· web viewrevised . comparative performance of . aspergillus . galactomannan. elisa. and. pcr...
TRANSCRIPT
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, [email protected]
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