicion is very high, and appropriate non-culture-based tests are done, most patients do not survive. IA is frequently an incidental finding at autopsy. For example, it was the most common major error in missed infections in ICU deaths in the United Kingdom (1); in Italy, only 11% of cases in AIDS patients were diagnosed in life (2). Major improvements in diagnosis are required to improve outcomes, and these are now available for LMICs. Does the test meet a medical need? Access to Aspergillus antigen is critical to diagnosing IA. As LMICs establish cancer treatment programmes and intensive care, they will face an increasing number of IA cases. These will only be effectively diagnosed through Aspergillus antigen detection, as culture is insensitive, biopsy with histopathology requires high levels of skill and radiology is often non-specific. How the test is used IA is a multimodal diagnosis and galactomannan (GM) is a critical component of this, although not the only one. It is therefore challenging to separate out GM’s specific role from other tests, particularly imaging, except in ICU/ventilated patients, where it is pivotal.
Public health relevance
Prevalence The clinical presentation of IA is relatively rapid, ranging from days to a few weeks. Fever is uncommon and symptoms often absent until late in the course. While neutropenia is the best known risk group, IA is also increasingly recognized in other patients, notably those hospitalized with chronic obstructive pulmonary disease (COPD), lung cancer and occasionally other solid tumours, or liver failure (4%), as well as those receiving corticosteroid therapy and transplant recipients (notably allogeneic stem cell and lung, but also renal and heart recipients). In lung transplant recipients, tracheobronchial disease is the earliest and most common manifestation of IA, usually diagnosed when ulcerations or pseudomembranes are noted on surveillance bronchoscopy. IA is present in 5% of renal and hepatic transplant recipients. In medical ICUs 2–5% of patients develop IA. The widespread use of steroids for COPD exacerbations may contribute to the increased rate of IA, but COPD is an independent risk factor. The IA rate in patients hospitalized for COPD is 1.3% in Spain (3) and 3.9% in Southern China (3). In patients with lung cancer, the IA attack rate is 2.6% based on data from China (4). IA is found in 4% of all patients who die of AIDS, based on a mean of autopsy studies, and in 19% of cases with complicating severe influenza (5). All other immunocompromising conditions and the rare non-immunocompromised patients (such as post-influenza) comprise 5% of cases. According to conservative burden estimates, IA occurs in 13% of acute myeloid leukaemia cases and 50% of all cases in haematology patients. Aspergillus contamination of air filters in ICUs is linked to nosocomial IA. Socioeconomic impact Data from HICs reveal that IA has a high economic cost in ICU patients. Patients with IA have high death rates and longer hospital stays. Early diagnosis is expected to improve outcomes and reduce lengths of hospitalization (6).
WHO or other clinical guidelines relevant to the test
Guidelines were published by IDSA in 2016 (7); ESCMID, the European Confederation of Medical Mycology (ECMM) and the European Respiratory Society (ERS) in 2017 (8); and the American Thoracic Society in 2019 (9). All recommend the use of serum or bronchoalveolar lavage (BAL) GM antigen – EIA based on rat-derived EB-A2 monoclonal immunoglobulin M (IgM) capture and detector antibody, and EIA as well as LFIAs based on glycoprotein/GM-like antigens, respectively – to screen for and diagnose IA in patients with haematologic malignancies (both neutropenic and non-neutropenic; the sensitivity is lower in the latter group) and haematopoietic stem cell transplant recipients (except those on anti-Aspergillus prophylaxis). With a lower strength of recommendation, the guidelines also recommend use in solid organ transplant recipients, and in other patient groups. Optical density index cut-offs recommended for these purposes are 0.5–1. IDSA and ESCMID-ECMM-ERS both recommend that this test should not be performed on patients receiving anti-mould-prophylaxis.
Evidence for diagnostic accuracy
The GM EIA tests yield quantitative results, with a linear range of measurement that varies with each batch of testing and with the spectrophotometer used. Diagnostic thresholds well within this linear range have been proposed, but there are no universally agreed thresholds for predicting the outcome. False negative results can be seen in patients receiving antifungal treatment. False positives have also been reported in patients with severe mucositis and in the presence of Histoplasma spp. antigen and other fungal antigens in serum or BAL. Two Cochrane systematic reviews of diagnostic test accuracy have studied GM detection in serum and in BAL using EIA as the index test and against elaborate EORTC/MSG criteria as reference standard. The first, by Leeflang et al. (10) looked at GM detection in serum and included 50 studies, containing 5660 neutropenic patients, of whom 586 had proven or probable IA. When using Oswestry Disability Index (ODI) cut-offs of 0.5, 1.0 and 1.5, the sensitivities of the test were 78% (95% CI: 70–85), 71% (95% CI: 63–78) and 63% (95% CI: 49–78), respectively; the specificities were 85% (95% CI: 78–91), 90% (95% CI: 86–93) and 93% (95% CI: 89–97), respectively. The second Cochrane systematic review, by de Heer et al. (11), looked at 17 studies on GM detection in BAL. The diagnostic performance at ODI cut-off of 0.5 showed sensitivity of 88% (95% CI: 75–100) and specificity 81% (95% CI: 71–91). An ODI cut-off of 1.0 yielded better specificity, i.e. sensitivity of 78% (95% CI: 61–95) and specificity of 93% (95% CI: 87–98). At ODI cut-offs of 1.5 or higher, the heterogeneity in specificity decreased significantly and was invariably > 90%. LFA tests yield qualitative results. These tests detect different fungal targets, using either the GM antigen or the mouse monoclonal antibody JF5 to detect Aspergillus-specific antigens. Systematic reviews have been done on the new LFA tests (12), and some landscape reviews are also published (13). The systematic review by Pan et al. (12) included seven primary studies and showed that pooled sensitivity, specificity and diagnostic odds ratio for the proven/probable versus no IA cases were 86% (95 % CI: 76–93), 93% (95 % CI: 89–96) and 65.94 (95 % CI: 27.21–159.81) in the lateral flow device test using BAL fluid. The performance in BAL was better than in serum. Zhang et al. (14) describe performance of the Aspergillus glycoprotein LFA in combination with other tests.
Evidence for clinical usefulness and impact
GM is a critical component of the methods used to diagnose IA (clinical signs, imaging signs) and is best used in combination with other diagnostic modalities. GM testing has a high positive predictive value for diagnosing IA. In immunocompromised patients suspected of having IA, at a cut-off of 1.0, serum GM testing had positive and negative likelihood ratios of 6.6 (95% CI: 3.4–12.5) and 0.24 (95% CI: 0.11–0.5), respectively, with a diagnostic odds ratio (DOR) of 28 (95% CI: 9–83). In BAL, at a cut-off of 1.0, GM positive and negative likelihood ratios were 14.3 (95% CI: 7.2–28.5) and 0.11 (95% CI: 0.04–0.26), respectively, with a DOR of 134 (95% CI: 43–420) (9). These data show further evidence of clinical accuracy, but also indicate, in the absence of outcome studies, the role of these IVDs in prompting more diagnosis and thus earlier treatment. Dabas et al. (15) studied serum GM in a heterogenous group of 235 critically ill patients and found that for 37% of patients, IA diagnoses were made earlier with serum GM than with radiology. The review of 13 studies by Zhang et al. (14) also concluded that adding Aspergillus GM or LFA to the diagnostic algorithm confirmed more IA cases. Serial screening for serum GM in prolonged neutropenia and in allogeneic stem cell transplant recipients (those not receiving mould-active prophylaxis) during the early engraftment phase has a high sensitivity and negative predictive value for IA. But the inability of a negative GM to exclude non-Aspergillus mould infections is well known. Thresholds are not well established to predict outcome; for GM as a prognostic marker, some evidence points to GM levels falling with effective antifungal treatment.
Evidence for economic impact and/or cost–effectiveness
In the absence of a screening test for IA, patients at risk receive antifungal prophylaxis and/or empiric treatment. GM tests can be used to trigger pre-emptive treatment with the potential to influence cost as well as antifungal exposure. In a meta-analysis of testing to screen vs empiric treatment (16), using the test resulted in minimal cost savings but significantly reduced exposure to antifungal agents. In another comparative study (17), pre-emptive treatment after GM testing was 5% less cost-effective than empiric treatment and reduced antifungal treatment days.
Ethical issues, equity and human rights issues
The test targets immunocompromised patients and generates no ethical, equity or human rights concerns.
1. Winters B, Custer J, Galvagno SM, Colantuoni E, Kapoor SG, et al. Diagnostic errors in the intensive care unit: a systematic review of autopsy studies. BMJ Qual Saf. 2012;21(11):894–902. doi:10.1136/bmjqs-2012-000803. 2. Antinori S, Corbellino M, Necchi A, Corradini P, Vismara C, et al. Immune reconstitution inflammatory syndrome associated with Aspergillus terreus pulmonary infection in an autologous stem cell transplant recipient. Transpl Infect Dis. 2010;12(1):64–68. doi:10.1111/j.1399-3062.2009.00460.x. 3. Xu H, Li L, Huang WJ, Wang LX, Li WF, et al. Invasive pulmonary aspergillosis in patients with chronic obstructive pulmonary disease: a case control study from China. Clin Microbiol Infect. 2012;18(4):403–408. doi:10.1111/j.1469-0691.2011.03503.x. 4. Yan X, Li M, Jiang M, Zou LQ, Luo F, Jiang Y. Clinical characteristics of 45 patients with invasive pulmonary aspergillosis: retrospective analysis of 1711 lung cancer cases. Cancer. 2009;115(21):5018–5025. doi:10.1002/cncr.24559. 5. Schauwvlieghe AF, Rijnders BJ, Philips N, Verwijs R, Vanderbeke L, et al. Invasive aspergillosis in patients admitted to the intensive care unit with severe influenza: a retrospective cohort study. Lancet Respir Med. 2018;6(10):782–792. doi:10.1016/S2213-2600(18)30274-1. 6. Baddley JW, Stephens JM, Ji X, Gao X, Schlamm HT, et al. Aspergillosis in intensive care unit (ICU) patients: epidemiology and economic outcomes. BMC Infect Dis. 2013;13:29. doi:10.1186/1471-2334-13-29. 7. Patterson TF, Thompson III GR, Denning DW, Fishman JA, Hadley S, et al. Practice guidelines for the diagnosis and management of aspergillosis: 2016 update by the Infectious Diseases Society of America. Clin Infect Dis. 2016;63(4):e1–e60. doi:10.1093/cid/ciw326. 8. Ullmann AJ, Aguado JM, Arikan-Akdagli S, Denning DW, Groll AH, et al. Diagnosis and management of Aspergillus diseases: executive summary of the 2017 ESCMID-ECMM-ERS guideline. Clin Microbiol Infect. 2018;24:e1–e38. doi:10.1016/j.cmi.2018.01.002. 9. Hage CA, Carmona EM, Epelbaum O, Evans SE, Gabe LM, et al. Microbiological laboratory testing in the diagnosis of fungal infections in pulmonary and critical care practice. An official American Thoracic Society clinical practice guideline. Am J Respir Crit Care Med. 2019;200(5):535–550. doi:10.1164/rccm.201906-1185ST. 10. Leeflang MM, Debets‐Ossenkopp YJ, Visser CE, Scholten RJ, Hooft L, et al. Galactomannan detection for invasive aspergillosis in immunocompromised patients. Cochrane Database Syst Rev. 2008;4:CD007394. doi:10.1002/14651858.CD007394. 11. de Heer K, Gerritsen MG, Visser CE, Leeflang MM. Galactomannan detection in broncho‐alveolar lavage fluid for invasive aspergillosis in immunocompromised patients. Cochrane Database Syst Rev. 2019;5:CD012399. doi:10.1002/14651858.CD012399.pub2. 12. Pan Z, Fu M, Zhang J, Zhou H, Fu Y, et al. Diagnostic accuracy of a novel lateral-flow device in invasive aspergillosis: a meta-analysis. J Med Microbiol. 2015;64(7):702–707. doi:10.1099/jmm.0.000092. 13. Heldt S, Hoenigl M. Lateral flow assays for the diagnosis of invasive aspergillosis: current status. Current fungal infection reports. 2017;11(2):45–51. doi:10.1007/s12281-017-0275-8. 14. Zhang L, Guo Z, Xie S, Zhou J, Chen G, et al. The performance of galactomannan in combination with 1,3-β-D-glucan or aspergillus-lateral flow device for the diagnosis of invasive aspergillosis: evidences from 13 studies. Diagn Microbiol Infect Dis. 2019;93(1):44–53. doi:10.1016/j.diagmicrobio.2018.08.005. 15. Dabas Y, Mohan A, Xess I. Serum galactomannan antigen as a prognostic and diagnostic marker for invasive aspergillosis in heterogeneous medicine ICU patient population. PloS One. 2018;13(4). doi:10.1371/journal.pone.0196196. 16. Fung M, Kim J, Marty FM, Schwarzinger M, Koo S. Meta-analysis and cost comparison of empirical versus pre-emptive antifungal strategies in hematologic malignancy patients with high-risk febrile neutropenia. PLoS One. 2015;10(11). doi:10.1371/journal.pone.0140930. 17. Walker BS, Schmidt RL, Tantravahi S, Kim K, Hanson KE. Cost‐effectiveness of antifungal prophylaxis, preemptive therapy, or empiric treatment following allogeneic hematopoietic stem cell transplant. Transpl Infect Dis. 2019;21(5):e13148. doi:10.1111/tid.13148.