0
Original Research: LUNG INFECTION |

Accuracy of BAL Galactomannan in Diagnosing Invasive Aspergillosis: A Bivariate Metaanalysis and Systematic Review FREE TO VIEW

Ya-Ling Guo, MD; Yi-Qiang Chen, MD, PhD; Ke Wang, MD; Shou-Ming Qin, MD; Cong Wu, MD; Jin-Liang Kong, MD, PhD
Author and Funding Information

From the Department of Respiratory Diseases (Dr Guo), Nursing School, Guangxi Medical University; and the Institute of Respiratory Diseases (Drs Chen, Wang, Qin, Wu, and Kong), First Affiliated Hospital, Guangxi Medical University, Nanning, Guangxi, Republic of China.

Correspondence to: Ke Wang, MD, Institute of Respiratory Diseases, First Affiliated Hospital, Guangxi Medical University, Nanning, Guangxi, ROC; e-mail: walk7632@yahoo.com.cn


Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (http://www.chestpubs.org/site/misc/reprints.xhtml)


© 2010 American College of Chest Physicians


Chest. 2010;138(4):817-824. doi:10.1378/chest.10-0488
Text Size: A A A
Published online

Background:  A serum galactomannan (GM) assay has been approved for diagnosing invasive aspergillosis (IA). However, the role of the BAL-GM assay has not been well established. Therefore, we conducted a metaanalysis to determine the overall accuracy of BAL-GM in the diagnosis of IA.

Methods:  After a systematic review of English-language studies, the sensitivity (SEN), specificity (SPE), and positive and negative likelihood ratios (PLR and NLR, respectively) of BAL-GM for the diagnosis of IA were pooled using a bivariate metaanalysis. Hierarchic summary receiver operating characteristic curves were used to summarize overall test performance. Potential between-study heterogeneity was explored by subgroup analyses. We calculated posttest probability to evaluate clinical usefulness.

Results:  Twelve reports, including 13 studies, met our inclusion criteria. The summary estimates of the BAL-GM assay for proven or probable IA were as follows: SEN, 0.90 (95% CI, 0.79-0.96); SPE, 0.94 (95% CI, 0.90-0.96); PLR, 14.87 (95% CI, 8.89-24.90); and NLR, 0.10 (95% CI, 0.04-0.24). The four summary estimates of the BAL-GM assay for proven IA were 0.94 (95% CI, 0.86-0.98), 0.79 (95% CI, 0.68-0.86), 4.41 (95% CI, 2.87-6.77), and 0.07 (95% CI, 0.03-0.09), respectively. Significant heterogeneity was present.

Conclusions:  BAL-GM determination is a sensitive and specific test for the diagnosis of proven and probable IA. The measurement of BAL-GM is thus likely to be a useful tool for diagnosing IA. Further studies focused on the impact of treatment agents are needed.

Figures in this Article

Invasive aspergillosis (IA) has emerged as a common, opportunistic fungal infection among immunocompromised patients and has been associated with considerable morbidity and mortality.1-4 The prevalence of IA ranges from 1% to 15%, and mortality can exceed 90%.2 The high mortality rate results partly from difficulties in establishing an early diagnosis because of nonspecific clinical manifestations, delayed radiologic findings, and poor yield of cultures.5 Nowadays, diagnostic tools using galactomannan (GM) have become the focus of clinical study.5,6

GM is a cell-wall component cell released by the Aspergillus species during hyphal growth.5 The available commercial and widespread acceptable assay is Platelia sandwich enzyme-linked immunosorbent assay (BioRad; Marnes-La-Coquette, France).5-8 Serum GM detection has been validated as a moderately useful diagnosis tool.6-8 However, BAL fluid (BALF)-GM assay appears to be more sensitive.9 The rationale is that the bronchial tree of patients with the pulmonary form, the most common presentation of IA, has higher fungal burden. Because the role of the BAL-GM assay for diagnosing IA in at-risk patients has not been well established, we undertook a metaanalysis to assess overall accuracy.

Study Identification and Selection

Two investigators (J. L. K. and Y. L. G.) searched the MEDLINE and EMBASE databases for relevant articles published up to January 2010. Search terms included “aspergillosis,” “aspergillus,” “aspergill*,” “bronchoalveolar lavage,” and “galactomannan,” and the syntax for the MEDLINE searches was as follows: “aspergillosis” OR “aspergillus” OR “aspergill*” AND “bronchoalveolar lavage” AND “galactomannan.” The searches were limited to English publications on human subjects. We screened the reference lists of included studies and related publications. The results were then hand searched for eligible trials; we did not include abstracts or meeting proceedings. Results were arbitrated by a second investigator (K. W.). Full-text publications concerning Platelia on BALF were included if (1) they used European Organization for Research and Treatment of Cancer/Mycoses Study Group (EORTC/MSG) or revised EORTC/MSG criteria as a reference standard,1,6 (2) they provided data for two-by-two tables, and (3) they included immunocompromised or at-risk patients. Studies with fewer than 10 patients were excluded to avoid selection bias.

Data Extraction and Quality Assessment

Two investigators (Y. Q. C. and S. M. Q.) independently extracted the following information: population, study, and assay characteristics; β-lactam and antifungal therapy; methodologic quality; incorporation bias of applying a positive BAL-GM to the reference standard; and data for two-by-two tables. When the same population was reported in several publications, we retained only the most informative article or complete study, to avoid duplication of information. We also asked the investigators for additional information. Any disagreements were resolved by discussion with a third author (C. W.). We assessed study quality using the modified quality assessment for studies of diagnostic accuracy (QUADAS) tool8,10 and the standards for reporting diagnostic accuracy (STARD) tool.11 As mentioned, we requested the relevant information from the authors if data were unreported. The “unreported” items were treated as “no” if we received no answers.

Data Synthesis

According to the revised EORTC/MSG criteria, our reference standards classified patients into four groups: proven IA, probable IA, possible IA, and no IA.6 Considering that recent published guidelines emphasize the importance of managing probable and proven IA similarly,2 we mainly constructed and analyzed two-by-two tables (proven or probable IA vs possible or no IA). We also constructed other two-by-two tables (proven IA vs probable, possible, or no IA). If several cutoffs were reported in one study, we used the cutoff that offered the best test performance.

By using a bivariate regression approach, we estimated the overall sensitivity (SEN) and specificity (SPE) with 95% CIs as the main outcome measures, and constructed hierarchic summary receiver operating characteristic (SROC) curves.12 Based on random-effects models, this bivariate approach accounts for potential between-study heterogeneity and incorporates the possible correlation between the SEN and the SPE. By using the pooled SEN and SPE, we also calculated positive and negative likelihood ratios (PLR and NLR, respectively).12,13

We assessed statistically significant heterogeneity using the I2 statistic.14 Potential between-study heterogeneity was explored by subgroup analyses.15,16 Covariates requiring that at least 80% of studies reported on a particular item were analyzed: population (only hematologic malignancy vs mixed/other), design (cohort vs case-control), data collection (prospective vs retrospective), sampling method (consecutive/random vs nonconsecutive/nonrandom/unreported), revised EORTC/MSG criteria in 2008 (yes vs no), blinded status (yes vs no/unreported), incorporation bias (yes vs no), positive samples required (single vs two samples), and antifungal intervention (yes vs no/unreported). We assessed the potential publication bias using the funnel plots of Deeks et al.17 The prevalence of 12%8 was used for the calculation of posttest probability (PTP) with Fagan18 nomograms. All analyses were performed using STATA, version 10 (Stata Corp; College Station, TX) with the program “Midas.”19 All statistical tests were two sided, with P values of .05 denoting statistical significance.

Eligible Study Characteristics and Quality Assessment

Of the 503 references identified, we eventually pooled 12 reports, including 13 eligible studies (flowchart not shown).20-31Tables 1 and 2 list the characteristics and quality assessment of the eligible studies. The scores of quality issues of the STARD tool varied from 12 to 18. Quality assessment is shown with a bar graph according to the modified QUADAS criteria (Fig 1).

Table Graphic Jump Location
Table 1 —Characteristics and Quality of 13 Studies Included in the Metaanalysis of Diagnosis of IA using BAL-GM

BAL-GM = BAL-galactomannan; BC = bronchogenic carcinoma; CoR = consecutive or random; EORTC/MSG = European Organization of the Research and Treatment of Cancer/Mycoses Study Group; HM = hematologic malignancy; HSCT = hematopoietic stem cell transplant; IA = invasive aspergillosis; IC = immunocompromised; IFI = invasive fungal infection; SOT = solid-organ transplant.

a 

Incorporation of bias by applying a positive BAL-GM result to the diagnostic criteria.

b 

The second study in Becker’s report.

c 

Healthy controls were excluded.

Table Graphic Jump Location
Table 2 —Characteristics and Results of 13 Studies Included in the Metaanalysis of Diagnosis of IA Using BAL-GM

“Samples” required refers to the number of samples required for positivity. STARD = standards for reporting diagnostic accuracy. See Table 1 for expansion of other abbreviations.

a 

The second study in Becker’s report.

b 

Healthy controls were excluded.

Figure Jump LinkFigure 1. Overall quality assessment of all 13 included studies. Data are presented as stacked bars for each quality item, including modified quality assessment for studies of diagnostic accuracy (QUADAS) criteria.Grahic Jump Location
Analyses for Proven or Probable Cases
SEN, SPE, PLR, NLR, and SROC Curve:

We included 13 studies, of which four had a cutoff value of 0.5, seven had a cutoff value of 1.0, and the remaining two had cutoff values of 1.5 and 2, respectively. The SEN ranged from 0.59 to 1.00, whereas the SPE ranged from 0.76 to 1.00 (Forest plots not shown). The smaller numbers of IA cases in different studies and the variation due to chance mainly led to the wider range of SEN. The PLR and NLR and associated 95% CI were 14.87 (8.89-24.90) and 0.10 (0.04-0.24), respectively. I2 values of all measures were > 50%, indicating significant heterogeneity for these indexes.

Hierarchic SROC curves are displayed in Figure 2. The SROC curve represents the relationship between SEN and SPE across studies, recognizing that different thresholds may have been used. Because the bivariate approach estimates the strength and the shape of the correlation between SEN and SPE, we can draw a 95% confidence ellipse and a 95% prediction ellipse. The area under the SROC curve (AUC) was 0.97 (0.95-0.98) (Fig 2A), indicating that the BAL-GM assay has a high discrimination ability.

Figure Jump LinkFigure 2. SROC curves from the bivariate model for proven or probable cases (A) and proven cases (B), respectively. The smaller region (confidence contour) contains likely combinations of the mean value of sensitivity and specificity. The wider region (prediction contour) demonstrates more uncertainty as to where the likely values of sensitivity and specificity might occur for individual studies. SROC = summary receiver operating characteristic.Grahic Jump Location
Investigations of Heterogeneity, Publication Bias, and PTP

Subgroup analyses are shown in Figure 3. The SPE was significantly lower with some covariates, such as antifungal intervention. However, the effects of these covariates were not significant in the SEN. The funnel plots of Deeks et al17 failed to find any publication bias (P = .79) (figure not shown). The nomogram of Fagan18 demonstrated that the BAL-GM assay increased the probability of IA nearly sixfold when results were positive and decreased the probability to 1% when negative (Fig 4).

Figure Jump LinkFigure 3. Forest plot of subgroup analyses for sensitivity and specificity. Antifungal = antifungal intervention; Bias = the incorporation bias; Consec = consecutive or random; Criteria2008 = revised EORTC/MSG criteria in 2008; HM = hematologic malignancy; OneSample = one single sample testing for positivity; Prosp = prospective design. *P < .05; **P < .01; ***P < .001.Grahic Jump Location
Figure Jump LinkFigure 4. Fagan’s nomogram for calculating posttest probabilities (PTPs). A straight edge was used to link the pretest probability of invasive aspergillosis (IA) with the PTP, by crossing the likelihood ratio line at a point that describes the results obtained.Grahic Jump Location
Analyses for Proven Cases

For proven cases, we included 10 studies in which two studies had a cutoff value of 0.5, six had a cutoff value 1.0, and the remaining two had cutoff values of 1.5 and 2, respectively. The SEN, SPE, PLR, NLR, and AUC with associated 95% CIs were 0.94 (0.86-0.98), 0.79 (0.68-0.86), 4.41 (2.87-6.77), 0.07 (0.03-0.19), and 0.95 (0.93-0.97), respectively. The increased SEN was related to that most proven cases had positive BAL-GM results, whereas the decreased SPE was probably due to those probable cases regarded as negative. SROC curves are displayed in Figure 2B. Significant heterogeneity was present. In the subgroup analyses for proven cases, the SPE was significantly lower, with two covariates, including only hematologic malignancy population and a single positive sample. There was no publication bias.

IA has emerged as an important cause of life-threatening infections among immunocompromised or at-risk patients. As the most common form of IA, pulmonary aspergillosis accounts for > 50% of all cases and may be the source of dissemination to other critical organs.32 Invasive procedures relying on histopathologic or cytopathologic evidence are still considered the gold standard for establishing the diagnosis of IA.6 However, these procedures are rarely performed, especially in critically ill patients or in those with thrombocytopenia.

Bronchoscopy with BAL seems to be a feasible diagnostic tool, with high yield and low complication rates in immunocompromised patients.33-35 However, the diagnostic yield of BALF culture may have a low SEN. If culture-independent methods such as GM detection are used, the yield of BAL can be increased.35 Variable performances of the BAL-GM assay have been found in different at-risk patient populations. The goal of the present metaanalysis was to establish the overall accuracy of the BAL-GM assay for the diagnosis of IA.

Overall, we drew the conclusion that the BAL-GM assay was more powerful than serum GM detection, which has been evaluated in two metaanalyses.7,8 Although there were some methodologic limitations in Pfeiffer’s study,12,36 both metaanalyses drew the similar conclusion that serum GM detection is moderately useful for diagnosing IA.7,8 In the study of Leeflang et al8 in which only proven plus probable cases were considered, the decreasing threshold from 1.5 to 0.5 increased the overall SEN by 17% (from 62% to 79%) but decreased the overall SPE by 13% (from 95% to 82%). For the overall analysis, we used a single cutoff to gain in simplicity and transparency, and performed the metaanalysis with bivariate regression models. Adopting the revised EORTC/MSG criteria or similar criteria as a reference standard, the BAL-GM assay had a higher accuracy, with both the SEN and the SPE ≥ 90%. The elevated SEN of GM results could be partly related to the bronchial tree of IA cases, which had a higher fungal burden. Further, the hyphae release more GM and antigenic determinants than the conidia,23,25,37 which may enable increased SPE and help differentiate between colonization and invasive disease. Because the study of Leeflang et al8 did not report the AUC index, we had to compare the summary points in SROC curves between the study of Leeflang et al8 and ours through visual examination. The more accurate the test, the closer the curve approached the top left-hand corner of the graph (AUC = 1). As seen in Figure 2A, our summary points were closer to the top left-hand corner than those of the study of Leeflang et al,8 indicating that the BAL-GM assay has a better discriminative ability.

Likelihood ratios are metrics that take into account the interaction between the SEN and the SPE in their calculation, and PLR > 10 and NLR < 0.1 are considered convincing evidence to rule in or rule out diagnoses, respectively, in most circumstances.13 Although two metaanalyses of serum GM assay did not report these results, the conclusions of Leeflang’s metaanalysis and another recent study suggested that serum GM did not effectively rule in or rule out IA.8,38 In our metaanalysis, for proven or probable cases, PLR and NLR succeeded in passing the threshold index and generated large and often conclusive shifts from pretest to PTP. For example, in an at-risk case with a disease prevalence of 12%, the probability of proven or probable IA reaches 67% when results are positive, whereas the probability decreases to 1% when negative.

An exploration of the reasons for heterogeneity, rather than the computation of summary measures, is an important goal of metaanalyses.15,16 The subgroup analyses identified several study qualities that account for some of the observed heterogeneity in our results. However, we noted that the differences in these items did not reach statistical significance for the SEN, indicating that quality problems had little influence on the true-positivity of BAL-GM test. Further studies should aim to improve study quality and thus decrease the risk of bias.

One probable cause of the heterogeneity is that different cutoffs were used in different studies. Although an optimal cutoff for positive BAL-GM has not been determined, some studies used different cutoffs to classify a continuous measurement as either positive or negative, whereas others reported intermediate results. However, we found that the proportion of variation due to threshold effects was very low (1%), suggesting that choosing only one of the reported cutoffs to incorporate in the metaanalyses was feasible. In other situations, variations in cutoffs due to differences in the BAL-GM assays may occur between studies. For example, some studies used a single sample above the cutoff as proving the test positive, where others used two consecutive samples. For the subgroup analyses of proven or probable cases, two consecutive samples had the higher value of SPE (0.94). If clinical conditions are feasible, we recommend performing two BAL-GM samples to rule out diagnoses for at-risk patients.

We also provided the pooled performance measures at different cutoffs, as seen in Table 3. However, the highest SEN value of BAL-GM was only 0.86, when studies with a cutoff of 0.5 were included. The discrepancy in the results between the overall analysis and the cutoff subgroup analyses was mostly related to the number of included studies. When a cutoff of 0.5 was used, we included eight studies.21-23,25-29 However, the false-negative values were very small in the remaining five studies, in which three had a value of zero and the remaining two had a value of one and four.20,24,30,31 If all these studies had used a cutoff value of 0.5, the false-negative values may have been lower or unchanged, leading to increased or unchanged SEN values. In the reanalysis of 13 studies with a cutoff of 0.5, we found that the pooled SEN was at least 0.92, which was greater than that of the overall analysis. The discrepancy in the value of pooled NLR resulted from similar reasons.

Table Graphic Jump Location
Table 3 —Pooled Sensitivity and Specificity of the Included Studies for Proven or Probable IA

NLR = negative likelihood ratio; PLR = positive likelihood ratio; SEN = sensitivity; SPE = specificity. See Table 1 for expansion of other abbreviation.

The use of prophylactic or empiric antifungal treatment may be another possible cause of variable test performance.2 However, the effects of antifungal intervention on the presence of BAL-GM for diagnosing IA were controversial. Some studies found a trend toward better performance of the BAL-GM assay in patients receiving antifungal therapy,21,23,28 other studies failed to reveal any significant impact on the BAL-GM assay,24,29 and still others showed that antifungal agents seemed to lower the SEN of BAL-GM.20,25,27 For proven or probable cases, the antifungal intervention had some influence on the SEN of BAL-GM, with 0.87 and 0.98 for those with and without treatment, respectively (P = .23). The results of BAL-GM in our analyses might be explained on the basis that administration of antifungal drugs may lower the residual fungal burden in lung tissue and therefore diminish the SEN of enzyme-linked immunosorbent assay.39,40 The reasons for the discrepancies among individual studies could be partly related to the various performance characteristics of the antifungal drugs applied in the different studies, as mentioned in an experimental model.39 In a rabbit IA model, BAL-GM values were significantly reduced in animals treated with triazoles or polyenes, whereas they were persistently elevated in echinocandin-treated animals.39 We also found that antifungal agents seemed to lower the SPE of BAL-GM, with 0.92 and 0.98 for those with and without treatment, respectively (P = .02). This statistical significance may be the result of a clinical (therapeutic) bias, because possible IA cases receiving antifungal treatment may have been infected with other fungi containing a cross-reactive GM, or may have received β-lactam antibiotics. The β-lactams agents, which contain fermentation products of Penicillium spp, could cause false-positive results9; however, we failed to perform subgroup analyses for lack of eligible data in our analysis.

Our study had several important limitations. First, we acknowledged that the metaanalyses included a relatively small number of patients in the individual studies. Although we aimed to retrieve additional data from investigators, and failed to find any publication bias, it is unavoidable that some missing and unpublished data may still exist. Second, the exclusion of non-English-language studies and studies with fewer than 10 patients may have led to a bias in effect size. However, the addition of these exclusive studies to our 13 studies suggests that the overall results were similar. Further, misclassification bias can occur. IA is not always diagnosed by either histopathologic or cytopathologic examination. Actually, the definition of probable IA based on the presence of the host factor, clinical criterion, and mycologic criterion has been expanded, whereas the definition of possible IA has been diminished.6 This issue regarding accuracy of diagnosis can cause misclassification and discrepancy, leading to biased results. Lastly, we could not perform neutropenic subgroup analyses because of a lack of eligible data. Neutropenic status is an important covariate because negative or lower GM seemed to occur more often in moderately neutropenic and nonneutropenic patients.41 Further studies focused on the impact of treatment agents and neutropenic status are needed.

In conclusion, the current metaanalyses suggest that the BAL-GM assay can be used as a major diagnostic method with excellent accuracy. Because the BAL-GM assay is not absolutely sensitive and specific for IA, the results of BAL-GM should be interpreted in parallel with clinical findings for IA cases.

Author contributions:Dr Guo: contributed to the conception and design of the study, acquisition and interpretation of the data, and drafting of the article, and gave final approval of the version to be published.

Dr Chen: contributed to acquisition, analysis, and interpretation of the data and revision of the article, and gave final approval of the version to be published.

Dr Wang: contributed to the conception and design of the study, analysis and interpretation of the data, and drafting and revision of the article, and gave final approval of the version to be published.

Dr Qin: contributed to acquisition, analysis, and interpretation of the data and revision of the article, and gave final approval of the version to be published.

Dr Wu: contributed to acquisition, analysis, and interpretation of the data and revision of the article, and gave final approval of the version to be published.

Dr Kong: contributed to acquisition, analysis, and interpretation of the data and revision of the article, and gave final approval of the version to be published.

Financial/nonfinancial disclosures: The authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

AUC

area under the curve

BALF

BAL fluid

EORTC/MSG

European Organization for Research and Treatment of Cancer/Mycoses Study Group

GM

galactomannan

IA

invasive aspergillosis

NLR

negative likelihood ratio

PLR

positive likelihood ratio

PTP

posttest probability

QUADAS

quality assessment for studies of diagnostic accuracy

SEN

sensitivity

SPE

specificity

SROC

summary receiver operating characteristic

STARD

standards for reporting diagnostic accuracy

Ascioglu S, Rex JH, de Pauw B, et al; Invasive Fungal Infections Cooperative Group of the European Organization for Research and Treatment of Cancer Invasive Fungal Infections Cooperative Group of the European Organization for Research and Treatment of Cancer Mycoses Study Group of the National Institute of Allergy and Infectious Diseases Mycoses Study Group of the National Institute of Allergy and Infectious Diseases Defining opportunistic invasive fungal infections in immunocompromised patients with cancer and hematopoietic stem cell transplants: an international consensus. Clin Infect Dis. 2002;341:7-14. [CrossRef] [PubMed]
 
Walsh TJ, Anaissie EJ, Denning DW, et al; Infectious Diseases Society of America Infectious Diseases Society of America Treatment of aspergillosis: clinical practice guidelines of the Infectious Diseases Society of America. Clin Infect Dis. 2008;463:327-360. [CrossRef] [PubMed]
 
Meersseman W, Lagrou K, Maertens J, et al. Galactomannan in bronchoalveolar lavage fluid: a tool for diagnosing aspergillosis in intensive care unit patients. Am J Respir Crit Care Med. 2008;1771:27-34. [CrossRef] [PubMed]
 
Cornillet A, Camus C, Nimubona S, et al. Comparison of epidemiological, clinical, and biological features of invasive aspergillosis in neutropenic and nonneutropenic patients: a 6-year survey. Clin Infect Dis. 2006;435:577-584. [CrossRef] [PubMed]
 
Hope WW, Walsh TJ, Denning DW. Laboratory diagnosis of invasive aspergillosis. Lancet Infect Dis. 2005;510:609-622. [CrossRef] [PubMed]
 
De Pauw B, Walsh TJ, Donnelly JP, et al; European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group 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;4612:1813-1821. [CrossRef] [PubMed]
 
Pfeiffer CD, Fine JP, Safdar N. Diagnosis of invasive aspergillosis using a galactomannan assay: a meta-analysis. Clin Infect Dis. 2006;4210:1417-1427. [CrossRef] [PubMed]
 
Leeflang MM, Debets-Ossenkopp YJ, Visser CE, et al. Galactomannan detection for invasive aspergillosis in immunocompromized patients. Cochrane Database Syst Rev. 2008;4:CD007394
 
Wheat LJ, Walsh TJ. Diagnosis of invasive aspergillosis by galactomannan antigenemia detection using an enzyme immunoassay. Eur J Clin Microbiol Infect Dis. 2008;274:245-251. [CrossRef] [PubMed]
 
Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol. 2003;3:25. [CrossRef] [PubMed]
 
Bossuyt PM, Reitsma JB, Bruns DE, et al; Standards for Reporting of Diagnostic Accuracy Standards for Reporting of Diagnostic Accuracy Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. BMJ. 2003;3267379:41-44. [CrossRef] [PubMed]
 
Reitsma JB, Glas AS, Rutjes AW, Scholten RJ, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol. 2005;5810:982-990. [CrossRef] [PubMed]
 
Jaeschke R, Guyatt GH, Sackett DL. Users’ guides to the medical literature. III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? The Evidence-Based Medicine Working Group. JAMA. 1994;2719:703-707. [CrossRef] [PubMed]
 
Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;2111:1539-1558. [CrossRef] [PubMed]
 
Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heterogeneity in systematic reviews of diagnostic tests. Stat Med. 2002;2111:1525-1537. [CrossRef] [PubMed]
 
Petitti DB. Approaches to heterogeneity in meta-analysis. Stat Med. 2001;2023:3625-3633. [CrossRef] [PubMed]
 
Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol. 2005;589:882-893. [CrossRef] [PubMed]
 
Fagan TJ. Letter: Nomogram for Bayes theorem. N Engl J Med. 1975;2935:257. [PubMed]
 
Dwamena Ben A. Midas: A Program for Meta-analytical Integration of Diagnostic Accuracy Studies in Stata. 2007; Ann Arbor, MI Division of Nuclear Medicine, Department of Radiology, University of Michigan Medical School
 
Becker MJ, Lugtenburg EJ, Cornelissen JJ, Van Der Schee C, Hoogsteden HC, De Marie S. Galactomannan detection in computerized tomography-based broncho-alveolar lavage fluid and serum in haematological patients at risk for invasive pulmonary aspergillosis. Br J Haematol. 2003;1213:448-457. [CrossRef] [PubMed]
 
Bergeron A, Belle A, Sulahian A, et al. Contribution of galactomannan antigen detection in BAL to the diagnosis of invasive pulmonary aspergillosis in patients with hematologic malignancies. Chest. 2010;1372:410-415. [CrossRef] [PubMed]
 
Clancy CJ, Jaber RA, Leather HL, et al. Bronchoalveolar lavage galactomannan in diagnosis of invasive pulmonary aspergillosis among solid-organ transplant recipients. J Clin Microbiol. 2007;456:1759-1765. [CrossRef] [PubMed]
 
Desai R, Ross LA, Hoffman JA. The role of bronchoalveolar lavage galactomannan in the diagnosis of pediatric invasive aspergillosis. Pediatr Infect Dis J. 2009;284:283-286. [CrossRef] [PubMed]
 
Fréalle E, Decrucq K, Botterel F, et al. Diagnosis of invasive aspergillosis using bronchoalveolar lavage in haematology patients: influence of bronchoalveolar lavage human DNA content on real-time PCR performance. Eur J Clin Microbiol Infect Dis. 2009;283:223-232. [CrossRef] [PubMed]
 
Husain S, Paterson DL, Studer SM, et al. Aspergillus galactomannan antigen in the bronchoalveolar lavage fluid for the diagnosis of invasive aspergillosis in lung transplant recipients. Transplantation. 2007;8310:1330-1336. [CrossRef] [PubMed]
 
Kimura S, Odawara J, Aoki T, Yamakura M, Takeuchi M, Matsue K. Detection of sputum Aspergillus galactomannan for diagnosis of invasive pulmonary aspergillosis in haematological patients. Int J Hematol. 2009;904:463-470. [CrossRef] [PubMed]
 
Maertens J, Maertens V, Theunissen K, et al. Bronchoalveolar lavage fluid galactomannan for the diagnosis of invasive pulmonary aspergillosis in patients with hematologic diseases. Clin Infect Dis. 2009;4911:1688-1693. [CrossRef] [PubMed]
 
Musher B, Fredricks D, Leisenring W, Balajee SA, Smith C, Marr KA. Aspergillus galactomannan enzyme immunoassay and quantitative PCR for diagnosis of invasive aspergillosis with bronchoalveolar lavage fluid. J Clin Microbiol. 2004;4212:5517-5522. [CrossRef] [PubMed]
 
Penack O, Rempf P, Graf B, Blau IW, Thiel E. Aspergillus galactomannan testing in patients with long-term neutropenia: implications for clinical management. Ann Oncol. 2008;195:984-989. [CrossRef] [PubMed]
 
Sanguinetti M, Posteraro B, Pagano L, et al. Comparison of real-time PCR, conventional PCR, and galactomannan antigen detection by enzyme-linked immunosorbent assay using bronchoalveolar lavage fluid samples from hematology patients for diagnosis of invasive pulmonary aspergillosis. J Clin Microbiol. 2003;418:3922-3925. [CrossRef] [PubMed]
 
Shahid M, Malik A, Bhargava R. Bronchogenic carcinoma and secondary aspergillosis—common yet unexplored: evaluation of the role of bronchoalveolar lavage-polymerase chain reaction and some nonvalidated serologic methods to establish early diagnosis. Cancer. 2008;1133:547-558. [CrossRef] [PubMed]
 
Rapp RP. Changing strategies for the management of invasive fungal infections. Pharmacotherapy. 2004;242 Pt 22Pt2:4S-28S. [CrossRef] [PubMed]
 
Hummel M, Rudert S, Hof H, Hehlmann R, Buchheidt D. Diagnostic yield of bronchoscopy with bronchoalveolar lavage in febrile patients with hematologic malignancies and pulmonary infiltrates. Ann Hematol. 2008;874:291-297. [CrossRef] [PubMed]
 
Kuehnhardt D, Hannemann M, Schmidt B, Heider U, Possinger K, Eucker J. Therapeutic implication of BAL in patients with neutropenia. Ann Hematol. 2009;8812:1249-1256. [CrossRef] [PubMed]
 
Hohenthal U, Itälä M, Salonen J, et al. Bronchoalveolar lavage in immunocompromised patients with haematological malignancy—value of new microbiological methods. Eur J Haematol. 2005;743:203-211. [CrossRef] [PubMed]
 
Leeflang MM, Debets-Ossenkopp YJ, Visser CE, Bossuyt PM. Meta-analysis of diagnostic test accuracy. Clin Infect Dis. 2006;439:1220-1221. [CrossRef] [PubMed]
 
Kauffman HF, Beaumont F, Meurs H, van der Heide S, de Vries K. Comparison of antibody measurements against Aspergillus fumigatus by means of double-diffusion and enzyme-linked immunosorbent assay (ELISA). J Allergy Clin Immunol. 1983;723:255-261. [CrossRef] [PubMed]
 
Hidalgo A, Parody R, Martino R, et al. Correlation between high-resolution computed tomography and galactomannan antigenemia in adult hematologic patients at risk for invasive aspergillosis. Eur J Radiol. 2009;711:55-60. [CrossRef] [PubMed]
 
Francesconi A, Kasai M, Petraitiene R, et al. Characterization and comparison of galactomannan enzyme immunoassay and quantitative real-time PCR assay for detection of Aspergillus fumigatus in bronchoalveolar lavage fluid from experimental invasive pulmonary aspergillosis. J Clin Microbiol. 2006;447:2475-2480. [CrossRef] [PubMed]
 
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;1903:641-649. [CrossRef] [PubMed]
 
Cordonnier C, Botterel F, Ben Amor R, et al. Correlation between galactomannan antigen levels in serum and neutrophil counts in haematological patients with invasive aspergillosis. Clin Microbiol Infect. 2009;151:81-86. [CrossRef] [PubMed]
 

Figures

Figure Jump LinkFigure 1. Overall quality assessment of all 13 included studies. Data are presented as stacked bars for each quality item, including modified quality assessment for studies of diagnostic accuracy (QUADAS) criteria.Grahic Jump Location
Figure Jump LinkFigure 2. SROC curves from the bivariate model for proven or probable cases (A) and proven cases (B), respectively. The smaller region (confidence contour) contains likely combinations of the mean value of sensitivity and specificity. The wider region (prediction contour) demonstrates more uncertainty as to where the likely values of sensitivity and specificity might occur for individual studies. SROC = summary receiver operating characteristic.Grahic Jump Location
Figure Jump LinkFigure 3. Forest plot of subgroup analyses for sensitivity and specificity. Antifungal = antifungal intervention; Bias = the incorporation bias; Consec = consecutive or random; Criteria2008 = revised EORTC/MSG criteria in 2008; HM = hematologic malignancy; OneSample = one single sample testing for positivity; Prosp = prospective design. *P < .05; **P < .01; ***P < .001.Grahic Jump Location
Figure Jump LinkFigure 4. Fagan’s nomogram for calculating posttest probabilities (PTPs). A straight edge was used to link the pretest probability of invasive aspergillosis (IA) with the PTP, by crossing the likelihood ratio line at a point that describes the results obtained.Grahic Jump Location

Tables

Table Graphic Jump Location
Table 1 —Characteristics and Quality of 13 Studies Included in the Metaanalysis of Diagnosis of IA using BAL-GM

BAL-GM = BAL-galactomannan; BC = bronchogenic carcinoma; CoR = consecutive or random; EORTC/MSG = European Organization of the Research and Treatment of Cancer/Mycoses Study Group; HM = hematologic malignancy; HSCT = hematopoietic stem cell transplant; IA = invasive aspergillosis; IC = immunocompromised; IFI = invasive fungal infection; SOT = solid-organ transplant.

a 

Incorporation of bias by applying a positive BAL-GM result to the diagnostic criteria.

b 

The second study in Becker’s report.

c 

Healthy controls were excluded.

Table Graphic Jump Location
Table 2 —Characteristics and Results of 13 Studies Included in the Metaanalysis of Diagnosis of IA Using BAL-GM

“Samples” required refers to the number of samples required for positivity. STARD = standards for reporting diagnostic accuracy. See Table 1 for expansion of other abbreviations.

a 

The second study in Becker’s report.

b 

Healthy controls were excluded.

Table Graphic Jump Location
Table 3 —Pooled Sensitivity and Specificity of the Included Studies for Proven or Probable IA

NLR = negative likelihood ratio; PLR = positive likelihood ratio; SEN = sensitivity; SPE = specificity. See Table 1 for expansion of other abbreviation.

References

Ascioglu S, Rex JH, de Pauw B, et al; Invasive Fungal Infections Cooperative Group of the European Organization for Research and Treatment of Cancer Invasive Fungal Infections Cooperative Group of the European Organization for Research and Treatment of Cancer Mycoses Study Group of the National Institute of Allergy and Infectious Diseases Mycoses Study Group of the National Institute of Allergy and Infectious Diseases Defining opportunistic invasive fungal infections in immunocompromised patients with cancer and hematopoietic stem cell transplants: an international consensus. Clin Infect Dis. 2002;341:7-14. [CrossRef] [PubMed]
 
Walsh TJ, Anaissie EJ, Denning DW, et al; Infectious Diseases Society of America Infectious Diseases Society of America Treatment of aspergillosis: clinical practice guidelines of the Infectious Diseases Society of America. Clin Infect Dis. 2008;463:327-360. [CrossRef] [PubMed]
 
Meersseman W, Lagrou K, Maertens J, et al. Galactomannan in bronchoalveolar lavage fluid: a tool for diagnosing aspergillosis in intensive care unit patients. Am J Respir Crit Care Med. 2008;1771:27-34. [CrossRef] [PubMed]
 
Cornillet A, Camus C, Nimubona S, et al. Comparison of epidemiological, clinical, and biological features of invasive aspergillosis in neutropenic and nonneutropenic patients: a 6-year survey. Clin Infect Dis. 2006;435:577-584. [CrossRef] [PubMed]
 
Hope WW, Walsh TJ, Denning DW. Laboratory diagnosis of invasive aspergillosis. Lancet Infect Dis. 2005;510:609-622. [CrossRef] [PubMed]
 
De Pauw B, Walsh TJ, Donnelly JP, et al; European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group 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;4612:1813-1821. [CrossRef] [PubMed]
 
Pfeiffer CD, Fine JP, Safdar N. Diagnosis of invasive aspergillosis using a galactomannan assay: a meta-analysis. Clin Infect Dis. 2006;4210:1417-1427. [CrossRef] [PubMed]
 
Leeflang MM, Debets-Ossenkopp YJ, Visser CE, et al. Galactomannan detection for invasive aspergillosis in immunocompromized patients. Cochrane Database Syst Rev. 2008;4:CD007394
 
Wheat LJ, Walsh TJ. Diagnosis of invasive aspergillosis by galactomannan antigenemia detection using an enzyme immunoassay. Eur J Clin Microbiol Infect Dis. 2008;274:245-251. [CrossRef] [PubMed]
 
Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol. 2003;3:25. [CrossRef] [PubMed]
 
Bossuyt PM, Reitsma JB, Bruns DE, et al; Standards for Reporting of Diagnostic Accuracy Standards for Reporting of Diagnostic Accuracy Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. BMJ. 2003;3267379:41-44. [CrossRef] [PubMed]
 
Reitsma JB, Glas AS, Rutjes AW, Scholten RJ, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol. 2005;5810:982-990. [CrossRef] [PubMed]
 
Jaeschke R, Guyatt GH, Sackett DL. Users’ guides to the medical literature. III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? The Evidence-Based Medicine Working Group. JAMA. 1994;2719:703-707. [CrossRef] [PubMed]
 
Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;2111:1539-1558. [CrossRef] [PubMed]
 
Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heterogeneity in systematic reviews of diagnostic tests. Stat Med. 2002;2111:1525-1537. [CrossRef] [PubMed]
 
Petitti DB. Approaches to heterogeneity in meta-analysis. Stat Med. 2001;2023:3625-3633. [CrossRef] [PubMed]
 
Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol. 2005;589:882-893. [CrossRef] [PubMed]
 
Fagan TJ. Letter: Nomogram for Bayes theorem. N Engl J Med. 1975;2935:257. [PubMed]
 
Dwamena Ben A. Midas: A Program for Meta-analytical Integration of Diagnostic Accuracy Studies in Stata. 2007; Ann Arbor, MI Division of Nuclear Medicine, Department of Radiology, University of Michigan Medical School
 
Becker MJ, Lugtenburg EJ, Cornelissen JJ, Van Der Schee C, Hoogsteden HC, De Marie S. Galactomannan detection in computerized tomography-based broncho-alveolar lavage fluid and serum in haematological patients at risk for invasive pulmonary aspergillosis. Br J Haematol. 2003;1213:448-457. [CrossRef] [PubMed]
 
Bergeron A, Belle A, Sulahian A, et al. Contribution of galactomannan antigen detection in BAL to the diagnosis of invasive pulmonary aspergillosis in patients with hematologic malignancies. Chest. 2010;1372:410-415. [CrossRef] [PubMed]
 
Clancy CJ, Jaber RA, Leather HL, et al. Bronchoalveolar lavage galactomannan in diagnosis of invasive pulmonary aspergillosis among solid-organ transplant recipients. J Clin Microbiol. 2007;456:1759-1765. [CrossRef] [PubMed]
 
Desai R, Ross LA, Hoffman JA. The role of bronchoalveolar lavage galactomannan in the diagnosis of pediatric invasive aspergillosis. Pediatr Infect Dis J. 2009;284:283-286. [CrossRef] [PubMed]
 
Fréalle E, Decrucq K, Botterel F, et al. Diagnosis of invasive aspergillosis using bronchoalveolar lavage in haematology patients: influence of bronchoalveolar lavage human DNA content on real-time PCR performance. Eur J Clin Microbiol Infect Dis. 2009;283:223-232. [CrossRef] [PubMed]
 
Husain S, Paterson DL, Studer SM, et al. Aspergillus galactomannan antigen in the bronchoalveolar lavage fluid for the diagnosis of invasive aspergillosis in lung transplant recipients. Transplantation. 2007;8310:1330-1336. [CrossRef] [PubMed]
 
Kimura S, Odawara J, Aoki T, Yamakura M, Takeuchi M, Matsue K. Detection of sputum Aspergillus galactomannan for diagnosis of invasive pulmonary aspergillosis in haematological patients. Int J Hematol. 2009;904:463-470. [CrossRef] [PubMed]
 
Maertens J, Maertens V, Theunissen K, et al. Bronchoalveolar lavage fluid galactomannan for the diagnosis of invasive pulmonary aspergillosis in patients with hematologic diseases. Clin Infect Dis. 2009;4911:1688-1693. [CrossRef] [PubMed]
 
Musher B, Fredricks D, Leisenring W, Balajee SA, Smith C, Marr KA. Aspergillus galactomannan enzyme immunoassay and quantitative PCR for diagnosis of invasive aspergillosis with bronchoalveolar lavage fluid. J Clin Microbiol. 2004;4212:5517-5522. [CrossRef] [PubMed]
 
Penack O, Rempf P, Graf B, Blau IW, Thiel E. Aspergillus galactomannan testing in patients with long-term neutropenia: implications for clinical management. Ann Oncol. 2008;195:984-989. [CrossRef] [PubMed]
 
Sanguinetti M, Posteraro B, Pagano L, et al. Comparison of real-time PCR, conventional PCR, and galactomannan antigen detection by enzyme-linked immunosorbent assay using bronchoalveolar lavage fluid samples from hematology patients for diagnosis of invasive pulmonary aspergillosis. J Clin Microbiol. 2003;418:3922-3925. [CrossRef] [PubMed]
 
Shahid M, Malik A, Bhargava R. Bronchogenic carcinoma and secondary aspergillosis—common yet unexplored: evaluation of the role of bronchoalveolar lavage-polymerase chain reaction and some nonvalidated serologic methods to establish early diagnosis. Cancer. 2008;1133:547-558. [CrossRef] [PubMed]
 
Rapp RP. Changing strategies for the management of invasive fungal infections. Pharmacotherapy. 2004;242 Pt 22Pt2:4S-28S. [CrossRef] [PubMed]
 
Hummel M, Rudert S, Hof H, Hehlmann R, Buchheidt D. Diagnostic yield of bronchoscopy with bronchoalveolar lavage in febrile patients with hematologic malignancies and pulmonary infiltrates. Ann Hematol. 2008;874:291-297. [CrossRef] [PubMed]
 
Kuehnhardt D, Hannemann M, Schmidt B, Heider U, Possinger K, Eucker J. Therapeutic implication of BAL in patients with neutropenia. Ann Hematol. 2009;8812:1249-1256. [CrossRef] [PubMed]
 
Hohenthal U, Itälä M, Salonen J, et al. Bronchoalveolar lavage in immunocompromised patients with haematological malignancy—value of new microbiological methods. Eur J Haematol. 2005;743:203-211. [CrossRef] [PubMed]
 
Leeflang MM, Debets-Ossenkopp YJ, Visser CE, Bossuyt PM. Meta-analysis of diagnostic test accuracy. Clin Infect Dis. 2006;439:1220-1221. [CrossRef] [PubMed]
 
Kauffman HF, Beaumont F, Meurs H, van der Heide S, de Vries K. Comparison of antibody measurements against Aspergillus fumigatus by means of double-diffusion and enzyme-linked immunosorbent assay (ELISA). J Allergy Clin Immunol. 1983;723:255-261. [CrossRef] [PubMed]
 
Hidalgo A, Parody R, Martino R, et al. Correlation between high-resolution computed tomography and galactomannan antigenemia in adult hematologic patients at risk for invasive aspergillosis. Eur J Radiol. 2009;711:55-60. [CrossRef] [PubMed]
 
Francesconi A, Kasai M, Petraitiene R, et al. Characterization and comparison of galactomannan enzyme immunoassay and quantitative real-time PCR assay for detection of Aspergillus fumigatus in bronchoalveolar lavage fluid from experimental invasive pulmonary aspergillosis. J Clin Microbiol. 2006;447:2475-2480. [CrossRef] [PubMed]
 
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;1903:641-649. [CrossRef] [PubMed]
 
Cordonnier C, Botterel F, Ben Amor R, et al. Correlation between galactomannan antigen levels in serum and neutrophil counts in haematological patients with invasive aspergillosis. Clin Microbiol Infect. 2009;151:81-86. [CrossRef] [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543