|Year : 2013 | Volume
| Issue : 1 | Page : 34-39
Variables affecting the performance of galactomannan assay in high-risk patients at a Tertiary Care Centre in India
S Khanna1, JK Oberoi1, S Datta1, S Aggarwal2, C Wattal1
1 Department of Clinical Microbiology, Sir Ganga Ram Hospital, Rajinder Nagar, New Delhi, India
2 Department of Medical Oncology, Sir Ganga Ram Hospital, Rajinder Nagar, New Delhi, India
|Date of Submission||30-Mar-2012|
|Date of Acceptance||15-Oct-2012|
|Date of Web Publication||15-Mar-2013|
Department of Clinical Microbiology, Sir Ganga Ram Hospital, Rajinder Nagar, New Delhi
Source of Support: None, Conflict of Interest: None
Background: Diagnosis of invasive aspergillosis (IA) in immunocompromised patients using galactomannan ELISA (GM-ELISA) has shown variable sensitivity and specificity. Objectives: To assess the diagnostic performance of GM-ELISA and analyze the effect of decreasing the cut off value, neutropenia, antifungals and piperacillin-tazobactam (PTZ). Prognostic value using 30 day all-cause mortality was also determined. Materials and Methods: Serum samples from 81 patients categorized into "proven," "probable," and "possible," categories based on revised EORTC/MSG definitions were tested by GM-ELISA. Results: Sensitivity of GM-ELISA in proven, probable and possible cases was 91.7%, 84.6% and 83.3% respectively. At an index cut-off value of 0.5 an increased sensitivity with minimal loss of specificity was observed. Use of antifungals demonstrated a decrease in sensitivity in proven and possible cases whereas it remained unaffected in probable category. Specificity increased from 75% to 100% with a positivity criterion of >2 consecutive samples. Although an increase in specificity was observed in patients not receiving PTZ, it was not statistically significant. Serial GM index values increased significantly in neutropenic patients and were associated with a poor prognosis. Conclusions: GM-ELISA may be a useful diagnostic and prognostic modality for the detection of IA in high risk patients.
Keywords: Galactomannan, invasive aspergillosis, immunocompromised patients, variables
|How to cite this article:|
Khanna S, Oberoi J K, Datta S, Aggarwal S, Wattal C. Variables affecting the performance of galactomannan assay in high-risk patients at a Tertiary Care Centre in India. Indian J Med Microbiol 2013;31:34-9
|How to cite this URL:|
Khanna S, Oberoi J K, Datta S, Aggarwal S, Wattal C. Variables affecting the performance of galactomannan assay in high-risk patients at a Tertiary Care Centre in India. Indian J Med Microbiol [serial online] 2013 [cited 2020 Apr 1];31:34-9. Available from: http://www.ijmm.org/text.asp?2013/31/1/34/108717
| ~ Introduction|| |
Over the last two decades, invasive aspergillosis (IA) has emerged as a major life threatening infection in immunocompromised patients,  occurring in 8-15% of patients undergoing allogenic haemopoietic stem cell (HSCT) and solid-organ transplantation, , and 33-43% patients with malignancy.  Despite advances in therapy, mortality may occur in up to three-fourth of transplant recipients. 
Early diagnosis and appropriate antifungal therapy is important in reducing mortality and morbidity due to IA. However, conventional microbiological and serological techniques are often insufficient to ensure early diagnosis and monitoring invasive infections. In addition, clinical and radiological signs are usually insensitive or nonspecific.  Though a diagnosis of IA by tissue biopsy remains the gold standard, it is invasive and may be potentially harmful, especially among patients with coagulopathy and thrombocytopenia.
The use of extracellular enzymes and metabolites for diagnosing IA is limited due to complexities in the measurements by gas liquid chromatography and mass spectroscopy.  Although PCR-based molecular diagnostic tests show a low positive but high negative predictive value, lack of standardization remains a major barrier for its widespread application.  Thus, the current focus is on detection of circulating antigens, particularly galactomannan (GM), a heat stable heteropolysaccharide cell-wall component of Aspergillus, produced pathologically in vivo in serum, CSF, urine, or bronchoalveolar lavage.  ELISA based GM detection of Aspergillus has been reported as superior to both latex agglutination test and radioimmunoassay.  A double-sandwich ELISA, incorporating the β-1-5 galactofuranose-specific EBA2 rat monoclonal antibody as both the detector and acceptor for galactomannan  has a high sensitivity for diagnosing IA, with a threshold of 0.5-1.0 ng/ml serum.
In light of the above, this study determined the utility of GM assay in diagnosis of IA in high-risk patients and analyzed the variables that could affect the performance of this assay. To our knowledge, such a study is yet to be documented from India.
| ~ Materials and Methods|| |
This single-centre prospective study was conducted in samples, from adult and pediatric patients, obtained from 2006 to 2009 in the Department of Clinical Microbiology at Sir Gangaram hospital, a tertiary health care center with an active transplant program in New Delhi, India. The study was approved as per protocol by the Institutional Review Board of the institution. A total of 81 patients of either sex with suspected IA were tested for the presence of GM antigen by one-stage sandwich immuno-enzymatic technique. The patients were sub-categorized into "proven," "probable," and "possible," cases based on host factors, microbiological and clinical criteria, as per Invasive Fungal Infections Co-operative Group (IFICG) of the European Organization for Research and Treatment of Cancer and Mycoses study group case definitions (EORTC/MSG).  Details of patients including clinical illness, underlying risk factors, radiological and histopathology findings, laboratory investigations, treatment history and outcome were recorded. A total of 25 serum samples from healthy blood donors were also tested for the presence of galactomannan as controls. Eligible patients displayed at least one of the following host factors: Malignancy on chemotherapy within the last 3 months; neutropenia (<500 cells/ml); HSCT; solid organ transplant recipients, particularly those with fulminant hepatic failure or requiring hemodialysis, chronic steroid use (≥4 mg methylprednisolone a day for at least a week in past 3 weeks, or during ICU stay for at least 5 days, or a cumulative dose of ≥250 mg of methylprednisolone in past 3 months); immunocompromised patients of any etiology presenting with persistent fever despite first line antibiotic therapy or with suspected pulmonary infection based on clinical signs and symptoms, or new chest X-ray abnormalities; and those suffering from HIV infection. Informed consent for diagnostic testing was taken from the patients at the time of admission to hospital.
| ~ Samples|| |
Serum samples collected in plain vials were evaluated for the presence of GM antigen. Relevant samples from the respiratory tract, biopsy specimens and body fluids etc., were also evaluated for presence of Aspergillus by microscopy [10% Potassium hydroxide wet mount] and culture on Sabouraud's dextrose agar with and without antibiotics [(chloramphenicol (50 mg/liter), cycloheximide (500 mg/liter), Himedia], incubated at 37 0 C and 25 0 C for a maximum period of 4 weeks]. A positive sample showed the presence of septate hyphae with classical acute angle branching while a negative smear was defined as the absence of hyphae or presence of fungi other than Aspergillus spp. Culture was positive if it showed growth resembling Aspergillus spp. The isolates were identified based on macroscopic and microscopic morphological characteristics, and standard mycological procedures were followed. 
Semi-quantitative determination of Aspergillus GM antigen in serum was done using the Platelia Aspergillus EIA kit (Bio-Rad, France) as per manufacturer's instructions. An OD index of >0.5 was considered galactomannan positive. All positive samples were retested using non treated serum and/or by testing another sample taken from the same patient, if possible. Variables such as effect of decreasing the cut off value, neutropenia, antifungals, and piperacillin-tazobactam were analyzed in different groups. Consent from the patients was taken at the time of admission.
Sensitivity, specificity, positive and negative predictive values were calculated using the 2 × 2 contingency tables. Receiver operator curves (ROC) were calculated using the SPSS 17.00 software. Non-parametric Mann-Whitney U test for quantitative variables was used to evaluate the effect of neutrophil counts on the OD index values. The prognostic value of GM antigen detection in different groups was determined using 30 day all-cause mortality data.
| ~ Results|| |
All 81 patients were sub-categorized as proven IA (12, 14.8%), probable IA (39, 48.2%), possible IA (22, 27.2%), or no IA (8, 9.9%) as per EORTC/MSG criteria. Positive GM-ELISA was observed in 63 (77.7%) while a negative index value was obtained in 18 (22.2%) patients. In 24 (29.6%) patients, GM positivity was limited to a single sample. All 25 control samples from healthy blood donors tested negative for GM antigen. The number of patients, demographic data, samples, patient distribution, and antifungal use according to classification are summarized in [Table 1].
|Table 1: Patient characteristics and serum samples distribution based on classifi cation as per European Organization for Research and treatment of cancer and Mycoses study group criteria|
Click here to view
Based on microbiological criteria, IA was present in 51 (63%) patients of which all were smear positive while culture positivity for Aspergillus sp. was present in 48 (94.1%) patients [Table 2].
|Table 2: Distribution of microscopy and culture results, in 81 high-risk patients|
Click here to view
Outcome was determined on the basis of 30-day all-cause mortality. Of the 81 patients included in the study, after 30 days of inclusion, mortality was seen in 50 (61.7%) patients.
| ~ Galactomannan Antigen Test Performance|| |
Sensitivity and specificity
Eleven patients out of 12 (91.6%), 33 of 39 (84.6%) and 17 of 22 (77.27%) patients were positive for GM in patients diagnosed as proven, probable and possible IA, respectively. Single samples from 2 patients in the "no IA" category were positive whereas 6 (75%) repeatedly tested negative for GM antigen. [Table 1] The sensitivity and specificity of the GM ELISA were calculated in different groups by 3 methods [Table 3].
|Table 3: Sensitivity, specifi city and predictive values according to different estimates|
Click here to view
Calculated the overall sensitivity, specificity, and predictive values of the GM test in each respective category of patients. The sensitivity and specificity were 91.7% and 100%; 84.6% and 100 % and 83.3% and 50% in proven, probable and possible categories, respectively. With one single positive GM test, specificity was 75% whereas when the criterion for GM positivity was defined as positivity of two samples, then the specificity became 100%.
Calculated the sensitivity in only "proven' cases as proportion of true positives, and specificity in only "no IA" cases as proportion of true negatives (proven versus "No IA").
Assumed that all patients of proven and probable IA were true positives and that the patients in "no IA" group were true negatives (proven and probable versus "No IA"). These estimates changed the sensitivity, positive and negative predictive values dramatically [Table 3].
Effect of decreasing the cut off value
Receiver operator characteristic (ROC) curves using SPSS 17.00 software were calculated to illustrate the trade off in rates of true positive results (sensitivity) versus false positive results (1-specificity) as the cutoff for the test was shifted from high (1.0) to low (0.1) OD indices. [Figure 1] The diagnostic accuracy, as given by the area under ROC curves (AUC), was 0.710. In order to determine the optimal index cutoff, sensitivity and the specificity of the test at various cutoffs were calculated. At an OD cutoff index of 0.5 there was an increase of sensitivity up to 90.4% with minimal loss of specificity up to 62.5%.
Effect of mold active antifungals
Out of 81 patients, 28 (34.6%) were receiving mold active antifungals (itraconazole, amphotericin B formulations, and voriconazole). Sensitivity in the proven and possible groups decreased from 100% to 87.5% and 81.3% to 66.7%, respectively in patients receiving antifungals as compared to those not receiving antifungals. No difference however, was observed in the sensitivity in the probable group (84.7%) with the administration of antifungals. [Table 1]
Effect of Piperacillin-tazobactam
To analyze the effect of PTZ on GM antigen test, 8 patients with no evidence of Aspergillus infection by microbiological criteria ("no IA" categories) were considered from the total study patients (n = 81). Out of 8 patients, 3 (37.5%) were receiving PTZ at the time of testing for GM antigen. Of these, 1 (33.3%) tested positive, while 2 (66.6%) tested negative for GM antigen. Among the 5 (62.5%) patients not receiving PTZ, 1 (20%) patient tested positive while a negative test was seen in 4 (80%) patients. Using statistical analysis, effect of PTZ on the performance of GM antigen test was seen by specificity of the test. The odds ratio for this effect was calculated to be 1.55 (95% CI 0.48-4.99). Although an increase in specificity was observed in the group of patients not receiving PTZ, it was not statistically significant (P = 0.55).
Effect of neutropenia
The diagnostic yield of GM antigen assay in neutropenic patients (group 1) was compared with non-neutropenic patients (group 2) having microbiologically documented IA. For this purpose, a total of 51 patients categorized as "proven IA" and "probable IA" were studied. According to the neutrophil counts obtained within 48 hours of a diagnosis of IA, the patients were assigned to group 1 (<500 polymorphonuclear neutrophils (PMN)/mm 3 ; n = 18, 35.3%) and group 2 (>500 PMN/mm 3 ; n = 33, 64.7%). When the two groups were compared using the non-parametric Mann Whitney U test for quantitative variables, the median GM indices were significantly higher in group 1 (2.7, range 0.03-5.1) than in group 2 (1.04, range 0.01-4.8) (P < 0.01).
Of the 81 patients, multiple samples were tested in 57 (70.4%) patients. Of these, serial positive GM OD index values were obtained in 37 (64.9%) patients. There was a significantly higher rate of poor outcome (in terms of 30-day all-cause mortality) in patients with serially increasing GM index values (24, 64.9%) as compared to those with decreasing GM index values (2, 5.4%). ( P = 0.02).
| ~ Discussion|| |
The results of the present study, in which the sensitivities of the GM ELISA were 91.7% in proven cases and 84.6% in probable cases of IA, correlated well with the previous reports , with sensitivities ranging from 56-100%. ,, As reported previously, , our study too demonstrated higher sensitivity in proven IA compared to probable or possible IA. A lower sensitivity of 83.3% in "possible IA" could be because of the low likelihood of Aspergillus infection in these patients. Earlier studies too show a low sensitivity of 25-50% in possible IA. 
In our study, the high specificity in the proven or probable IA could be attributed to high likelihood of IA being present as proven by microscopy and/or culture of various respiratory and non-respiratory samples. However, in "possible IA", the specificity was reduced as false positive results were obtained in two patients. Of these, one had received allogenic HSCT 10 days before and the other was receiving PTZ at the time of diagnosis. Both these conditions could cause false positive GM-ELISA , thus lowering the specificity in this category. Similarly, the specificity in "no IA" category was also lower (75%), than previously reported (84-100%), , possibly due to false positivity in a child who had received HSCT and an adult patient on PTZ therapy. The low numbers of negative patients could also have been the reason for low specificity of the test. However, if two consecutive GM-ELISA readings were taken as determinant of positivity, the specificity and PPV became 100%. Maertens et al.  have also reported an increase in the specificity of GM-ELISA from 85.4% to 98.8% using the criterion of 2 positive ELISA results.
There were also several false-negative cases of IA in the present study. Possible cause of these could be the limited angio-invasion or encapsulation of the process, use of prophylactic mold-active antifungals, low level release of GM antigen, or presence of anti-Aspergillus antibodies. In our study, the patients profile varied from those who were severely neutropenic (PMN <500 cell/mm 3 ) for prolonged duration to the non- neutropenic (PMN >500 cells/mm 3 ). All, except one, who tested negative for GM-ELISA were non-neutropenic and were capable of mounting an antibody response against Aspergillus antigens. In an earlier study, the presence of anti-Aspergillus antibodies led to decrease in sensitivity from 60% to 35%.  Although tests for the detection of anti-Aspergillus antibodies were not performed by us, these antibodies could hypothetically explain the false-negative GM-ELISA.
False positivity of GM-ELISA with PTZ has been observed in previous studies. , which was also seen by us. In our study, although the specificity decreased from 80% to 67% in "no IA" category in patients receiving PTZ, this was not statistically significant. This could be due to small sample size in this category of patients. The effect of PTZ was not evaluated in other categories as presence of IA in these groups could cause overestimation of a false positive test, if the patient was on PTZ therapy.
The PPV of the test obtained in our study was 85-100% and this co-related well with previous studies.  In the possible category, negative predictive value was low (40%) as three patients reported a false negative test. All of these patients were receiving mold active antifungals, which could explain these false negative results. In previous reviews, it has been postulated that mold-active antifungals lower the GM antigen load by altering its release kinetics.  Negative predictive values could not be calculated in the other categories due to the absence of true negative patients in them.
The selection of the cutoff to define positivity for classification as true positives is a major variable affecting the sensitivity and specificity of the test. An elliptical ROC curve confirmed that the test had a good performance with high true positive values and low false positive values, even at low cutoffs. In concordance with earlier studies, a high overall sensitivity and an acceptable specificity was obtained by using GM OD index at cutoff of 0.5 to define a case of IA. 
In our study, GM sensitivities were lower in patients receiving mold-active antifungal therapy. These antifungals can decrease the levels of circulating GM by causing fungal localization in tissues as reported earlier by Marr et al.  In addition, amphotericin B causes slow or inefficient release of GM by altering its pharmacokinetics. Administration of antifungals to patients in the probable category did not affect the sensitivity (84.6%) of the test. It could be due to the high fungal burden, releasing galactomannan, which was not affected by the antifungals given to the patients.
In accordance to studies done before, , GM indices were significantly higher in neutropenic patients (<500 PMN/mm3) with proven and probable IA, suggesting that neutropenia could affect the diagnostic yield of GM-ELISA. The better yield in such patients could be due to higher fungal burden at the time of diagnosis or because of disseminated IA that occurs due to lack of neutrophils causing a deficient inflammatory process. The macrophage deficiency often associated with prolonged chemotherapy-induced neutropenia could also impact GM clearance from the circulation.  Presence of Aspergillus colonization rather than deep infection in some non-neutropenic patients could also lead to lower GM index values.
The overall 30-day all-cause mortality in the study population was 61.7%. In agreement with previous studies,  those with serially rising GM index values had a greater fungal burden and fared poorly compared to those with decreasing values. Serial determination of GM-ELISA could, therefore, serve as a useful prognostic indicator of IA.
| ~ Conclusions|| |
The study demonstrated that GM-ELISA may be a useful diagnostic modality with a high sensitivity and specificity for the detection of IA in high risk patients. An OD index positivity cutoff of 0.5 allowed for an increase in sensitivity with a minimal loss of specificity. Sensitivity decreased with mold active antifungals. Specificity increased with serial sampling and taking a positivity criterion of ≥2 consecutive samples, and decreased if fungus derived antibiotics were used. Serial GM index values increased significantly in neutropenic patients and were associated with a poor prognosis.
| ~ References|| |
|1.||Denning DW. Invasive aspergillosis. Clin Infect Dis 1998;26:781-803. |
|2.||Singh N, Paterson DL. Aspergillus infections in transplant recipients. Clin Microbiol Rev 2005;18:44-69. |
|3.||Paterson DL, Singh N. Invasive aspergillosis in transplant recipients. Medicine (Baltimore) 1999;78:123-38. |
|4.||Herbrecht R, Moghaddam A, Mahmal I, Natarajaname S, Fornecker LM, Letscherbru V. Invasive aspergillosis in the hematologic and immunologic patient: New findings and key questions in leukemia. Med Mycol 2005;43:S239-242. |
|5.||Fukuda T, Boeckh M, Carter RA, Sandmaier BM, Maris MB, Maloney DG, et al. Risks and outcomes of invasive fungal infections in recipients of allogeneic hematopoietic stem cell transplants after nonmyeloablative conditioning. Blood 2003;102:827-33. |
|6.||Latge JP. Aspergillus fumigatus and aspergillosis. Clin Microbiol Rev1999;12:310-50. |
|7.||Klont RR, Mennink-Kersten MASH, Verweij PE. Utility of Aspergillus antigen detection in specimens other than serum specimens. Clin Infect Dis 2004;39:1467-74. |
|8.||Verweij PE, Stynen D, Rijs AJ, de Pauw BE, Hoogkamp-Korstanje JA, Meis JF, et al. Sandwich enzyme-linked immunosorbent assay compared with Pastorex latex agglutination test for diagnosing invasive aspergillosis in immunocompromised patients. J Clin Microbiol 1995;33:1912-4. |
|9.||Morelle W, Bernard M, Debeaupuis JP, Buitrago M, Tabouret M, Latge JP. Galactomannoproteins of Aspergillus fumigatus. Eukaryot Cell 2005;41:308-16. |
|10.||Pauw BD, Walsh TJ, Donnelly JP, Stevens DA, Edwards JE, Calandra T, et al. Revised Definitions of Invasive Fungal Disease from the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group. Clin Infect Dis 2008;46:1813-21. |
|11.||Larone DH. Medically important fungi: A guide to identification. 3 rd edition. Washington DC: American Society for Microbiology; 2005. |
|12.||Maertens J, Van Eldere J, Verhaegen J, Verbeken E, Verschakelen J, Boogaerts M. Use of circulating galactomannan screening for early diagnosis of invasive aspergillosis in allogeneic stem cell transplant recipients. J Infect Dis 2002;186:1297-306. |
|13.||Fortun J, Martin-Davila P, Alvarez ME, Sanchez-Sousa A, Quereda C, Navas E, et al. Aspergillus antigenemia sandwich-enzyme immunoassay test as a serodiagnostic method for invasive aspergillosis in liver transplant recipients. Transplantation 2001;71:145-9. |
|14.||Herbrecht R, Letscher-Bru V, Oprea C, Lioure B, Waller J, Campos F, et al. Aspergillus galactomannan detection in the diagnosis of invasive aspergillosis in cancer patients. J Clin Oncol 2002;20:1898-906. |
|15.||Mori AY, Kanda Y, Oshima K, Kako S, Shinohara A, Nakasone H, et al. False-positive Aspergillus galactomannan antigenaemia after haematopoietic stem cell transplantation. J Antimicrob Chemother 2008;61:411-6. |
|16.||Adam O, Auperin A, Wilquin F, Bourhis JH, Gachot J, Chachaty E. Treatment with piperacillin-tazobactam and false-positive Aspergillus galactomannan antigen test results for patients with hematological malignancies. Clin Infect Dis 2004;38:917-20. |
|17.||Fortun J, Martin-Davila P, Alvarez ME, Norman F, Sanchez-Sousa A, Gajate L, et al. False positive results of Aspergillus galactomannan antigenemia in liver transplant recipients. Transplantation 2009;87:256-60. |
|18.||Marr KA, Balajee SA, McLaughlin L, Tabouret M, Bentsen C, Walsh TJ. Detection of galactomannan antigenemia by enzyme immunoassay for the diagnosis of invasive aspergillosis: variables that affect performance. J Infect Dis 2004;190:641-9. |
|19.||Cordonnier C, Botterel F, Ben Amor R, Pautas C, Maury S, Kuentz M, et al. Correlation between galactomannan antigen levels in serum and neutrophil counts in haematological patients with invasive aspergillosis. Clin Microbiol Infect 2009;15:81-6. |
|20.||Boutboul F, Alberti C, Leblanc T, Sulahian A, Gluckman E, Derouin F, et al. Invasive aspergillosis in allogenic stem cell transplant recipients: Increasing antigenemia is associated with progressive disease. Clin Infect Dis 2002;34:939-43. |
[Table 1], [Table 2], [Table 3]
|This article has been cited by|
||Pulmonary fungal infections- recent updates
| ||Marwan M. Azar,Joshua Malo,Kenneth S. Knox,Chadi A. Hage |
| ||Current Respiratory Care Reports. 2014; |
|[Pubmed] | [DOI]|
||Threshold of galactomannan antigenemia positivity for early diagnosis of invasive aspergillosis in neutropenic children
| ||Veronique Dinand,Madasu Anjan,Jaswinder Kaur Oberoi,Shilpi Khanna,Satya Prakash Yadav,Chand Wattal,Anupam Sachdeva |
| ||Journal of Microbiology, Immunology and Infection. 2014; |
|[Pubmed] | [DOI]|
||The Future of Fungal Serology
| ||Mark D. Lindsley |
| ||Current Fungal Infection Reports. 2013; 7(3): 161 |
|[Pubmed] | [DOI]|