1

Improved standardisation of the Bio-Rad Platelia Aspergillus galactomannan (GM) antigen 2

sandwich enzyme immunoassay (EIA) using the DS2® (Dynex) ELISA processing system. 3

4

R L Gorton1#, P L White3, E Bagkeris4, D Cotterall1, R Desai1, T McHugh2, C C Kibbler1 5

6

1UK Clinical Mycology Network (UKCMN) Regional laboratory, Department of 7 Microbiology, Royal Free Hospital, Pond Street, Hampstead, London, NW3 2QG 8

2Centre for Clinical Microbiology, UCL, Royal Free Campus, London 9

3UKCMN Regional laboratory, Public Health Wales, Microbiology Cardiff. UHW, Heath 10 Park, Cardiff CF14 4XW. 11

4Population, Policy & Practice Program, UCL Institute of Child Health, 30 Guilford Street, 12 London WC1N 1EH, UK 13

14

Running title; standardisation of Bio-Rad Platelia Aspergillus EIA 15

16

Corresponding author Rebecca Gorton: rebecca.gorton@nhs.net 17

18

19

20

21

22

23

24

25

26

JCM Accepted Manuscript Posted Online 15 April 2015J. Clin. Microbiol. doi:10.1128/JCM.00157-15Copyright © 2015, American Society for Microbiology. All Rights Reserved.

Abstract 27

The Galactomannan EIA is widely utilized in the diagnosis of Invasive Aspergillosis (IA). 28

There is inconsistent reproducibility of results between centres when the assay is processed 29

manually. Automation of EIA assays can reduce variation. This study investigated the semi-30

automation of the GM-EIA on the DS2 (Dynex) platform in three stages; (i) DS2 GM-EIA 31

method validation with experimental samples (ii) DS2 re-testing of case defined clinical 32

samples and (iii) a 12 month audit of DS2 GM-EIA performance 33

In stage (i) Bland-Altman analysis demonstrated reduced variance between ODI values for 34

samples processed on two DS2 platforms (mean difference -0.02, LOA -0.19-0.14), compared 35

with the variance between samples processed manually and on a DS2 platform (mean 36

difference 0.02, limits of agreement (LOA) -0.25-0.3). In stage (ii) 100% (14/14) qualitative 37

agreement was observed for serum samples from cases of IA with no significant change in 38

the ODI values when processed on the DS2 platform. A significant decrease in the ODI 39

values was observed for control serum samples on the DS2 platform (difference: 0.01, 40

p=0.042). In Stage (iii) a significant reduction in the frequency of equivocal results from 41

5.56% (136/2443) to 1.56% (15/961) was observed after DS2 automation (Difference: 4.0%, 42

95% CI: 2.7, 5.2, p: <0.01) with an equivalent increased in negative results. 43

This study demonstrates that GM-EIA automation may reduce inter-site variability. 44

Automation does not impact on the repeatability of true positive results but contributes to a 45

reduction in false-positive (equivocal) GM-EIA results, reducing the need to re-test a 46

significant proportion of samples. 47

48

49

50

51

52

53

54

Introduction 55

The Bio-Rad Platelia Aspergillus galactomannan antigen sandwich enzyme immunoassay 56

(GM-EIA) is widely used as a screening method for prospective surveillance of invasive 57

aspergillosis (IA), in patients at high risk of disease. The assay is well established, reflected 58

by its recommendation in European Organisation for Research and Treatment of Cancer 59

(EORTC) consensus criteria for defining invasive fungal disease (5). Despite this the 60

diagnostic performance of the GM-EIA is variable, with meta-analyses showing combined 61

sensitivity raging between 0.71-0.78 and specificity ranging between 0.81-0.89 (11,14). The 62

false positivity experienced using the GM-EIA has been associated with antimicrobial 63

treatment (e.g. piperacillin-tazobactam), other invasive fungal diseases (e.g. Fusariosis) and 64

even after ingestion of an ice-pop (7,8,15,18). Consequently, the GM-EIA should not be used 65

as a stand-alone diagnostic test but is an important component of the the diagnostic strategy 66

for managing IA (5). 67

Although initially, the GM-EIA reproducibility was reported to be excellent between 68

laboratories (16), recent reports document a lack of reproducibility when repeat testing 69

positive samples (2,13). In particular, samples with an optical density index (ODI) on or 70

around the positivity threshold (≅0.5) of the assay regularly found to be negative on repeat 71

testing (10,16). The storage conditions of specimens appears to impact on reproducibility 72

with a significant decline in sample ODI reported after storage at -80 °C (10). IA diagnosis 73

also appears to be important, with non-reproducibility more frequently observed when 74

retesting false positive samples from patients without IA (6,10). A significant correlation 75

between serum albumin concentration and difference in ODI value on retesting has also 76

recently been reported. A larger reduction in indices was observed when retesting serum with 77

increasing albumin concentration (10). 78

Whilst storage and disease status have been shown to affect GM-EIA reproducibility other 79

factors such as human error, environmental contamination at the point of testing, and 80

variability in local testing conditions between laboratories may also impact on the assay 81

performance and reproducibility. The GM-EIA assay is performed manually by most 82

laboratories and consequently is susceptible fluctuations in the environmental temperature 83

and operator variability. Any measures taken to standardise the GM-EIA assay across 84

laboratories would be advantageous with the aim of minimising operator and environmental 85

influence. Automation of the GM-EIA on an ELISA processing system will assist in 86

standardising the assay. There are several ELISA processing systems available and although 87

unpublished the GM-EIA has been automated using the EVOLIS ELISA processing system 88

(Biorad, written communication). This study aimed to evaluate the automation of the GM-89

EIA on an alternative open platform, the DS2 (Dynex Technologies) ELISA processing 90

system to test the impact of the DS2 in reducing test variability. 91

Methods 92

Study design: the study was conducted in three stages (i) validating DS2 GM-EIA 93

methodology, comparing results from experimental samples processed on two DS2 platforms 94

against manual processing (ii) DS2 re-testing of positive/negative clinical samples based on 95

case definition of IA and (iii) an audit of DS2 GM-EIA performance over a 12 month period 96

of use in a routine diagnostic setting. 97

(i) Stage 1 - DS2 method validation 98

Prior to implementation in clinical practice the DS2 system was evaluated using control 99

serum samples and pooled serum spiked with increasing volumes (10, 20, 30, 40 and 50µl) of 100

galactomannan positive control (supplied by BIORAD). Samples were processed on the day 101

of spiking, manually and on two identical DS2 platforms to allow analysis of inter-platform 102

reliability. Throughout the validation stage we collected OD values for the GM-EIA assay 103

threshold controls The OD values were then used to define new threshold control 104

characteristics for DS2 processing with an acceptable OD range of >0.3 and <1.2 and a 105

maximum OD difference between the two threshold controls of 0.5 (data not presented). Both 106

optical density (OD) and optical density index (ODI) values were analysed in this stage of the 107

study. 108

(ii) Stage 2 – Re-analysis of samples defined by disease status 109

To assess clinical performance of the DS2 processing the ODI values obtained from manual 110

and DS2 processing for case and control patients sera were compared. 111

Positive sera; 112

Positive sera are not frequently encountered therefore, an electronic search of samples 113

processed through the year 2012 identified 17 positive serum samples (ODI value >0.5) for 114

inclusion. Serum samples had been stored at -80°C and were tested manually and on the DS2 115

platform within 24 hours of thawing. Samples were assigned to a case/control definition prior 116

re-testing on the DS2. Fourteen samples were obtained from six patients (range 1-5 samples) 117

with IA. Three samples were from control patients with no evidence of IA. 118

Negative sera; 119

Data were collected for serum samples tested prospectively manually and by DS2 over one 120

month of testing. Only data for samples for which the disease status of the patient was known 121

were included, in total 55 negative sera were included, 53 samples came from control patients 122

and two samples from patients with IA. 123

(iii) Stage 3 - post implementation clinical evaluation 124

The impact of DS2 automation was assessed through a twelve-month audit of qualitative 125

clinical results (positive, equivocal and negative interpreted from the ODI values) reported 126

pre DS2 implementation from 1st of January – 31st December 2009 and post DS2 127

implementation 1st of January – 31st December 2013. GM-EIA threshold control values were 128

also collected to evaluate the performance of the adjusted DS2 threshold control parameters 129

for DS2 processing. In the pre DS2 timeframe samples were tested on the day of receipt or 130

stored in the fridge until tested (Jan-March), then protocol changed to store samples in the 131

freezer at -20°C (April – Dec) prior to testing, when the test was not being performed on the 132

day of receipt. In the post DS2 implementation time frame samples were either tested on the 133

day of receipt or stored at -20°C until the day of testing. 134

GM-EIA processing; GM-EIA kits from the same LOT number were utilised during stage 1 135

with the same operator performing all analyses. For stage 2 and 3 of the study the operators 136

varied and kits from different batches were used. Throughout all stages serum was added to 137

treatment solution in a 3:1 ratio (either 360µl of serum and 120 µl of treatment solution or 138

300 µl serum and 100 µl of treatment solution, for automated and manual processing, 139

respectively). Sample pre-treatment was performed as per manufacturer’s instructions in a 140

Class 2 biological safety cabinet. Liquid handling processes were identical for manual and 141

DS2 processing and were performed as per manufacturer’s instructions. During stage (i) all 142

manually processed plates were washed manually (five washes) by the operator. During the 143

second stage of the study manual plates were washed using a bench top plate washer (5 144

stringent washes). The DS2 wash cycle program was as follows; a five cycle super sweep 145

wash strip wise with constant timing set to four dispense loops, a bottom wash with 250ul 146

and a final aspirate cycle with 380ul of wash buffer (full DS2 program is available to 147

download in S2 supplementary material). 148

GM-EIA results interpretation; the optical density (OD) for samples and controls was 149

measured at 450nm/620 nm. The validity of each run was determined as per manufacturer’s 150

instructions. For manually processed plates the threshold control had to be within the 151

prerequisite OD limits >0.3 and <0.8. For DS2 processing the upper range of acceptance for 152

threshold controls was increased to <1.2 with a control equation written into the DS2 153

program to ensure that the difference between the threshold control OD measurements was 154

no greater than 0.5, as allowed by the manufacturer. ODI values were calculated as per 155

manufacturer’s instructions; any ODI value above 0.5 was considered positive. 156

Statistical analysis; Data normality was tested using the Kolmogorov-Smirnov and Shapiro-157

Wilks tests. For stage (i) of the study Bland - Altman plots (1) were used to demonstrate 158

agreement between ODI results from manual processing and the two DS2 platforms. The 159

Bland–Altman analysis calculates the mean difference between ODI values from the 160

processing methods (the ‘bias’), and 95% limits of agreement as the mean difference around 161

the bias (2 SD). It is expected that the 95% limits include 95% of differences between the two 162

measurement methods. Spearman’s rho was used to test correlations and Kappa statistic was 163

calculated to test inter-assay qualitative agreement between all samples and for cases and 164

controls individually. In stage (ii) of the study agreement between results was determined in 165

both a qualitative manner (positive versus negative result) and quantitative manner 166

(comparison of median initial ODI versus retest ODI). Wilcoxon signed-rank test was used to 167

measure difference across ODI values. Differences were determined at the 0.05 level of 168

significance. In stage (iii) of the study frequencies for negative, equivocal and positive values 169

were calculated for a year pre DS2 processing and a year of DS2 use in the diagnostic setting. 170

Analysis of difference between the frequencies was conducted using Fisher’s exact test. 171

172

173

Results 174

Stage (i) - DS2 validation. 175

Threshold controls; Data collected for 10 threshold controls was used to define the 176

threshold control limits for DS2 testing, table 1. A median OD 0.48 (Inter quartile range 177

(IQR): 0.42, 0.53) was observed for GM-EIA plates processed manually. Median OD’s of 178

0.88 (IQR: 0.76, 1.0) and 0.79 (range 0.68, 1.1) were observed for GM-EIA plates processed 179

on the two DS2 platforms respectively, table 1. 180

181

Sample population; 121 samples were processed (20 GM spiked serum samples and 101 182

blank serum samples). All 20 GM spiked serum samples were positive (ODI ≥0.5) by manual 183

and DS2 GM-EIA processing and all 101 negative serum samples were confirmed as 184

negative on repeat testing by manual processing and on the two DS2 platforms. Observed 185

qualitative agreement was 100% (95% CI: 96.9-100) with a Kappa statistic of 1.0 between 186

the manual results and those obtained on the two DS2 platforms. 187

188

The characteristics of the OD and ODI values for the 121 samples measured manually and on 189

the DS2 platforms are presented in tables 1 and 2. To assess the relationship between OD and 190

ODI values linear regression was performed and Spearman’s rho correlation coefficients 191

calculated. A strong positive significant correlation was observed between the OD’s for 192

samples processed manually and on the first (r=0.67 p=<0.01) and second DS2 platforms 193

(r=0.63 p=<0.01), figure S1 (a) and (b). A very strong positive correlation of OD 194

measurements was observed between the two DS2 platforms (r=0.78 p=<0.01) figure S1 (c). 195

196

Figure 1 displays linear regression and Spearman’s correlation coefficients for ODI values 197

measured manually and on the two DS2 platforms. When applying Spearman’s rho to ODI 198

values generated by manual and DS2 processing again a strong positive significant 199

correlation was observed between manual processing and the first (r=0.72 p=<0.01) and 200

second DS2 (r=0.73 p=<0.01) platform, figures 1 (a) and (b). A very strong positive 201

correlation was observed between indices generated on the two DS2 platforms (r=0.84 202

p=<0.01) figure 1 (c). 203

204

Given that Spearman’s correlation measures the strength of a monotonic relationship between 205

the data rather than agreement Bland-Altman plots (1,3) were generated between manual 206

processing and DS2 processing. In Bland-Altman analysis the difference between ODI values 207

is plotted against the average ODI values for the two processing methods being compared, 208

figure 2 (a-c). 209

210

In figure 2 Bland-Altman plots demonstrate greater variation between ODI values measured 211

on the DS2 platforms and manual processing than the variation between ODI values 212

measured on the two DS2 platforms. The LOA (represented as two standard deviations 213

around the mean) was approximately halved for ODI values obtained from the two DS2 214

platforms (mean difference -0.02, LOA -0.19-0.14) compared with manual processing and 215

first DS2 platform (mean difference 0.02, limits of agreement (LOA) -0.25-0.3) or manual 216

processing and second DS2 platform (mean difference -0.00, LOA -0.27-0.26). The observed 217

differences for the three methods fell within two standard deviations of the mean for 94.2% 218

(114/121) of the samples between manual and the first DS2 platform, 95% (115/121) of the 219

samples between manual and the second DS2 platform and 93.4% (113/121) of the samples 220

between the two DS2 platforms. 221

222 Stage (ii) Re-analysis of samples defined by disease status 223 224 In total seventy-two samples were retested by DS2 which included 58 samples obtained from 225

42 control patients (range 1-3 samples) at high risk but with no evidence of fungal disease 226

and 14 samples (range 1-5 samples) were obtained from six patients defined as cases of 227

probable IA, with radiological evidence of sinusitis and/or pulmonary disease. 228

Of the 72 samples, 55 (53 from control patients and two from cases) were initially negative 229

by manual processing. Fifty two samples remained negative by DS2 processing (94.5% 230

observed agreement 95% CI: 85.2 98.1). Three samples were positive by DS2 processing on 231

repeat, all from control patients and all had an initial ODI value in the range of 0.4 - 0.49, 232

close to the positivity threshold of 0.5. Both of the negative samples from cases were in the 233

range of 0.4 - 0.49, when manually tested and on repeat testing by DS2. Seventeen samples 234

were initially positive when tested manually, 12 samples from cases of IA and 5 samples 235

from control patients. On repeat 94.1% (16/17) were positive when re-tested by the DS2 236

(95% CI: 73.0 99.0). Analysis by disease status of samples initially testing positive (ODI 237

>0.5) by manual processing demonstrated that for cases 100% (12/12) of the samples 238

remained positive on repeat by DS2 processing. For controls 80% (4/5) of positive samples 239

remained positive on repeat testing. One sample positive manually, with an ODI value of 240

0.51, was negative by DS2 repeat testing with an ODI value of 0.45. 241

242

Observed qualitative agreement between the manual processing and DS2 was 94.4% (68/72, 243

95% CI: 86.6, 97.8 with a Kappa statistic of 0.64. Observed agreement for controls was 244

93.1% (54/58, 95% CI: 83.6, 97.3) with a Kappa of 0.63. For cases 100% (14/14, 95% CI: 245

78.5, 100) observed agreement was achieved between the manual and DS2 processing, but it 246

was not possible to determine a Kappa statistic due to low sample numbers. The median GM 247

ODI (when tested manually and on the DS2 platform) for the overall population, cases and 248

controls are shown in table 3. On the DS2 platform the GM ODI for the overall population 249

was not significantly different (p=0.116) to the initial ODI obtained by manual testing. 250

Analysis by disease status shows that there was no significant difference between ODI values 251

when testing samples for cases by DS2. When testing samples from control patients with no 252

evidence of IFD a significant reduction in ODI values (p=0.042) was observed. Spearman’s 253

correlation coefficient between original and retest ODI overall was positive r=0.59, 254

representing a strong positive correlation between manual and DS2 testing (p: <0.001). 255

256

Stage (iii) - post implementation clinical evaluation 257

Threshold control analysis; A strong positive significant Spearman’s correlation of 0.89 258

p<0.01 was achieved between threshold control limit 1 and threshold control limit 2. The 259

median OD value was 0.72 (IQR: 0.62, 0.8) for threshold control limit 1 and 0.7 (IQR: 0.61, 260

0.79) for threshold control limit 2. In figure 3 the manufacturer’s recommended optical 261

density range for threshold controls (>0.3 and <0.8) are indicated by the dotted lines. The 262

optical density range defined for DS2 automation in this study was >0.3 and <1.2 as shown as 263

solid lines in figure 3. 264

Pre and post DS2 GM-EIA audit; Analysis of results reported form the GM-EIA over two 265

12 month time period’s pre and post DS2 automation showed that the number of positive 266

results decreased slightly from 4.01% (98/2443) in 2009 to 3.54% (34/961) in 2013 but this 267

decline was not statistically significant (Difference: 0.47%, 95% CI: -1.06, 1.78, p=0.555). 268

Equivocal reports (samples that tested positive initially and negative repeat) decreased almost 269

five fold from 5.56% (136/2443) in 2009 to 1.56% (15/961) in 2013 (Difference: 4.0%, 95% 270

CI: 2.7 5.2, p: <0.01). Negative reports increased from 90.42% (2209/2443) in 2009 to 94.9% 271

(912/961) in 2013 (95% CI: 2.6-6.2, p: <0.001). 272

273 Discussion 274 275 This study has demonstrated that the use of an automated liquid handling platform, such as 276

the DS2, can replace manual processing of the GM-EIA. Almost complete agreement (97.9% 277

189/193) was observed between qualitative results (negative/positive) across simulated and 278

clinical samples combined when processed manually and on the DS2 platform. Manual 279

processing is responsible for variation within the GM-EIA assay. In stage (i), where GM-EIA 280

was performed by a single operator, OD and ODI values measured manually were compared 281

against those measured on the two DS2 platforms. A stronger positive correlation was 282

observed between OD and ODI values measured on the two DS2 platforms. This indicates 283

that manual processing was responsible for the reduced correlation when comparing manual 284

and DS2 OD and ODI values. Bland-Altman plots demonstrated that bias across all three 285

comparisons was minimal (0.02). However; limits of agreement around the bias were almost 286

halved when comparing ODI values measured on the two DS2 platforms. This supports the 287

view that automation of the GM-EIA across laboratories will contribute to a reduction in 288

variance, an issue that has been reported in several other studies (2,12,16). This is further 289

supported by other studies reporting a decrease in variance within EIA assays after 290

automation (17,20). 291

When testing samples defined by disease status in stage (ii) of the study there was no 292

significant difference in ODI values for cases by DS2 (difference: 0 P: 0.082), but for 293

samples from patients with no evidence of IFD a significant reduction in ODI values was 294

observed when tested on the DS2 platform. Qualitative results (positive/negative) differed 295

between manual and DS2 processing for four samples, all from control patients (negative to 296

positive in three samples and positive to negative in one sample). Importantly all four 297

samples had ODI values between 0.4-0.49, close to the positivity threshold of the assay. As 298

these patients had no other evidence of invasive fungal disease these samples represent false 299

positives. The change in the qualitative results for these four samples is not likely to be a 300

result of DS2 automation but a result of the variable measured ODI seen in false positive 301

samples, which has been reported in a similar study (10). A lack of repeatability for samples 302

with ODI values close to the positivity threshold of the assay has also been reported by Marr 303

et al, 2006 when 10.2% of positive serum samples included in their study, all with an ODI 304

between 0.5-0.7, tested negative on repeat (16). 305

False positivity using the GM assay has been reported in several case reports, associated with 306

factors such as the use of Piperacillin-tazobactam, which is an antimicrobial frequently used 307

to treat neutropenic sepsis (4,7,15). Weaker non-specific binding of cross reacting epitopes in 308

false positive sample could cause increased variation in ODI values upon repeat testing and 309

explain the lack of repeatability observed for control samples. In a recent report Kimpton et al 310

suggested that molecules causing false positive results bind with less affinity to the EB-A2 311

antibody after freezing and storage which leads to them being negative on repeat (10). 312

Conversely, the target molecule Galactomannan binds with high affinity, which leads to 313

repeatability of results for true positive samples. The complete qualitative agreement seen for 314

results from positive serum samples spiked with control GM in stage (i) of the study and 315

positive samples form cases of IA in stage (ii) of the study supports this hypothesis. 316

Furthermore, two ‘negative’ samples from cases of IA with ODI values close to the positivity 317

threshold (0.45 & 0.46 by manual testing and 0.46 &0.47 by DS2 testing), were likely to 318

represent low level positive samples as measured by both processing methods. To control for 319

this the manufacturer’s recommendations for GM-EIA testing are that all positive GM 320

samples are retested. In our centre we adhere to these recommendations and account for 321

positive GM samples that test negative on repeat as being equivocal. 322

After implementation of the DS2 GM-EIA processing we observed a significant decrease in 323

equivocal results (initially positive samples testing negative on repeat) reported from our 324

laboratory from 5.56% to 1.56%, with an equivalent increase in negative results reported in 325

this time period. This suggests that automation contributed to a reduction in false positivity. 326

The reduction in equivocal results using the DS2 may be a result of increased stringency of 327

the washing program. Weakly bound cross reacting molecules may be washed away more 328

effectively. However, the observed reduction in equivocal results could also be due to an 329

effect of sample storage at -20°C prior to testing, which was introduced for all samples not 330

being tested on the day of receipt in our centre around the time of DS2 automation. Low 331

reproducibility of GM positive results after storage at -20°C has been highlighted in two 332

studies (9,13) however, the long term stability of galactomannan in serum at -20° C has been 333

demonstrated for up to 5 years (19). This suggests freezing may lead to a reduction in false 334

positivity but not affect genuinely positive samples. In this study there was no significant 335

difference in the frequency of positive samples (those testing positive twice) after DS2 336

automation and freezing measures. 337

Finally this is the first study to descibe a protocol for GM-EIA automation and highlights an 338

important technical issue. During the validation stage of this study the measured optical 339

density on both DS2 platforms was significantly higher across the entire GM-EIA plate for 340

samples and controls compared with manual processing (difference in median 0.058 341

P<0.01/difference 0.061 P<0.01). Using the optical density threshold control limits defined 342

by the manufaturer on the DS2 consequently resulted in a failure of the assay’s quality 343

control and so the limits for threshold controls required adjustment. The threshold control 344

limits (OD values of >0.3 and <0.8) control for intra-assay variance, by ensuring that the 345

difference in OD between the two threshold controls does not exceed 0.5. The manufacturer’s 346

recommended <0.8 OD upper limit was frequently exceeded during the validation stage of 347

this study, but crucially the difference between the OD of the two cut off controls was 348

negligible (range 0.002-0.088, data not presented). To enable the quality control of the GM-349

EIA to pass on the DS2 the threshold control upper limit was increased to <1.2 with an 350

additional measure written into the DS2 program to reject an OD difference >0.5 between the 351

two threshold controls. Increasing the threshold control upper limit did not impact on the 352

final qualitative results obtained by DS2 processing. The ODI value is calculated by division 353

of the sample OD by the mean OD value of the two threshold controls. As the OD increase 354

was proportional across samples and controls the effect was normalised by this calculation. 355

This is supported by the absolute agreement of results from samples processed manually and 356

by DS2. Threshold control data obtained from a 12-month audit of 92 GM-EIA runs also 357

demonstrated that the OD did not exceed 1.2, as demonstrated in figure 2, and that the largest 358

difference between the two threshold controls was again negligible (range 0.002-0.112). There 359

is strong evidence from this study that the allowed OD variation of 0.5 between threshold 360

controls could be significantly reduced. 361

362

In conclusion this study has demonstrated that the GM-EIA assay can be automated on an 363

ELISA processing system, such as the DS2, without impacting on the assay performance and 364

will contribute to a reduction in variance of results using the assay. Automation of EIA assays 365

leads to increased standardisation, time saving benefits for the laboratory and allows for more 366

robust comparison of results across centres that employ automated platforms for ELISA 367

processing. 368

369

370

371

Acknowledgements 372

This work was supported by the National Institute for Health Research (NIHR) as part of the 373

Chief Scientific Officers NIHR Ph.D. funded Fellowship scheme. 374

C. C. Kibbler has received honoraria from Astellas, Gilead, MSD and Pfizer. 375

P.L.W. is a founding member of the EAPCRI, received project funding from Myconostica, 376

Luminex, and Renishaw Diagnostics, was sponsored by Myconostica, MSD, and Gilead 377

Sciences to attend international meetings, and was a consultant for Renishaw Diagnostics 378

Limited 379

380 Reference List 381

382

1. Altman D G and Bland J M. Measurement in medicine; the analysis of method comparison 383 studies. 32, 307-317. 1983. Statistician. 384

385 2. Bizzini, A., O. Marchetti, and P. Meylan. 2012. Response to: lack of intra-laboratory 386

reproducibility in using Platelia Aspergillus enzyme immunoassay test for detection of 387 Aspergillus galactomannan antigen. Transpl.Infect.Dis. 14:218-219. 388

3. Bland, J. M. and D. G. Altman. 1986. Statistical methods for assessing agreement between two 389 methods of clinical measurement. Lancet 1:307-310. 390

4. Blijlevens, N. M., J. P. Donnelly, J. F. Meis, P. E. Verweij, and B. E. de Pauw. 2002. Aspergillus 391 galactomannan antigen levels in allogeneic haematopoietic stem cell transplant recipients 392 given total parenteral nutrition. Transpl.Infect.Dis. 4:64-65.. 393

5. De, P. B., T. J. Walsh, J. P. Donnelly, D. A. Stevens, J. E. Edwards, T. Calandra, P. G. Pappas, J. 394 Maertens, O. Lortholary, C. A. Kauffman, D. W. Denning, T. F. Patterson, G. Maschmeyer, J. 395 Bille, W. E. Dismukes, R. Herbrecht, W. W. Hope, C. C. Kibbler, B. J. Kullberg, K. A. Marr, P. 396 Munoz, F. C. Odds, J. R. Perfect, A. Restrepo, M. Ruhnke, B. H. Segal, J. D. Sobel, T. C. Sorrell, 397 C. Viscoli, J. R. Wingard, T. Zaoutis, and J. E. Bennett. 2008. Revised definitions of invasive 398 fungal disease from the European Organization for Research and Treatment of Cancer/Invasive 399 Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious 400 Diseases Mycoses Study Group (EORTC/MSG) Consensus Group. Clin.Infect.Dis. 46:1813-1821. 401

6. Furfaro, E., M. Mikulska, F. Miletich, and C. Viscoli. 2012. Galactomannan: testing the same 402 sample twice? Transpl.Infect.Dis. 14:E38-E39. 403

7. Gerlinger, M. P., P. Rousselot, S. Rigaudeau, C. Billon, S. Touratier, S. Castaigne, and O. Eloy. 404 2012. False positive galactomannan Platelia due to piperacillin-tazobactam. Med.Mal Infect. 405 42:10-14. 406

8. Guigue, N., J. Menotti, and P. Ribaud. 2013. False positive galactomannan test after ice-pop 407 ingestion. N.Engl.J.Med. 369:97-98.. 408

9. Johnson, G. L., S. J. Sarker, K. Hill, D. A. Tsitsikas, A. Morin, S. A. Bustin, and S. G. Agrawal. 409 2013. Significant Decline in Galactomannan Signal during Storage of Clinical Serum Samples. 410 Int.J.Mol.Sci. 14:12970-12977. 411

10. Kimpton, G., W. P. Lewis, and R. A. Barnes. 2014. The effect of sample storage on the 412 performance and reproducibility of the galactomannan EIA test. Med.Mycol. 413

11. Leeflang, M. M., Y. J. Debets-Ossenkopp, C. E. Visser, R. J. Scholten, L. Hooft, H. A. Bijlmer, J. 414 B. Reitsma, P. M. Bossuyt, and C. M. Vandenbroucke-Grauls. 2008. Galactomannan detection 415 for invasive aspergillosis in immunocompromized patients. 416 Cochrane.Database.Syst.Rev.CD007394. 417

12. Oren, I., I. Avidor, and H. Sprecher. 2012. Lack of intra-laboratory reproducibility in using 418 Platelia Aspergillus enzyme immunoassay test for detection of Aspergillus galactomannan 419 antigen. Transpl.Infect.Dis. 14:107-109. 420

13. Pedroza, K. C., S. B. de Matos, D. L. de Moura, M. B. Oliveira, M. A. Araujo, R. J. Nascimento, 421 and F. W. Lima. 2013. Reproducibility of positive results for the detection of serum 422 galactomannan by Platelia aspergillus EIA. Mycopathologia 176:295-297. 423

14. Pfeiffer, C. D., J. P. Fine, and N. Safdar. 2006. Diagnosis of invasive aspergillosis using a 424 galactomannan assay: a meta-analysis. Clin.Infect.Dis. 42:1417-1427. 425

15. Tortorano, A. M., M. C. Esposto, A. Prigitano, A. Grancini, C. Ossi, C. Cavanna, and G. L. 426 Cascio. 2012. Cross-reactivity of Fusarium spp. in the Aspergillus Galactomannan enzyme-427 linked immunosorbent assay. J.Clin.Microbiol. 50:1051-1053. . 428

16. Upton, A., A. Gugel, W. Leisenring, A. Limaye, B. Alexander, R. Hayden, and K. A. Marr. 2005. 429 Reproducibility of low galactomannan enzyme immunoassay index values tested in multiple 430 laboratories. J.Clin.Microbiol. 43:4796-4800. 431

17. Van Campenhout, C. M., K. A. Van Cotthem, W. J. Stevens, and L. S. De Clerck. 2007. 432 Performance of automated measurement of antibodies to cyclic citrullinated peptide in the 433 routine clinical laboratory. Scand.J.Clin.Lab Invest 67:859-867. 434

18. Vergidis, P., R. C. Walker, D. R. Kaul, C. A. Kauffman, A. G. Freifeld, D. C. Slagle, A. K. Kressel, 435 and L. J. Wheat. 2012. False-positive Aspergillus galactomannan assay in solid organ transplant 436 recipients with histoplasmosis. Transpl.Infect.Dis. 14:213-217. 437

19. Wheat, L. J., M. H. Nguyen, B. D. Alexander, D. Denning, A. M. Caliendo, G. M. Lyon, L. R. 438 Baden, F. M. Marty, C. Clancy, E. Kirsch, P. Noth, J. Witt, M. Sugrue, and J. R. Wingard. 2014. 439 Long-term stability at -20 degrees c of Aspergillus galactomannan in serum and 440 bronchoalveolar lavage specimens. J.Clin.Microbiol. 52:2108-2111. 441

20. Whitworth, W. C., D. J. Goodwin, L. Racster, K. B. West, S. O. Chuke, L. J. Daniels, B. H. 442 Campbell, J. Bohanon, A. T. Jaffar, W. Drane, P. A. Sjoberg, and G. H. Mazurek. 2014. 443 Variability of the QuantiFERON(R)-TB gold in-tube test using automated and manual methods. 444 PLoS.One. 9:e86721. 445

446 447

448

Table 1. Galactomannan EIA OD characteristics when testing serum samples and threshold controls by manual, DS1 and DS2 processing

*Does not include threshold controls

Table 2. Comparison of Galactomannan EIA ODI characteristics when testing serum samples by manual, DS1 and DS2 processing Sample population

Overall* (n=121) Negative (n=101) Spiked Positive (n=20) Threshold controls (n=10)

Parameter Manual DS1 DS2 Manual DS1 DS2 Manual DS1 DS2 Manual DS1 DS2

Range 0.023-1.44 0.055-1.876 0.04-2.693 0.023-0.149 0.055-0.299 0.04-0.329 0.227-1.440 0.486-1.876 0.313-2.693 0.378-0.687 0.742-1.106 0.603-1.135

25th percentile 0.04 0.076 0.079 0.037 0.073 0.075 0.442 0.807 0.688 0.42 0.88 0.68

Median OD 0.048 0.106 0.109 0.045 0.093 0.102 0.627 1.058 1.141 0.48 0.76 0.79

75th percentile 0.078 0.151 0.159 0.057 0.127 0.127 0.912 1.341 1.580 0.53 1.0 1.1

Overall (n=121) Negative (n=101) Spiked Positive (n=20) Parameter Manual DS2A DS2B Manual DS2A DS2B Manual DS2A DS2B Range 0.048-3.282 0.057-2.374 0.056-2.380 0.048-0.339 0.057-0.32 0.056-0.416 0.50-3.282 0.51-2.374 0.501-2.380 25th percentile 0.085 0.09 0.097 0.078 0.088 0.093 0.894 0.887 0.868 Median index 0.101 0.115 0.132 0.093 0.105 0.12 1.337 1.209 1.317 75th percentile 0.160 0.167 0.177 0.118 0.137 0.151 0.83 1.626 1.79

Table 3. Comparison of Galactomannan EIA ODI reproducibility when testing case defined clinical serum samples manually and on the DS2 platform.

Population

Parameter Overall (n=72) Case (n=14) Control (n=58) Initial ODI Retest ODI Initial ODI Retest ODI Initial ODI Retest ODIRange 0.05-1.59 0.04-1.71 0.45-1.59 0.46-1.71 0.05-0.51 0.04-0.56 25th percentile 0.08 0.06 0.5 0.57 0.07 0.05 Median 0.11 0.10 0.85 0.88 0.10 0.09 75th percentile 0.48 0.5 1.1 1.15 0.21 0.11 Wilcoxson P value 0.116 0.972 0.042 Spearman’s rho Correlation coeff. P value

0.593 <0.01