REVISED

Comparative performance of Aspergillus galactomannan ELISA and PCR in sputum from

patients with ABPA and CPA

Running Head: Aspergillus galactomannan and DNA concentrations in sputum.

Samuel Fayemiwo, a, b, c, # Caroline B. Moore, b, c Philip Foden, d David W. Denning c , e and

Malcolm D. Richardson. b, c

Affiliations:

Department of Medical Microbiology and Parasitology, College of Medicine University of Ibadan,

Ibadan, Nigeria a ; Mycology Reference Centre, University Hospital of South Manchester, Manchester

Academic Health Sciences Centre, Manchester, U.K b ; Institute of inflammation and Repair, Faculty of

Medicine and Health Sciences, University of Manchester, UK c; Medical Statistics Department,

University Hospital of South Manchester, Manchester, UK d; National Aspergillosis Centre, University

Hospital of South Manchester, Manchester Academic Health Sciences Centre, University of

Manchester, UK e.

# Address correspondence to: Dr. Samuel Adetona Fayemiwo, dayteet@yahoo.com

1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

# Present Address: Department of Medical Microbiology and Parasitology, College of

Medicine, University of Ibadan, University College Hospital, Ibadan, Nigeria.

Authors’ contributions: DWD and MDR supervised the operational aspect of the study that

was conducted by SAF; SAF, CBM supervised the data entry; SAF, CBM, PF and DWD

undertook the analysis of data; SAF, DWD and MDR wrote the manuscript which was reviewed

by all authors.

This article contributes new data on the utility of Aspergillus galactomannan (GM) and

Aspergillus DNA detection in sputum from patients with allergic bronchopulmonary

aspergillosis (ABPA) and chronic pulmonary aspergillosis (CPA).

Keywords

Sputum; Galactomannan; Aspergillus DNA; Allergic bronchopulmonary aspergillosis; Chronic

pulmonary aspergillosis (CPA)

Highlights

o Receiver Operating Characteristic (ROC) curve analysis of sputum

galactomannan (GM) compared with PCR signal strength showed a cut-off

galactomannan index (GMI) of 6.5 in the Aspergillus GM assay.

o A GMI of 6.5 was associated with approximately 50% of strongly positive PCR

values

o Sputum characteristics could be used as good predictor of GMI values and PCR

signal

o The utility of GM in sputum for diagnosis or for following therapy is unclear

2

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

ABSTRACT

Objectives

Galactomannan (GM) and Aspergillus DNA detection are useful tools for the diagnosis of

invasive pulmonary aspergillosis (IPA), primarily in blood and bronchoscopy samples. This

study aimed to evaluate the utility of both markers for detection of Aspergillus in sputum from

patients with allergic bronchopulmonary aspergillosis (ABPA) and chronic pulmonary

aspergillosis (CPA).

Methods

ABPA or CPA demographic patient data were retrieved. This retrospective observational audit

included 159 patients with at least one sputum pair. 223 sputum sample pairs were analysed, as

well as six control samples for GM only. Real time PCR was performed following sputum DNA

extraction using the MycAssayTM Aspergillus kit and cycle thresholds were subtracted from 38 to

give positive values (transformed Ct, TCt).

Results

The mean age of the patients was 61.81 years (SD : +/- 11.06; range 29-100). One hundred and

twenty-six (79.2%) had CPA. Cultures were positive for fungi in 13.1% of the samples, and A.

fumigatus was the commonest (11.9%) fungus isolated. Receiver operating characteristic (ROC

curve) analysis of sputum GM comparing TCt of >0.0, and >2.0 to derive GMI cut-off values

showed a cut-off of 6.5. About 50% of sputa with strongly positive PCR values had GM values

>6.5. Two of six (33%) control samples had GM indices >6.5.

Conclusion

It is not clear that GM determinations in sputum are useful for diagnosis of either CPA or ABPA,

or following therapy.

3

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

INTRODUCTION

Aspergillus species may cause life-threatening respiratory infections including invasive

pulmonary aspergillosis, and allergic and chronic lung disease [1]. A small number of species of

Aspergillus are responsible for pulmonary aspergillosis including allergic bronchopulmonary

aspergillosis (ABPA) and chronic pulmonary aspergillosis (CPA) [1, 2]. While ABPA is a non-

invasive form of Aspergillus lung disease with hypersensitivity manifestations in patients who

have pre-existing atopy, asthma or cystic fibrosis [3], CPA manifests in patients who have

impaired immunity and chronic obstructive pulmonary diseases (COPD) [3]. Globally, about

6.8% of asthmatic patients and those with corticosteroid-complicated asthma develop ABPA [4].

Among cystic fibrosis patients, the prevalence of ABPA was found to be 2-15% [5], increasing

to 18 % among children [6]. The global burden of CPA complicating ABPA has been estimated

to be approximately 400,000 patients [7]. Most subjects with a simple aspergilloma are

asymptomatic except for haemoptysis which occurs in about 50-90% of cases [8]. Confirmation

of diagnosis of these infections requires assessment of clinical symptoms, positive Aspergillus

spp. culture from respiratory secretions, histological demonstration of fungal invasion, detection

of Aspergillus IgG antibody and other immunological tests as well as review of radiological

appearances [5, 9-11].

Most diagnostic tests for the detection of Aspergillus and other fungi utilise serum and

bronchoalveolar lavage (BAL) fluid for the confirmation of pulmonary aspergillosis and acute

invasive aspergillosis [11]. BAL requires bronchoscopy that is invasive and may be difficult to

obtain in critically ill patients [2]. However, sputum that is either expectorated or induced is non-

invasive and its diagnostic significance in term of volume yield, sensitivity and specificity cannot

be overemphasized [12-14].

4

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

The detection of galactomannan (GM), a polysaccharide that is released during Aspergillus

growth, is a useful and reliable non-invasive diagnostic test for screening and management of

aspergillosis and is often more sensitive than culture [13, 15]. The level of galactomannan in

body fluids such as serum and BAL is positively associated with the fungal burden of

Aspergillus [15]. Two studies have shown evaluated GM detection in sputum, one in patients

with invasive aspergillosis [16], one in cystic fibrosis [17]. The optimal cut-off for the

galactomannan index (GMI), as determined by constructing receiver operating characteristic

(ROC) curve, and for the diagnosis of invasive pulmonary aspergillosis had been set at equal or

> 0.5 for serum and 1.0 for BAL respectively [13, 16, 18-21]. There have been considerable

variations in the sensitivity and specificity of GMI detections in body fluids [15]. In BAL fluid,

the sensitivity ranges between 29 and 100% [1], whilst it ranges between 17 and 100% in serum

[13]. There is considerable controversy surrounding the optimal cut–off for respiratory secretion

GMI values. A cut-off value of 1.2 in sputum was proposed for Aspergillus GM in IPA among

haematological patients with sensitivity and specificity set at 100 and 62.2 % respectively [14,

16].

Real-Time PCR (RT-PCR) has now become an important diagnostic procedure for the diagnosis

of aspergillosis [12]. Detection of Aspergillus DNA has been applied to most body fluids

including blood, BAL and sputum specimens using commercially available kits, for example

MycAssay® Aspergillus [22] (Lab21, Cambridge, UK). Both GM and Aspergillus PCR have

shown higher sensitivity than culture for the detection of Aspergillus species in sputum in CF

patients [22, 23] and PCR in non-CF patients [2, 24]. A cycle threshold (Ct) value <36.0 is

regarded as positive for Aspergillus DNA in RT-PCR in respiratory fluids [25]. However, there

5

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

have been many inconsistencies in the reported RT-PCR Ct and GM index values for positive

detection of Aspergillus [16].

The survival of patients with aspergillosis is highly dependent on the ability to make an early and

accurate diagnosis. Since the isolation of organisms is rarely successful, this study was designed

to determine the GM index and PCR Ct values to evaluate the cut-off values for positive sputum

detection of Aspergillus in patients with ABPA and CPA.

MATERIALS AND METHODS

Study design

The audit was a retrospective observational study involving routinely collected sputum samples

from patients with ABPA and CPA, whether they have been treated or not. Baseline data for

patients with a diagnosis of ABPA or CPA, as previously described [2, 26], were collected and

included age, gender, sputum microscopy, fungal culture and current antifungal treatment. The

audit was registered with the University Hospital of South Manchester audit department.

Patient selection

Patients with ABPA or CPA who had submitted sputum samples for microscopy, fungal culture,

Aspergillus RT-PCR and detection of GM antigen by ELISA within a two weeks interval were

identified. Approximately 60 ABPA and CPA patients attend the National Aspergillosis Centre

(NAC) clinics each week, and some do not produce sputum. The pairs of values of results for

GM and PCR were recorded for each patient. Unselected sputa submitted by general

practitioners for routine bacterial culture were used as controls for GM only.

6

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

Laboratory procedures

The sputum samples were divided into aliquots for the routine direct microscopy, culture,

galactomannan testing and RT-PCR. Consistency of sputum specimen was also graded as

watery, salivary, mucoid, mucopurulent, and purulent and blood stained.

Microscopy: Sputum samples were digested with dithiothreitol (Sputasol®) [27] and then

examined under bright-field and fluorescent microscopy with calcofluor white. Aliquots of

digested sputum were cultured on Sabouraud dextrose agar plates (SAB) and incubated for a

total of 7 days at 300C and 370C [28]. Fungal identification was determined by conventional

mycological procedures, and by DNA sequencing when required [2, 5].

Galactomannan assay: GM levels were determined using a commercially available kit

(PlateliaTM Aspergillus, BIO-RAD, France). The kit has been validated only for serum and

bronchoalveolar lavage (BAL). A BAL sample result that was greater than a 1.0 index value was

considered to be positive while 0.5 was the validated index value for serum. However, the

performance for sputum has not been validated. The results interpreted according to the

manufacturer recommendations, are usually expressed as the ratio of optical densities (OD) of

either serum or BAL samples to the mean cut-off of the threshold control OD.

Aspergillus PCR assay: Aspergillus spp DNA was detected using the commercially available

real-time PCR diagnostic assay MycAssayTM Aspergillus (Lab21, Cambridge, UK) kit. The

control reagent was set at a cycle threshold (Ct) of 38.0. Interpretations of the Ct value

according to the manufacturer guidelines indicate that a Ct value of 36.0 or below is positive

while a Ct of 38.0 or greater is regarded as negative. However, based on local experience with

this assay an additional interpretation is that a Ct value between 36 and 38 has been assigned as a

7

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

weak positive which could represent aspergillosis or colonization. In the interpretation of the Ct

values, all Ct values were subtracted from 38 to provide a positive and increasing value

(‘Transformed Ct ‘(TCt)) correlating with the strength of signal and Aspergillus load. A TCt of

0.0 is equivalent to Ct of 38.0 or above.

Statistical analysis

The data generated were sorted and coded serially. The analysis was carried out using SPSS

version 17.0. Data was cleaned after entry by removing duplicated data that were misclassified as

disease-type. Summary statistics for both categorical and continuous variables were estimated.

Patients were selected based on their diagnosis of ABPA or CPA. Transformed Ct values were

set at either >0.0 or >2.0 denoting the positivity of Aspergillus DNA detection while GM index

cut-off values were set at 0.5-6.5. The sensitivity, specificity, positive predictive and negative

predictive values were calculated for each possibility of the cut-off values. The Youden J-

statistic (Youden Index (YI)), which is the maximum potential effectiveness of a biomarker was

used for the comparison of the different cut-off values [29], and its value ranges from 0.0 - 1.0.

Receiver Operating Characteristics (ROC) curves were constructed to compare and generate the

best cut-off values for GM index in the diagnosis of ABPA and CPA.

8

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

RESULTS

Patient Populations

During the audit period, a total of 159 patients who fulfilled the inclusion criteria and had at least

one sputum sample pair tested for GM antigen and Aspergillus DNA by RT-PCR were included

in the analysis. Other investigations retrieved from the database included results of direct

microscopy and culture. In addition subsequent follow-up visits with paired samples (visit 2 =39

patients; visit 3 = 13 patients; visit 4 = 4 patients) allowed a total 223 sputum samples that were

processed to be analysed. The mean age of the patients was 61.8 years +/- 11.1 (range 29-100).

Ninety-one (57.2%) of the patients were male (Table 1). One hundred and twenty-six patients

(79.2%) were diagnosed with CPA, while 33 had ABPA. The majority of the sputum samples

were muco-purulent, purulent and blood stained (69.7%), while others were mucoid and watery-

salivary (Table 1). Twenty-one (13.2%) of the sputum samples from the patients yielded a

positive culture of fungi, and Aspergillus fumigatus was the commonest species isolated (11.9%).

The PCR results and diagnostic accuracy of GM are shown in Tables 2 & 3 and Figs 1 & 3.

We were able to assess GM in six control samples, two of whom had COPD, three were not on

antibotics and one was on doxycycline, and background data was absent from two samples. The

GMI ranged from 0.9 to 11.2 (Mean 4.7). Two samples had high GMIs – one at 7.7 and one at

11.2.

Correlations between the values of the two biomarkers, GM index and TCt values, are illustrated

by scatter plots for all patients and each of the diseases. Fig 1 (all patients; both blue (ABPA)

and green (CPA) dots) shows a Spearman’s rank correlation of 0.40. This value is a moderately

positive correlation and is statistically significant (P<0.001). When these diseases are analysed

9

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

separately as shown in fig 1(green (CPA) and blue (ABPA) dots) , there is a moderate positive

Spearman’s rank correlation of 0.43 that is statistically significant (P <0.001) for CPA, while for

ABPA there is a weak positive Spearman’s rank correlation of 0.27 that is not statistically

significant (P = 0.14).

When TCt categories are compared with the GM index, the spread of the data is similar in each

of the groups, though it is clear that the median GM value is higher in the categories with larger

TCt values (Kruskal-Wallis P <0.001). There is no perfect discrimination between the two larger

TCt groups in the median of their GM index (Fig 3A). The median GM index for a TCt of 0.0

was 1.85 while the median GM index values for TCt of 0.1-1.99 (weakly positive) and ≥ 2.0 are

4.68 and 6.66 respectively. The degree of positivity is directly proportional to the GM index

values and the transformed cycle thresholds.

Galactomannan Index and PCR results for second visit samples

Thirthy-nine patients produced duplicate samples on their second visit (Supplementary data –

Table S1). Six (15.4%) of the sputum samples from the patients yielded positive microscopy and

positive culture of fungi on their second visit. The mean GM values of patients with positive

microscopy and cultures was 7.96. Out of the 39 duplicate samples, fifteen (38.5%) had TCt

>0.0 in sample 1 and 20 (51.3%) in sample 2 while 12 (30.8%) had TCt > 2.0 in sample 1 and 13

(33.3%) in sample 2. For the GMI values, thirty-two (82.1%) had GM > 0.5 in sample 1 and 36

(92.3%) in sample 2 while 13 (33.3%) had GM ≥6.5 in sample 1 and 18 (46.2%) (Supplementary

data – Table S2 and S3).

Figure S1 (Supplementary data) shows a non-significant (P=0.11), weak positive Spearman’s

rank correlation of 0.26 between the values of GM and TCt from the second visits. There is a

10

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

significant, moderate positive Spearman’s rank correlation of 0.54 between the Tct for the 1st and

2nd visits (P<0.001) as well as a significant Pearson correlation of 0.32 between the GM for the

duplicate samples in the 1st and 2nd visits (P= 0.047). There was a non-significant increase in GM

between the visits, with a mean of 4.54 at the first visit and a mean of 5.91 at the second visit

(paired t-test P=0.064). There was a small, non-significant increase in the TCt values between

the visits, with a median of 0 at the first visit and a median of 0.1 at the second visit (Wilcoxon

signed-rank test P=0.40).

Analysis of criteria for the detection of diseases by GM

We considered three possible constructs to define positive and negative samples to determine the

cut-off for GM. The first one focussed on narrow criteria for “true positive“. In this case, a “true

positive” was a patient with PCR TCt >0.0, along with at least one of positive microscopy and

positive culture. This was termed construct A, and 16 patients (15 CPA and 1 ABPA) were found

to be truly positive based on these criteria, and 143 negative. For this construct, the area under

the ROC curve was 0.732 (Figure 2). The second construct (B) was the same as construct A

regarding a “true positive” except that the PCR TCt >2.0 and 14 patients (13 CPA and 01 ABPA)

were positive. For this construct the area under the ROC curve was 0.757 (Figure 2). The third

construct for GM analysis that was considered as “true negative“ included only those samples in

which microscopy and culture were negative with a PCR TCt of 0.0. In this construct, eighty-

seven patients were truly negative and the area under the ROC curve was 0.729.

However, there were a few patients who were considered to be negative but had some

combination of positive microscopy, positive culture and PCR TCt that resulted in them not

being classified as a “true positive” or a “true negative”. These categories of patients (56 patients

11

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

from 159 samples in construct A and 58 patients from 159 samples in construct B) were

considered as intermediates (Supplementary data , Table S4) .

Construction of ROC curves to determine cut-off value

The optimal cut-offs for both the GM and PCR TCt values for the sputum specimens were

determined by the construction of ROC curves ( Figure 2) for all patients and their clinical

diagnosis. The best possible cut-off for both A and B constructs was a GMI of 6.5. At this cut-

off, the sensitivity and specificity of sputum GM were 68.8% and 79.7% for construct A while it

was 71.4% and 79.7% respectively for construct B. The Youden Index (J) for both A and B are

0.49 and 0.507 respectively (Table 3). If a cut-off level of 1.0 was considered, the sensitivity and

specificity would be 87.5% and 31.5% for construct A while construct B would be 92.9% and

31.7% respectively. With the “true negative“, the optimal cut-off was considered to be 4.5 from

the ROC. The sensitivity, specificity and Youden Index were 59.7%, 79.3% and 0.39

respectively (Table 3 and Figure 2.). When the category of patients considered as intermediates

were analysed using the GMI cut-off value of 6.5, twenty (35.7%) of them were ‘positive.’

A GMI of 6.5 was still the best possible cut-off for both A and B constructs for the duplicate

sputum specimens at the second visit but with lower specificities. At this cut-off value, the

sensitivity and specificity of sputum GM were 80.0% and 58.8%. for construct A whereas it was

100.0% and 58.3% respectively for construct B. Two of six control samples were positive using

this cut-off, confirming poor specificity. The sensitivity and specificity of sputum GM for the 3rd

visits were 100.0% and 72.7% for both construct A and B.

12

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

Analysis of GM index, microscopy and culture

The GMI was compared with the microscopy and culture results for all patients. The median

level of GM for the patients with positive fungal culture group was 6.11, and there is a positive

correlation between the two (Kruskal-Wallis P=0.002). The GM values were relatively similar

for the negative or normal flora group (median 3.23, IQR 0.69-6.22) and not done group (median

1.97, IQR 0.26-4.91), while the positive culture group had higher GM values (median 7.19, IQR

3.11-9.25) (Figure 3B).

There is also a statistically significant difference in the GMI between microscopy groups

(Kruskal-Wallis P=0.003). Twelve of the fifteen (80%) in the group where fungal elements were

seen had a positive culture of Aspergillus spp. The samples with positive microscopy had the

highest GM values (median 7.65, IQR 3.66-9.56) and the microscopy-negative and not done

samples had lower GM values (negative: median 3.87, IQR 0.70-6.43, not done: median 2.07,

IQR 0.26-4.68) (Figure 3C).

Relationship of sample characteristics, GM indices and PCR

The consistency of the sputum samples according to the Mycology Reference Centre Manchester

(MRCM) classification was associated with the predictive value of the two biomarkers as evident

in Table 4. Mucopurulent, purulent and blood-stained sputa were associated with higher levels of

both GMI and TCt. There was a significant association between purulent sputa and increased

level of GMI and detection of Aspergillus genomic DNA (P <0.05). A GMI of 6.5 was

associated with about 50% of strongly positive PCR values. The average mean GMI associated

with mucoid, mucopurulent, purulent and blood-stained sputa was 4.93 (Fig. 3D and 3E).

13

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

DISCUSSION

In sputum, the gold standard to confirm the presence of Aspergillus spp remains microscopy and

culture, but these methods are not highly sensitive, however positivity in microscopy and culture

of sputum is not always indicative of disease [13].

In our study, there was no perfect relationship between the GMI and PCR for the detection of

Aspergillus in the sputum of either patients with ABPA or CPA. This is reflected in the mean

GMI level of 3.81, and the mean transformed Ct of 1.49. This result does not concur with

previous studies on the diagnosis of pulmonary aspergillosis that validated a GM index of 1.0 for

BAL [18, 19, 30] and 1.2 for sputum (16). In the present study, forty-seven (29.6%) patients had

GMI < 1.0 whilst about 60.1% of them were negative for Aspergillus DNA. Even though PCR is

highly sensitive and specific, most of the patients in whom Aspergillus DNA was not detected in

their sputum had detectable GM, with a GMI which ranged from 2.0 to 14.0. A similar situation

is seen with our ‘control’ samples in which 33% had high GM indices. In fact the term control is

potentially misleading as GPs in the UK tend to submit samples for problematic patients and the

two patients with high GM indices could have some form of pulmonary or airways aspergillosis,

such as Aspergillus bronchitis. The prior study evaluating GM in sputum from patients suspected

of having invasive aspergillosis indicated that GMI sensitivity may be improved by testing

multiple samples and two consecutive positive GMI readings should be considered diagnostic

[20]. It was also discovered that PCR was more sensitive than GMI in our study as confirmed by

others [12, 31] though both biomarkers were significantly correlated.

The significance of the two biomarkers for the detection of Aspergillus antigen and DNA in the

sputum of patients cannot be over-emphasised. The mean GMI value was higher among CPA

patients than the ABPA. This might not be unconnected with the fact that CPA is associated with

14

282283284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

lung damage and pre-existing lung diseases that allow the airway or damaged lung tissue to be

colonized with Aspergillus spp. [9, 16, 32, 33]. ABPA, on the other hand, is characterised by

intense inflammation of the airway and colonization by conidia, hyphae and hyphal fragments of

Aspergillus [34, 35]. This fact is also supported in our study where seventeen CPA patients

produced a positive sputum culture of A. fumigatus, while only three ABPA patients had positive

sputum cultures. This is also in agreement with previous studies where some of the patients with

Aspergillus isolation from respiratory secretions could not fulfil the diagnostic criteria of

ABPA[36].

The characteristics and consistencies of sputum that have been adjudged to be a non-invasive

alternative to bronchoscopy and BAL in the diagnosis of respiratory infections were studied in

this audit. The sputum classification adopted by the MRCM was easier to assess and evaluate

than the previous Geckler classification used in a previous study [16]. Most of the sputa were

purulent and mucopurulent in nature. Purulent and blood stained sputum samples were

significantly associated with positive cultures and strong PCR signals among CPA than ABPA

patients, while GMI greater than 6.5 was more often positive in purulent and blood-stained

samples as well as muco-purulent sputum samples. Both PCR signals and GM were usually

negative in watery-salivary and mucoid samples, raising the prospect for laboratories of rejecting

these samples for further analysis or at least reporting the sputum character with the result,

indirectly suggesting that a negative result is a potential false negative.

Since we have no working gold standard for the diagnostic utility of GMI in our patient groups,

we created three possible constructs for the analysis of GM values. Construct A “true positivity

“considered TCt of >0.0 with positive microscopy or culture while Construct B considered

positivity with TCt of ≥2.0, in addition to positive culture or microscopy. These constructs only

15

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

captured about 10% of patients and so more than 90% of the patients were negative but might

reasonably considered not be “true negatives“. Not all the patients who had either microscopy or

culture positive were positive for both GM antigen and detection of Aspergillus DNA. These

categories of patients have not been truly captured. We need further studies to be able to define

the true positivity criteria for their diagnosis. We also considered “true negativity” with negative

PCR, GM less than 1.0 and TCt of 0.0. In this category, about 45% of patients were truly

negative. With all the inconsistencies in the results of the GM and PCR, we might need further

studies to validate the usefulness of measurement of GM antigen in the sputum for the screening

and diagnosis of CPA or ABPA.

To minimize these effects, we constructed the ROC curve for the two constructs and true

negativity to find the optimal cut-off for the positive GM. The optimal cut-off value of sputum

GMI as estimated from the ROC curve for both ABPA and CPA in the two possible constructs

was 6.5 (Figure 2). Using the cut-off value of 6.5, the sensitivity and specificity of

galactomannan antigen detection were 68.8 and 79.7% for construct A and 71.4 and 79.3% for

construct B respectively. However, the PPV was below 30% and the NPV were between 96-97%

(Table 3). This finding is in contrast with the previous study where 1.2 was the optimal cut-off

for sputum in the diagnosis of IPA [16]. Our small ‘control’ group may not be true controls.

Obtaining sputum from normal people, who would not naturally produce sputum, is challenging,

and bronchoscopy or induced sputum samples are not a reasonable substitute for this purpose.

In this study, a GMI of 1.0 was associated with high sensitivity above 90%, specificity below

30% and higher negative predictive value. However, this cut-off couldn’t have been optimal

because of the lower level of Youden J Index which was less than 0.2 in both constructs [29].

This inconsistency was also supported by varied sensitivities that ranged from 27-100% in serum

16

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

and 60-100% in sputum [15, 16, 20]. The reasons for these observations might be linked to

sample variability, the load of organism, other medications, antifungal resistance in the context

of antifungal therapy and patient host status [16, 20]. In order to validate the optimal cut-off for

Aspergillus GM in sputum, it is suggested that long term prospective studies involving larger

populations of ABPA or CPA patients with diverse predisposing factors should be carried out. In

addition to the non-invasive nature of sputum collection, a previous study had confirmed that

GM could be more easily detected in sputum than in other respiratory fluids [16]. Since there

was no adequate control group in the present study, it was technically difficult to assign the

optimal cut–off values for the GMI.

Aspergillus spp especially A. fumigatus and Penicillium spp. were isolated in about 13.2 % of the

sputa cultured from CPA and ABPA patients in our study. We had only three culture positive

sputa in ABPA patients. This rate was low compared to that found in previous studies where the

culture rate from sputum was as high as 56-80% [2, 26]. High volume culture that was adopted

by the MRCM could improve significantly the yield of Aspergilli from sputum samples [12].

There is a positive correlation between the median GM of 6.11 and positive culture of sputum.

This result supported the optimal cut-off value of 6.5 of our GM analysis. More than 50% of

patients with culture positive sputum had a GMI greater than 6.0.

A strongly positive PCR with TCt of ≥ 2.0 was significantly associated with positive sputum

culture. A GM value of >6.5 was strongly associated with about 50% of strongly positive PCR

results. As noted previously, a few of the positive sputum cultures did not concur with the

corresponding positive detection of Aspergillus DNA by PCR. Due to variability in the

sensitivity of PCR and increased significance of adopting TCt of >0.0 as weak signal of

17

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

Aspergillus detection, there is a need for urgent standardization and validation of PCR tests

across all laboratories as advocated by other studies [2, 15] .

Sputum consistencies could be a good predictor of positivity for microscopy, culture, GMI

values and Aspergillus detection by PCR. Mucopurulent, purulent and blood-stained sputum had

higher GM and PCR values. However, our study has shown that higher mean GMI values and

TCt values for PCR were significantly associated with purulent sputum samples.

In conclusion, mucopurulent, purulent and blood stained sputum had higher GM and PCR

values. A GMI value of >6.5 was associated with ~50% of strongly positive PCR values. We

have no control group in our study which is required for setting the optimal cut-off. There are

inconsistencies for both GM and PCR, so within patient reproducibility may be problematic (data

not shown). Based on the results presented in this study, it is not clear whether GM is useful in

sputum for diagnosis or for following therapy given the wide range of values, which did not

correspond with PCR or culture.

Future direction

To be able to have the definitive optimal cut-off value for the detection of Aspergillus DNA by

PCR and galactomannan antigen, more samples are needed for both index cases and the control

groups. It is not clear what the source of the GM found in such high quantities in sputum is. In

the future studies, possible confounding variables such as antibiotics (piperacillin-tazobactam,

ampicillin), many opportunistic fungi, some solutions containing gluconate, food containing

cereals products and soya beans should be evaluated [37, 38].

18

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

Limitations of the study

We didn’t have an adequate control group that could have been used to set up the optimal cut-off

values for both PCR and GM antigen detection on sputum.

Acknowledgement

We would like to acknowledge Europe Gilead Sciences Ltd. (Educational Scholarship) for

providing support for the University of Manchester Masters Medical Mycology degree

programme. We appreciate the support of Dr. Riina Rautemaa-Richardson, Dr. Lilyann Novak-

Frazer, Mr. Rikesh Masania and Mrs. Julie Morris (medical statistician).

Possible conflicts of interest

Dr. Fayemiwo or SAF has received full financial support from Europe Gilead Sciences Ltd for

his M.Sc. degree programme in Medical Mycology and has been paid for talks on behalf of

AstraZeneca and GSK.

Caroline Moore or CBM has received a travel grant from Astellas, been paid for talks on behalf

of Pfizer and has received grant support from Pfizer.

Philip Foden or PF doesn’t have anything to declare.

Dr Denning or DWD holds Founder shares in F2G Ltd a University of Manchester spin-out

antifungal discovery company, in Novocyt which markets the Myconostica real-time molecular

assays and has current grant support from the National Institute of Allergy and Infectious

Diseases, National Institute of Health Research, NorthWest Lung Centre Charity, Medical

19

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

Research Council, Global Action Fund for Fungal Infections and the Fungal Infection Trust. He

acts or has recently acted as a consultant to T2 Biosystems, GSK, Sigma Tau, Oxon

Epidemiology, Basilea, Pulmatrix and Pulmicort. In the last 3 years, he has been paid for talks on

behalf of Astellas, Dynamiker, Gilead, Merck and Pfizer. He is also a member of the Infectious

Disease Society of America Aspergillosis Guidelines and European Society for Clinical

Microbiology and Infectious Diseases Aspergillosis Guidelines groups.

Malcolm Richardson or MDR acts a consultant for Gilead Science eEurope, Astellas, MSD,

Pfizer and Basilea. He is a member of the European Society for Clinical Microbiology and

Infectious Diseases Aspergillosis Guidelines writing groups.

20

415

416

417

418

419

420

421

422

423

REFERENCES

1. Hope WW, Walsh TJ, Denning DW. Laboratory diagnosis of invasive aspergillosis. The

Lancet Infectious diseases 2005; 5:609-22.

2. Denning DW, Park S, Lass-Florl C, et al. High-frequency triazole resistance found In

nonculturable Aspergillus fumigatus from lungs of patients with chronic fungal disease.

Clinical infectious diseases : an official publication of the Infectious Diseases Society of

America 2011; 52:1123-9.

3. Kousha M, Tadi R, Soubani AO. Pulmonary aspergillosis: a clinical review. European

respiratory review : an official journal of the European Respiratory Society 2011; 20:156-

74.

4. Agarwal R, Chakrabarti A, Shah A, et al. Allergic bronchopulmonary aspergillosis: review of

literature and proposal of new diagnostic and classification criteria. Clinical and

experimental allergy : journal of the British Society for Allergy and Clinical Immunology

2013; 43:850-73.

5. Denning DW, Kibbler CC, Barnes RA. British Society for Medical Mycology proposed

standards of care for patients with invasive fungal infections. The Lancet Infectious

diseases 2003; 3:230-40.

6. Sharma VK, Raj D, Xess I, Lodha R, Kabra SK. Prevalence and risk factors for allergic

bronchopulmonary aspergillosis in Indian children with cystic fibrosis. Indian pediatrics

2014; 51:295-7.

7. Denning DW, Pleuvry A, Cole DC. Global burden of allergic bronchopulmonary aspergillosis

with asthma and its complication chronic pulmonary aspergillosis in adults. Medical

mycology 2013; 51:361-70.

21

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

8. Patterson KC, Strek ME. Diagnosis and treatment of pulmonary aspergillosis syndromes.

Chest 2014; 146:1358-68.

9. Zmeili OS, Soubani AO. Pulmonary aspergillosis: a clinical update. Qjm 2007; 100:317-34.

10. Saraceno JL, Phelps DT, Ferro TJ, Futerfas R, Schwartz DB. Chronic necrotizing pulmonary

aspergillosis: approach to management. Chest 1997; 112:541-8.

11. Agarwal R, Aggarwal AN, Sehgal IS, Dhooria S, Behera D, Chakrabarti A. Performance of

serum galactomannan in patients with allergic bronchopulmonary aspergillosis. Mycoses

2015; 58:408-12.

12. Fraczek MG, Kirwan MB, Moore CB, Morris J, Denning DW, Richardson MD. Volume

dependency for culture of fungi from respiratory secretions and increased sensitivity of

Aspergillus quantitative PCR. Mycoses 2014; 57:69-78.

13. Schelenz S, Barnes RA, Barton RC, et al. British Society for Medical Mycology best practice

recommendations for the diagnosis of serious fungal diseases. The Lancet Infectious

diseases 2015; 15:461-74.

14. Kimura S, Odawara A, Aoki T, Yamakura M, Takeuchi M, Matsunaga T. [Usefulness of

sputum galactomannan in diagnosis of pulmonary aspergillosis]. The Japanese journal of

antibiotics 2008; 61:99-101.

15. Fisher CE, Stevens AM, Leisenring W, Pergam SA, Boeckh M, Hohl TM. The serum

galactomannan index predicts mortality in hematopoietic stem cell transplant recipients

with invasive aspergillosis. Clinical infectious diseases : an official publication of the

Infectious Diseases Society of America 2013; 57:1001-4.

22

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

16. 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. International journal of hematology 2009; 90:463-70.

17. Baxter CG, Dunn G, Jones AM, et al. Novel immunologic classification of aspergillosis in

adult cystic fibrosis. The Journal of allergy and clinical immunology 2013; 132:560-

6.e10.

18. Khorvash F, Meidani M, Babaei L, Abbasi S, Ataei B, Yaran M. Galactomannan antigen

assay from bronchoalveolar lavage fluid in diagnosis of invasive pulmonary aspergillosis

in intensive care units patients. Advanced biomedical research 2014; 3:68.

19. Morrissey CO, Chen SC, Sorrell TC, et al. Galactomannan and PCR versus culture and

histology for directing use of antifungal treatment for invasive aspergillosis in high-risk

haematology patients: a randomised controlled trial. The Lancet Infectious diseases 2013;

13:519-28.

20. He H, Ding L, Chang S, Li F, Zhan Q. Value of consecutive galactomannan determinations

for the diagnosis and prognosis of invasive pulmonary aspergillosis in critically ill

chronic obstructive pulmonary disease. Medical mycology 2011; 49:345-51.

21. Richardson MD, Warnock DW. Fungal infection: diagnosis and management. John Wiley &

Sons, 2012.

22. Baxter CG, Rautemaa R, Jones AM, et al. Intravenous antibiotics reduce the presence of

Aspergillus in adult cystic fibrosis sputum. Thorax 2013; 68:652-7.

23. Baxter CG, Jones AM, Webb K, Denning DW. Homogenisation of cystic fibrosis sputum by

sonication--an essential step for Aspergillus PCR. J Microbiol Methods 2011; 85:75-81.

23

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

24. Langridge PJ, Sheehan RL, Denning DW. Microbial yield from physiotherapy assisted

sputum production in respiratory outpatients. BMC pulmonary medicine 2016; 16:23.

25. Tuon FF. A systematic literature review on the diagnosis of invasive aspergillosis using

polymerase chain reaction (PCR) from bronchoalveolar lavage clinical samples. Revista

iberoamericana de micologia 2007; 24:89-94.

26. Denning DW, Riniotis K, Dobrashian R, Sambatakou H. Chronic cavitary and fibrosing

pulmonary and pleural aspergillosis: case series, proposed nomenclature change, and

review. Clinical infectious diseases : an official publication of the Infectious Diseases

Society of America 2003; 1:S265-80.

27. Stokell JR, Khan A, Steck TR. Mechanical Homogenization Increases Bacterial

Homogeneity in Sputum. Journal of clinical microbiology 2014; 52:2340-5.

28. Denning DW, Pleuvry A, Cole DC. Global burden of chronic pulmonary aspergillosis

complicating sarcoidosis. The European respiratory journal 2013; 41:621-6.

29. Ruopp MD, Perkins NJ, Whitcomb BW, Schisterman EF. Youden Index and Optimal Cut‐

Point Estimated from Observations Affected by a Lower Limit of Detection. Biometrical

Journal 2008; 50:419-30.

30. Wichmann D, Kluge S. Invasive pulmonary aspergillosis in the ICU: an emerging disease?

Intensive care medicine 2013; 39:790.

31. Buchheidt D, Hummel M, Schleiermacher D, et al. Prospective clinical evaluation of a

LightCyclerTM‐mediated polymerase chain reaction assay, a nested‐PCR assay and a

galactomannan enzyme‐linked immunosorbent assay for detection of invasive

aspergillosis in neutropenic cancer patients and haematological stem cell transplant

recipients. British journal of haematology 2004; 125:196-202.

24

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

32. Smith NL, Denning DW. Underlying conditions in chronic pulmonary aspergillosis including

simple aspergilloma. The European respiratory journal 2011; 37:865-72.

33. Schweer KE, Bangard C, Hekmat K, Cornely OA. Chronic pulmonary aspergillosis. Mycoses

2014; 57:257-70.

34. Shah A, Panjabi C. Allergic aspergillosis of the respiratory tract. European respiratory review

: an official journal of the European Respiratory Society 2014; 23:8-29.

35. Wark P. Pathogenesis of allergic bronchopulmonary aspergillosis and an evidence-based

review of azoles in treatment. Respiratory medicine 2004; 98:915-23.

36. Liu JC, Modha DE, Gaillard EA. What is the clinical significance of filamentous fungi

positive sputum cultures in patients with cystic fibrosis? Journal of cystic fibrosis :

official journal of the European Cystic Fibrosis Society 2013; 12:187-93.

37. Girmenia C, Ballaro D, Martino P. Contamination by cardboard particles as a cause of false-

positive results for Aspergillus galactomannan enzyme immunoassays. Clinical and

vaccine immunology : CVI 2007; 14:929.

38. Lescher-Bru V, Cavalier A, Pernot-Marino E, et al. Aspergillus galactomannan antigen

detection with Platelia® Aspergillus: multiple positive antigenemia without Aspergillus

infection. J Mycol Med 1998; 8:112-3.

25

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

TABLES AND FIGURES

Table 1 Socio-demographics and characteristics of the sputum samples of the patients (GM –Galactomannan; TCt-Transformed Ct; FES- Fungal elements seen.)

Characteristics

Frequency Percentage (%)

Total Population 159 100

Age Group

21-30 1 0.6

31-40 6 3.8

41-50 17 10.7

51- 60 36 22.6

61-70 68 42.8

71-80 27 17.0

81-90 3 1.9

91-100 1 0.6

Sex

Male 91 57.2

Female 66 41.5

Unassigned 2 1.3

Diagnosis

ABPA 33 20.8

CPA 126 79.2

Microscopy

Fungal Element Not Seen (FENS)

89 56.0

Fungal Element Seen (FES)

15 9.4

ND 55 34.6

Culture

26

535

536537

538

Negative or normal flora

88 55.3

Aspergillus fumigatus 19 11.9

Aspergillus niger 1 0.6

Penicilium spp 1 0.6

Not done 50 31.4

Sputum positivity by PCR (TCt) Frequency Percentage (%)

TCt >0 64 40.3

TCt >2.0 45 28.3

Sputum consistency (n=76)

Watery 3 3.9

Salivary 1 1.3

Mucoid 19 25.0

Mucopurulent 15 19.7

Purulent 24 31.6

Blood stained 14 18.4

27

539

540

541

542

543

544

545

546

Table 2 Characteristics of CPA and ABPA patients (GM –Galactomannan; TCt-

Transformed Ct; FES- Fungal elements seen.)

CPA ABPA

N 126 33

M(%)/F(%)/U(%) 80 (63.5%)/44 (34.9%)/2 (1.6%)

22 (33.3%)/ 11(66.7%)

Median Age (Range)

63 (31-100) 64 (29-80)

GM (Mean/Median)

4.07/3.86 2.84/1.83

GMI >1.0 (%) 92(73.0%) 20 (60.6%)

TCt (Mean/Median)

1.48/0.00 1.53/0.00

TCt > 2.0 (%) 34(27.0%) 11(33.3%)

Microscopy (FES) N/%

15 (11.9%) 0 (%)

Postive Culture N/%

18 (14.3%) 3 (9.1%

28

547

548

549

550

551

552

553

554

555

Table 3 Results of the diagnostic accuracy of GM with different constructs of GM analysis

GM Sensitivity Specificity PPV NPV Youden’ J

statistics

1.0 Construct A 87.5 31.5 12.5 95.7 0.19

Construct B 92.9 31.7 11.6 97.9 0.25

True Negative 86.1 42.5 21.3 78.7 0.29

4.5 Construct A 68.8 65.0 18.0 94.9 0.34

Construct B 71.4 64.8 16.4 95.9 0.36

True Negative 59.7 79.3 70.5 70.4 0.39

6.5 Construct A 68.8 79.7 27.5 95.8 0.49

Construct B 71.4 79.3 25.0 96.6 0.507

True Negative 43.1 89.7 77.5 65.5 0.320

29

556

557

558

559

560

561

562

563

564

Table 4 Association of sputum consistency characteristics and biomarkers level. (GM –

Galactomannan and PCR (TCt) – Polymerase Chain Reaction - transformed Cycle threshold)

Characteristics N (%) Mean GMI (Range) PCR signal –

Mean-TCt (Range)

P-Value

Watery-salivary 4 (5.3) 2.3 (0.58-4.22) 0.26 (0.0-1.3) 0.171

Mucoid 19 (25.0) 3.05 (0.26-9.71) 1.4 (0.00-9.4) 0.446

Muco-purulent 15 (19.7) 6.32 (0.11-9.66) 2.1 (0.0-12.3) 0.271

Purulent 24 (31.6) 5.30 (0.59-14.79) 3.0 (0.0-14.2) 0.027

Blood stained 14 (18.4) 6.44 (1.62-12.44) 2.6 (0.0-8.2) 0.301

30

565

566

567

568

569

570

571

572

573

574

575

Figure 1: A scatter plot showing the correlation between Transformed Ct (TCt) and

Galactomannan (GM) index values for both CPA and ABPA patients. (CPA-Chronic pulmonary

aspergillosis; Ct- Cycle threshold; GM- Galactomannan; ABPA- Allergic broncho-pulmonary

aspergillosis)

31

576

577

578

579

580

581

582

583

584

0 20 40 60 80 1000

20

40

60

80

100

ROC curve

(100 - Specificity) %

Sen

sitiv

ity %

Construct A

Construct B

True Negative

Figure 2 Receiver operating characteristic (ROC) curve of sputum GM comparing

sensitivities and 1-specificity of true positivity with Tct >0.0, Tct >2.0 and true

negativity to derive GM index cut-off values. The arrow indicates optimal cut-off of 6.5.

Area under the curve (AUC) = 0.73 for Construct A, 0.76 for construct B and 0.73 for

true negative. CO = Best cut-off.

32

CO - 6.5

585

586

587

588

589

590

591

592

593

Figure 3A Box plots showing GM index values for the categories of transformed Ct.

Yellow line shows GMI of 6.5 while dashed line shows GM of 1.0. On the horizontal

axis, ‘0’ indicates negative PCR signal strength, ‘>0-2’ indicates weak PCR signal

strength, while ‘>2’ indicates strong PCR signal strength. The line across the box

represents the “ median “ , the box indicates the interquantile (IQ) range which contains

the middle 50% of the GM values, while whiskers are lines that extends from the upper

and lower edge of the box to the highest and lowest values which are not greater than

1.5times the IQ range.

33

594

595596597

598

599

600

601

602

603

604

FIGURE 3B Box plots showing GM index values for the culture groups. Yellow line

shows GMI of 6.5 while the green dashed line shows GMI of 1.0. (Galactomannan

index). The line across the box represents the “ median “ , the box indicates the

interquantile (IQ) range which contains the middle 50% of the GM values, while

whiskers are lines that extends from the upper and lower edge of the box to the highest

and lowest values which are not greater than 1.5times the IQ range. Outliers are values

outside the whiskers.

34

605

606

607608

609

610

611

612

613

614

615

Figure 3C Box plots showing GM index values for the microscopy groups. The yellow

line shows GMI of 6.5 while the dashed line shows GM of 1.0. On the horizontal axis,

‘FENS’ indicates Fungal elements not seen, ‘FES’ indicates Fungal element seen, while

‘ND’ indicates Not done. The line across the box represents the “ median “ , the box

indicates the interquantile (IQ) range which contains the middle 50% of the GM values,

while whiskers are lines that extends from the upper and lower edge of the box to the

highest and lowest values which are not greater than 1.5times the IQ range.

35

616

617618619

620

621

622

623

624

625

626

Figure 3D Box plots showing the relationship between GM index values and sputum

consistencies. Yellow line shows GMI of 6.5 while green dashed line shows GM of 1.0

The line across the box represents the “ median “ , the box indicates the interquantile

(IQ) range which contains the middle 50% of the GM values, while whiskers are lines

that extends from the upper and lower edge of the box to the highest and lowest values

which are not greater than 1.5times the IQ range.

36

627628

629630

631

632

633

634

635

636

Figure 3E Box plots showing the relationship between transformed Ct values and sputum

consistencies. Blue line shows transformed Ct (Tct) value of 2.0. (Ct- Cycle threshold).

The line across the box represents the “ median “ , the box indicates the interquantile

(IQ) range which contains the middle 50% of the GM values, while whiskers are lines

that extends from the upper and lower edge of the box to the highest and lowest values

which are not greater than 1.5times the IQ range.

37

637

638639640

641

642

643

644

645

646647