Can rapid integrated Polymerase Chain Reaction (PCR)-based diagnostics for Gastrointestinal pathogens improve routine hospital infection control practice? A diagnostic study

Participants

Samples were collected according to the MassCode standard operating procedures (SOPs) (see Appendix 1). In brief, all samples collected were initially sent for faecal culture and/or C. difficile toxin testing at the Oxford University Hospitals microbiology laboratory by hospital-based doctors or general practitioners (GPs) as a result of a suspected enteric infection. For negative samples, no pathogens were found using standard microbiological workflows and the samples were sent for discard. For known positive samples, one or more pathogens were found and the samples were sent for discard once testing was complete. The service microbiology laboratory in the Oxford University Hospitals uses an enzyme immunoassay (EIA) to identify C. difficile, which is well recognised to have suboptimal sensitivity and specificity. Separately to the MassCode study, all EIA-positive C. difficile samples were cultured in the parallel research laboratory, and only those positive for C. difficile on both EIA and culture were used as reference positives in this study. At this point, both sample groups were completely anonymised and collected for this investigation. As norovirus testing was not carried out by the microbiology laboratory unless an outbreak was suspected, positive norovirus samples were obtained through a separate investigation conducted by a National Institute for Health Research (NIHR) clinical fellow. These samples had been confirmed positive by PCR13,14 and were anonymised prior to collection. All samples were stored at 4 °C. For limit of detection testing, negative samples containing adequate quantities of faecal matter were selected at random and homogenised.

Reference standard

Initial diagnosis of faecal pathogens was performed according to Health Protection Agency (now Public Health England) guidelines. These are the current gold standard methods for investigating faecal specimens for enteric pathogens in the UK. Readers of the reference standard tests were qualified laboratory staff. Throughout all reference standard and index tests normal aseptic microbiological laboratory working practices were followed.

Limits of detection

Limit of detection experiments were designed and implemented for the four main bacterial targets of interest in the MassCode panel: C. difficile, S. enterica, C. jejuni and C. coli. As norovirus cannot be cultured it was not included in these experiments. National Collection of Type Cultures (NCTC) strains of each of the bacterial organisms were obtained. These were C. difficile (NCTC 13307), C. jejuni (NCTC 11168), C. coli (NCTC 11353) and S. enterica serovar Enteritidis (NCTC 13349). Each organism was cultured on Columbia blood agar (E&O Laboratories Ltd, Bonnybridge, UK). C. difficile was incubated for 48 hours at 37 °C under anaerobic conditions; S. enterica for 24 hours at 37 °C under aerobic conditions; and C. jejuni and C. coli for 24 hours at 42 °C under microaerophilic conditions. Resulting plates of growth were scraped into 200 µl of nutrient broth with 10% glycerol (E&O Laboratories Ltd). Serial 1 : 10 dilutions were made from the neat sample to 10^–15. Aliquots of 50 µl of each serial dilution were spread onto blood agar plates and incubated under the conditions previously described for each organism. Colonies were enumerated and the cultivable range selected for subsequent limit of detection experiments. Culture aliquots were stored at –80 °C prior to use.

Negative stool samples from five patients were initially chosen for use in the limit of detection testing across the different pathogens. Aliquots of 200 µl or 0.2 mg of faecal matter were placed into clean 2-ml vials and spiked with a known volume of culture stock from the previously described serial dilutions. In accordance with the MassCode SOPs (see Appendix 1), 8 µg of yeast transfer RNA (tRNA; Invitrogen, Carlsbad, CA, USA) carrier RNA and 10⁵ MS2 bacteriophage internal control (IC) were also added at this stage. Following preparation of the sample, nucleic acids were extracted and purified with the QIAamp DNA Stool Mini Kit (Qiagen, Limburg, the Netherlands) following the manufacturers protocol. This kit was chosen because of its specific design for processing faecal samples, and ability to isolate both DNA and RNA from bacteria and viruses, since in clinical practice whether a stool sample will contain a bacteria or a virus is unknown. Thus, processing workflows have to be sufficiently robust and sensitive for both types of pathogen.

Eluted DNA (10 µl) was reverse transcribed using MassCode Single Strand complementary DNA (cDNA) Synthesis Kits (Agilent, Santa Clara, CA) to enable detection of the MS2 IC; as well as to ensure any other RNAs were converted to cDNA in preparation for the method being used to detect pathogens such as norovirus in the future. All samples were stored at –20 °C prior to testing. Control samples included the patient stool samples without spiked culture for all organisms, and samples negative for contaminating human and other DNA (molecular-grade water).

MassCode sample processing and analysis was performed according to the SOPs described (see Appendix 1). A 14-plex primer mix targeting MS2, C. difficile, S. enterica, C. jejuni, C. coli, norovirus (two targets), E. coli O157 (two targets), Shigella spp., S. Typhi/Paratyphi, Giardia spp., Cryptosporidium spp. and rotavirus was used. Quantitative PCR (qPCR) assays using the same primer sequences as the MassCode panel, with the inclusion of a dual-labelled probe (Table 1), were used to verify unexpected results.

Quantitative PCRs were performed with Brilliant Multiplex MasterMix (Agilent). Each target organism was tested for in a duplex reaction with MS2. In addition, C. jejuni and C. coli could be tested for in triplex with MS2; the two norovirus gene targets could also be tested for in triplex with MS2. In-house optimisation for each qPCR assay was performed and the optimised primer and probe concentrations were used for each reaction. All qPCR runs included a 1 : 10 standard curve from 10⁸ copies of the target sequence to one copy, plus a positive control for MS2 (10⁴ copies of target sequence) and no template controls.

Product (2 µl) from the cDNA reaction was amplified in a reaction mixture consisting of target primers (forward and reverse primers), target probes, 12.5 µl of Multiplex Mastermix and nuclease-free water to a final volume of 25 µl. Amplification and reading was carried out with the Stratagene MX3005P^® (Agilent) under the following conditions:

95 °C for 10 minutes95 °C for 15 seconds60 °C for 1 minute}× 40

Following qPCR analysis, copy numbers for each culture serial dilution were generated and averaged across the five faecal samples. For MassCode, the lower limit of detection was defined as the sample CFU concentration at which more than half of the spiked stool samples (i.e. ≥ 3/5) could be detected, although the concentration at which all five of the spiked samples could be detected was also estimated. For qPCR, the lower limit of detection was defined as the sample CFU concentration at which copy numbers could be reliably calculated according to the standard curve included on each assay.

Optimisation of primer panel

Concurrently with limit of detection testing, the MassCode primer panel was tested against a set of anonymised clinical samples, collected as described above, retrieved essentially at random from laboratory discards. The laboratory worker was unblinded to the expected pathogens present in each sample to allow parallel processing of the samples by MassCode and the appropriate single qPCR assay(s). Samples were extracted with the QIAamp DNA Stool Mini Kit and reverse transcribed according to the MassCode SOP; samples were then stored at –20 °C prior to analysis. A total of 57 samples were S. enterica positive, 122 C. difficile positive, 119 Campylobacter spp. positive and 109 norovirus positive; 127 samples were negative.

MassCode testing was performed with the 14-plex primer mix. Samples that were unexpectedly positive or negative for target organisms were retested by qPCR. Where dual-labelled probe assays were not available, samples were tested with SYBR^® Green (Life Technologies, Carlsbad, CA, USA) qPCR under the following conditions: Product (2 µl) from the cDNA reaction was amplified in a reaction mixture consisting of target primers (forward and reverse primers), 10 µl SYBR Green MasterMix, and nuclease-free water to a final volume of 20 µl. Amplification and reading was carried out with the Stratagene MX3005 under the following conditions:

95 °C for 10 minutes95 °C for 15 seconds60 °C for 1 minute}× 40

Dissociation curve 95 °C for 1 minute followed by 55 °C (for 30 seconds) ramping to 95 °C (for 30 seconds).

The sensitivity of the MassCode assay was determined through calculation of the percentage of positive samples found by MassCode compared with the expected number of positives as previously determined by conventional laboratory methods. Specificity was determined by calculating the percentage of genuine negatives that were called as negative by MassCode.

A subinvestigation was also carried out to explore other options for S. enterica targets. The primer sets and conditions used were as published by Liu et al. ,9 targeting invA, and Malorny et al. ,20 targeting ttrRSBCA. Twenty S. enterica-positive samples, collected according to the protocol above, were tested by the Liu et al. 9 primer set; and 64 by the Malorny et al. 20 primer set.

Extraction optimisation

Following initial limit of detection and primer panel optimisation testing as performed above, a series of experiments designed to optimise nucleic acid extraction and purification were performed. The organisms chosen for optimisation experiments were C. difficile and S. enterica.

A homogenised negative stool sample, collected as described above, was used in 200-µl aliquots. Stool aliquots were spiked with 1 : 10 serial dilutions of culture above, at and below the limit of detection, as previously determined. The prevalent methods for nucleic acid purification were explored with a variety of pre-steps to facilitate lysis. Chemical lysis was carried out using Qiagen’s QIAamp DNA Stool Kit buffer ATL, and Stool Transport and Recovery (STAR) buffer plus chloroform (Roche, Basel, Switzerland). Physical lysis was carried out through heating, freeze–thaw, bead beating and sonication. Purification focused on the three major techniques of silica membrane spin column (QIAamp DNA Stool Kit), magnetic bead (MagMax, Life Technologies) and precipitation-based purification (Masterpure, Epicentre, Madison, WI, USA). In addition, two automated extraction platforms were tested: the magnetic bead-based QIAsymphony (Qiagen) and silica membrane Corbett (QIAxtractor) systems (Qiagen).

All samples were reverse transcribed according to the MassCode SOP post extraction and stored at –20 °C before qPCR analysis as described above (see Table 1). Control samples included stool samples without spiked culture for all extraction methods and negative samples (molecular-grade water) for all extraction methods. Quantitative PCRs were performed with the assays and controls described in Limits of detection.

Method validation

After optimisation of the primer panel and extraction methods, the MassCode assay was challenged against a batch of known positive clinical samples. All samples were reverse transcribed according to the MassCode SOP post extraction and stored at −20 °C before qPCR analysis as described above (see Table 1). A total of 20 samples positive for Campylobacter spp., 20 positive for S. enterica, 20 positive for C. difficile and 40 norovirus-positive samples were processed by the MassCode assay and by qPCR, where sufficient material remained.

Extraction efficiency

Further experiments investigating the extraction efficiency of target organisms alongside other Enterobacteriaceae were performed. American Type Culture Collection (ATCC) specimens of E. coli (ATCC 700928) and Klebsiella pneumoniae (ATCC 700721) were obtained. Both were plated on blood agar and incubated for 24 hours at 37 °C. Serial dilutions were made and enumerated as described previously. Known volume spikes of serially diluted culture were extracted following the optimised protocol, with K. pneumoniae also spiked into homogenised negative stool as described above. Culture spiking extraction experiments were also repeated with C. difficile, S. enterica and C. jejuni using the optimised extraction protocol.

All samples were reverse transcribed according to the MassCode SOP post extraction and stored at −20 °C before qPCR analysis as described above (see Table 1).

Limits of detection

Limit of detection testing was performed between December 2011 and March 2012. The negative stool samples used were collected in October 2011 immediately after reference standard testing from which no pathogens were found. Samples were completely anonymised, no patient information was obtained.

The results for limit of detection testing are shown in Table 2. MassCode limits of detection for spiked stool were determined to be 2.5 × 10³ CFU/µl for C. difficile, 2.5 × 10⁴ CFU/µl for S. enterica, 2.5 × 10³ CFU/µl for C. jejuni and 1.2 × 10² CFU/µl for C. coli. Quantitative PCR limits of detection for spiked stool were one order of magnitude lower for all organisms other than C. jejuni, where limits of detection were two orders of magnitude lower with qPCR. Fewer target copies were found and limits of detection were typically higher for pure culture than for spiked stools. Some evidence of contamination was observed within the S. enterica qPCR results.

Optimisation of primer panel

Processing and testing of known positive anonymised clinical samples was carried out between August 2011 and May 2012. The stool samples used were collected between November 2008 and October 2011. Collection parameters were as described earlier; no patient information was obtained.

Table 3 provides a summary of the results from initial clinical sample testing in August 2011. The percentage positive calls (sensitivity) varied between MassCode and qPCR. For MassCode, sensitivities varied between 40% for S. enterica and 83% for Campylobacter spp. For qPCR, sensitivities varied between 72% for S. enterica and 97% for C. difficile.

Table 4 breaks down these results by copy number as determined by qPCR. For C. difficile, 95% (62/65) of the samples undetected by MassCode had copy numbers below 100, but only 6% (4) were below detection limits (BDLs) for qPCR. For Campylobacter spp. and norovirus, all samples undetected by MassCode had copy numbers below 100, while 85% (17/20) and 90% (38/42) of samples, respectively, had copy numbers below 10; however, 25% (5) and 26% (11) were also not detected by qPCR, respectively. For S. enterica, 91% (31/34) of samples undetected by MassCode had copy numbers below 100 and 47% were also BDLs for qPCR.

Testing clinical samples also led to some unexpected positives, particularly among additional targets not included in phase 1 testing (Table 5). These were retested by qPCR where possible (if sufficient sample remained). The organisms most commonly found were Giardia spp., S. Typhi, and norovirus, with 61, 20, and 27 unexpected positives, respectively, (including indeterminate results) and most of these positives were found to be negative by qPCR testing, with 55, 20, and 22 of the samples confirmed as false positives, respectively.

A number of samples were also found to be positive by MassCode and qPCR, but not by the reference standard tests. This included 11 samples positive for Campylobacter spp., five positive for norovirus and six positive for Giardia spp.

During the course of these experiments, it was also observed that the ‘norovirus 1’ (see Table 1) primer set underperformed compared with the ‘norovirus 2’ primer set. Norovirus 1 amplification was achieved if norovirus 2 was also positive, while norovirus 2 amplification could also occur alone.

Testing of S. enterica-positive samples by alternative primer sets revealed little difference between detection using the MassCode primer set and the published alternatives. In particular, the alternative invA assay by Liu et al. 9 produced near-identical results to those obtained by qPCR with MassCode primers. The Malorny et al. 20 2ttrRSBCA primer set detected 43 out of 64 samples, whereas the MassCode primer set detected 46 of the same sample batch. Of the 18 samples not detected by MassCode, all were also negative by the Malorny et al. 20 assay.

Extraction optimisation

Following limit of detection testing and primer panel optimisation, experiments were performed between May 2012 and January 2013 to optimise the isolation of nucleic acids from faecal samples. The organisms chosen for investigation were C. difficile and S. enterica. Anonymised negative stool samples were collected between October 2011 and December 2012, immediately after reference standard testing from which no pathogens were found. No patient information was obtained.

Recovery (copy number compared with CFU spiked into the stool sample) for C. difficile increased from mean 5% following the QIAamp DNA Stool Kit with standard protocol to mean 689% using the QIAsymphony with STAR buffer and bead beating pre-steps (Table 6). (Bead beating plausibly increases yields to > 100% because it liberates DNA from spores.) Extraction efficiencies saw less improvement for S. enterica across all methods tested. However, some improvement in copy number yield was found, increasing from mean 0.2% using the QIAamp DNA Stool kit with standard protocol to mean 1% with the QIAsymphony plus STAR buffer and bead beating pre-steps. Other protocols also saw increases in yield to levels up to and above 1%, but these were less consistent across the CFU concentrations of S. enterica used.

Method validation

Following optimisation of extraction methods, a small batch of anonymised positive clinical samples was used to validate the new extraction method. These samples were collected between July 2009 and June 2012, immediately after reference standard testing. Samples were processed and tested during June and July 2012. No patient details were collected.

The percentage of positive samples (sensitivity) determined by MassCode increased for Campylobacter spp. and C. difficile (Table 7). In particular, C. difficile sensitivity increased from 47% to 95%. The MassCode sensitivity for S. enterica was slightly lower. A second batch of norovirus samples was processed as a result of concerns regarding the validity of the positive reference standard test on the initial batch. Although it was not possible to process the second norovirus batch by MassCode, all samples were positive by qPCR.

Copy number comparisons between the QIAamp DNA Stool Kit and optimised extraction method were performed, where possible (Table 8). Mean copy number change was positive for each organism.

Extraction efficiency

Extraction efficiencies of a panel of bacteria (C. difficile, C. jejuni, S. enterica, E. coli and K. pneumoniae) were evaluated through repeating limit of detection testing using 1 : 10 serial dilutions of culture and the optimised extraction method. Anonymised negative stool samples were collected between October 2011 and December 2012, immediately after reference standard testing from which no pathogens were found. No patient information was obtained.

All Gram-negative Enterobacteriaceae (S. enterica, E. coli and K. pneumoniae) had mean extraction efficiencies below 1%, whereas C. jejuni had an average extraction efficiency of 5% (Table 9). The Gram-positive bacterium, C. difficile, had the highest extraction efficiency at 625% on average.

Limits of detection

Limit of detection testing was performed for all four cultivable organisms targeted as part of the phase 1 testing. Limits of detection were one to two orders of magnitude higher for the MassCode assay than for qPCR. This finding was anticipated as amplification efficiency of individual targets decreases when performed as part of a large multiplex reaction. Nevertheless, the limits of detection observed were higher than expected.

Compared with other published data of limits of detection for enteric multiplex reactions, the limits found here are varied. Liu et al. 9 developed an in-house Luminex panel and presented limit of detection in CFU/g. Assuming a stool sample input of 0.2 g, the lowest concentration at which all five samples were MassCode positive were, therefore, 1.3 × 10⁴ CFU/g, 1.3 × 10⁶ CFU/g, 1.3 × 10⁵ CFU/g and 6 × 10² CFU/g for C. difficile, S. enterica, C. jejuni and C. coli, respectively. This is compared with 10³ CFU/g for Salmonella spp. and C. jejuni, and 10⁵ CFU/g for C. coli as reported previously. 9 However, the reported investigation used a different commercial DNA extraction kit than was used here.

In order to be detected by MassCode, S. enterica had to be spiked into stool at a CFU level one to two orders of magnitude higher than required for the other organisms tested. For S. enterica, qPCR copy numbers were also three orders of magnitude lower than the CFU of the spike, whereas for the majority of other samples copy numbers were one order of magnitude lower than the CFU of the spike. These findings suggested there may be a specific problem with S. enterica nucleic acid isolation and detection. Pathogen load has been reported at 10³ to 10⁹ CFU/g of stool, suggesting a single order of magnitude improvement in limits of detection for MassCode would detect most bacterial pathogens. 9,10

Other than for C. coli, for which an average copy number of 60 led to all five samples being detected by MassCode, all five samples were detected by MassCode when copy numbers averaged over 100. Therefore, copy number yield needed to be over 100 in order to reliably detect clinically positive samples. Overall, the data strongly indicated that, although limits of detection were insufficient with the currently implemented protocol, an increase in nucleic acid yield would significantly improve limits of detection for MassCode.

Optimisation of primer panel

Concurrently with limit of detection testing, a collection of 407 known positive clinical samples was processed by QIAamp DNA Stool Kit, MassCode 14-plex panel and the appropriate qPCR assay.

Three out of four targets fell below the 75% sensitivity required for the MassCode study to progress through phase 1 to phase 2. Validated duplex/triplex qPCR assays using the same primer pairs as were used in the MassCode multiplex mix (see Table 1) were also more sensitive than the MassCode multiplex, with only S. enterica falling below 75% sensitivity. Investigating how yield of nucleic acids was affecting the MassCode assay sensitivity showed that the majority of samples undetected by MassCode fell below the limits of detection for this assay (> 100 copies).

The exception to this was S. enterica: in 47% of samples that tested negative for S. enterica by MassCode, S. enterica was also below the limits of detection of qPCR. To confirm that the target sequence for S. enterica was not limiting its detection, some S. enterica samples were retested by qPCR with alternative primer sets. Use of these alternative qPCR assays did not lead to any additional detection of S. enterica or a decrease in qPCR cycle threshold, suggesting no additional target copies were detected. These data supported limit of detection testing, suggesting that yield of nucleic acids was the main limitation to MassCode sensitivity.

Although only C. difficile-, S. enterica-, Campylobacter spp-. and norovirus-positive samples were obtained for this pre-phase 1 investigation, use of the 14-plex MassCode primer panel yielded some unexpected positives (Table 5). Retesting these positive samples by qPCR suggested that some of the primer choices for the 14-plex panel may be resulting in false positives; in particular for norovirus, S. Typhi and Giardia spp. A proportion of unexpected positive samples were found to be genuinely positive by qPCR, illustrating how molecular methods may detect positive samples missed by reference standard tests. However, as shown previously, known positive samples in which nucleic acid yield was low were not detected by MassCode.

These data, along with the observation that the norovirus 1 primer set did not contribute to any additional diagnosis of norovirus, led to a recommendation that the norovirus 1 primer set should be removed from the primer panel in an effort to reduce false positives, creating a 13-plex primer panel. As only C. difficile, S. enterica, Campylobacter spp. and norovirus were directly under investigation, no recommendations were made regarding the additional primer sets in the panel.

The two main outcomes from the processing of the known positive sample collection were a recommended adjustment to the MassCode primer panel and the identification of nucleic acid yield as a major limitation to the sensitivity of MassCode. In order to improve yield of nucleic acids, a comprehensive investigation into nucleic acid isolation methods was instigated.

Extraction optimisation

In the light of previous results with a commercial extraction kit, a comparison between the dominant methods for nucleic acid extraction was performed to ensure the method was optimal for the MassCode assay. The results showed that through addition of extra lysis steps to standardised extraction protocols, nucleic acid yield could be dramatically increased (see Appendix 1, Table 48).

The most successful lysis method was bead beating. The addition of bead beating to the QIAamp DNA Stool Kit standard protocol resulted in C. difficile yields increasing from 5% to 191% and S. enterica yields increasing from 0.2% to 0.9% (see Table 6). Recent studies have also employed the QIAamp protocol modified to include additional lysis steps,10 implying that the increased nucleic acid yield has independently been found to improve detection assay sensitivity. The InhibitEX tablet step of the QIAamp protocol was also removed to test whether or not its use was resulting in DNA loss for S. enterica. The results suggested that use of the InhibitEX tablet did not improve nucleic acid yield for this organism. However, standard deviations (SDs) were higher without the InhibitEX tablet, suggesting that its use did improve the consistency of yields.

The comparison of a Gram-positive organism (C. difficile) and a Gram-negative organism (S. enterica) illustrated how the physiology of organisms affected what was the most efficient lysis method. However, this also varied across nucleic acid purification method. For example, coupled with QIAamp purification, sonication was one of the most efficient lysis methods for C. difficile. However, for S. enterica sonication was only the most efficient lysis method when coupled with purification by MagMax, which included bead beating as part of the standard protocol.

For both organisms, the MasterPure purification method failed to consistently yield any nucleic acids. This is because the method was not designed for stool extractions, although efforts were made to adapt the method for this purpose.

A comparison was also performed using QIAsymphony, an automated magnetic bead-based extraction system. STAR buffer, which includes chloroform compared with ASL buffer, and the standard lysis buffer included in the QIAamp DNA Stool Kit were compared. Both chemical lysis methods were combined with bead beating. Automated extraction methods are purported to be more efficient and consistent than manual methods. This was reflected in the results, with STAR buffer, bead beating and purification by the QIAsymphony found to be the most efficient extraction method, and also one of the most consistent, as shown by SDs.

Although a large increase in extraction efficiency for C. difficile was achieved, improvements were marginal for S. enterica. C. difficile yields improved dramatically as sporulation is likely to occur within the sample, and the increasingly vigorous lysis methods access the spore DNA. Spores are uncultivable through normal plating and culture methods; hence, copy number yield was higher than CFU input for the more efficient lysis methods. 21 The poor efficiency for Salmonella spp. suggests that their nucleic acids may be subject to effects that have failed to be identified through these experiments. For example, their nucleic acids may be susceptible to specific DNase degradation, or may bind to proteins and be discarded through the extraction process. Poor sensitivities for molecular-based tests for Salmonella spp. compared with culture-based tests have been noted previously. Schuurman et al. 12 tested a range of extraction methods for S. enterica in faecal samples, and found sensitivities using molecular methods were significantly lower than culture diagnosis. It has been reported that a manual extraction method developed by Boom et al. 22 is the most efficient; however, this method is prohibitively laborious for use in the clinical laboratory. 12

Furthermore, the wide range of pathogens targeted by MassCode and other large multiplex methods demands an extraction method suitable for all organisms. The introduction of more vigorous lysis methods increased the potential that more fragile nucleic acids, such as norovirus’ single-stranded RNA, may be lost during the process. As shown by Yang et al. 23 for respiratory pathogens, no single method is likely to be superior for all targeted organisms. The next stage of experiments aimed to validate the optimised extraction method for MassCode with the other phase 1 target organisms.

Method validation

In order to validate the optimised extraction method (STAR buffer with bead beating and purification with QIAsymphony) a batch of 80 positive clinical samples were extracted and analysed with qPCR and MassCode. An additional 20 norovirus samples were tested as, for the initial batch, the diagnosis of norovirus was found to be problematic (see below).

Overall, sensitivities for the MassCode assay improved for all organisms other than S. enterica. Norovirus also remained below the target 75% minimum sensitivity. However, only 80% of the first batch of norovirus samples were found to be positive by qPCR, most likely because of a faulty initial diagnosis of the infection. The second batch of norovirus samples were 100% positive by qPCR, leading to the confident prediction that more than 75% would be positive by MassCode.

The underperformance of S. enterica was anticipated, given the data presented above. The decrease in sensitivity compared with the original extraction method is likely because of the heterogeneity of stool samples, meaning that two separate extractions of the same clinical sample cannot be guaranteed to contain the same concentration of target organism.

In order to better predict how detection may improve using the new extraction method with a larger sample set, copy numbers of samples extracted by both the original and new extraction methods were compared. When newly detected samples are included, more samples increased in copy number than decreased with the new extraction method. Decreases in copy number are also likely to result from stool sample heterogeneity. Although some samples decreased in copy number, net percentage copy number change was calculated to be positive for all organisms. The data suggest that with a larger sample set the optimised extraction method would result in an increase in sensitivity for all organisms, as the improved nucleic acid yield would produce more samples above detection limits.

Although nucleic acid yields improved for S. enterica, the data still suggested that overall yield would be poor, and many samples would remain below the limits of detection. These results implied that poor extraction efficiency was intrinsic to Salmonella spp., potentially as a result of an unidentified biochemical or biological process. Limit of detection testing was repeated with other Enterobacteriaceae alongside non-Enterobacteriaceae to explore whether the problem was bacterial family wide or restricted to Salmonellae.

Extraction efficiency

The data suggest that members of the Enterobacteriaceae family are a challenging group of organisms to analyse directly from clinical samples by molecular methods. Although extraction efficiency of C. jejuni also appears poor at 5%, the other data suggest this yield provides sufficient DNA for molecular analysis, as MassCode sensitivity was good for Campylobacter spp.

Evidence regarding the efficiency of extraction methods is limited within the available literature, but it is possible to compare the limits of detection found here with other investigations. For example, a large multiplex PCR assay for E. coli directly from stool samples reported limits of detection of 10⁴–10⁵ CFU/ml. 24 The qPCR assay used as part of this investigation produced limits of detection of 4 × 10³; this limit of detection would be expected to increase by one to two orders of magnitude within a large multiplex assay.

Participants and sample collection

Phase 1 evaluation was conducted in the Oxford University Hospitals NHS Trust. All samples collected were initially sent to the service microbiology laboratory for faecal culture and/or C. difficile toxin testing by hospital-based doctors or GPs as a result of a suspected enteric infection. Consecutive samples positive for one of the three key pathogens being tested routinely (C. coli, C. jejuni and S. enterica) were put aside by routine microbiology staff rather than being sent for discard, as were samples that had been tested for C. difficile by toxin EIA in the service microbiology laboratory and that were negative for all pathogens. The only exclusion criterion was insufficient sample remaining after standard microbiological testing. During the period of the study, the service microbiology laboratory in the Oxford University Hospitals used an EIA to identify C. difficile manufactured by Meridian Bioscience, Inc. (Cincinnati, OH, USA). A large study conducted in 2011 independently of the manufacturer5 demonstrated that this EIA test has particularly poor sensitivity compared with the gold standard cell cytotoxicity assay (69.2%), considerably lower than an alternative EIA test (82.3%) which has been used in Oxford University Hospitals since April 2012 as part of a ‘two-step’ testing algorithm mandated by the UK Department of Health. 25 There are two reasons why EIA tests have been used historically by many trusts, and continue to be used within the ‘two-step’ testing algorithm to detect C. difficile: first, their turnaround time is shorter (≈ 1 day compared with 3 days for the cell cytotoxicity assay) and, second, their cost (consumables and staff) is lower, particularly important given the large number of tests performed (≈ 10,000 tests/year in Oxford University Hospitals). Separately to the MassCode study, all EIA-positive C. difficile samples were cultured in the parallel research laboratory under a separate research protocol. Samples positive for C. difficile on both EIA and culture were retrieved from this sample collection as reference positives for use in the MassCode study. As norovirus testing was not carried out by the microbiology laboratory unless an outbreak was suspected, qPCR-positive norovirus samples were obtained through a separate investigation conducted by a NIHR clinical fellow.

A research assistant not involved in the MassCode study therefore collected samples from the three sources (culture-positive C. difficile samples from the research laboratory, PCR-positive norovirus samples from a different study in the research laboratory and samples positive for Campylobacter spp., Salmonella spp. or negative for all pathogens from the service microbiology laboratory). The independent research assistant assigned each sample to one of 1000 pre-generated study numbers at random and maintained a list of which sample corresponded to which study number and what pathogen (if any) had been identified by the service microbiology laboratory. The research assistant produced a blinded random order of samples from the various sources for testing using the MassCode assay from this list (so that not all C. difficile- and norovirus-positive samples were processed in the same batches). As numbers of S. enterica-positive samples were lower than predicted, additional S. enterica-positive samples were sourced from Leeds, sent to the research assistant not involved with the MassCode study, assigned anonymous study numbers and periodically inserted into the workflow as for C. difficile and norovirus-positive samples. All samples were stored at 4 °C prior to processing.

All samples used in the phase 1 blinded evaluation were collected independently to those samples used to determine the extraction protocol (see Chapter 3).

Reference standard

Initial diagnosis of the target faecal pathogens was performed in accordance with Public Health England guidelines in the service microbiology laboratory. Approximately 1 g of faecal sample was inoculated into selenite broth and the broth inoculated onto xylose lysine deoxycholate (XLD) agar for culture of Salmonella spp. and Shigella spp., sorbitol MacConkey (SORB) agar for culture of E. coli O157 and Campylobacter-free blood (CAMP) agar for culture of Campylobacter spp. CAMP agar was incubated microaerophilically at 42 °C for 48 hours, XLD, SORB and the selenite broth, for culture of Salmonella spp. was incubated at 37 °C for 24 hours. After the 24-hour period, the selenite broth was inoculated onto chromogenic agar (SALM) for culture of Salmonella spp. This was incubated for a further 24 hours at 37 °C. Suspect colonies on XLD or SALM were inoculated onto analytical profile index (API^®) 10S, for identification of Salmonella spp., a Columbia agar slope for slide agglutination tests, and a MacConkey agar purity plate. If C. difficile infection was suspected, samples were subject to EIA testing for toxins A and B. Subsequent serological and sensitivity testing was performed as required for each organism identified. Throughout all reference standard and index tests, normal aseptic microbiological laboratory working practices were followed. Campylobacter was identified only to the species level, i.e. C. jejuni and C. coli were not distinguished by the reference standard testing.

Reference standard testing was performed by trainees and state-registered biomedical scientists (BMSs), but all results were confirmed by an experienced state-registered BMS before being passed onto a doctor. As reference tests were carried out before the MassCode assay was run, staff performing the reference assays did not know the results of the MassCode assay.

Blinded investigation

The full SOP for sample preparation and processing is detailed elsewhere (see Appendix 1). A total of 948 clinical samples were collected and extracted using the optimised protocol. This included 200 Campylobacter spp., 199 C. difficile, 60 S. enterica, 199 norovirus and 295 negative samples (some samples had more than one pathogen), compared with targets of 200, 200, 100, 200 and 300, respectively. Insufficient S. enterica-positive samples accrued during the period of the study. Samples were reverse transcribed and stored at −20 °C prior to amplification with the MassCode 13-plex primer mix. Following amplification, samples were cleaned according to the MassCode SOP and analysed by mass spectrometry (MS). All extraction batches included a water extraction as a control and all MS plates included MassCode calibrant controls.

All MassCode assays were conducted by a postdoctoral fellow and a research assistant who had received training on the MassCode assay from Agilent Technologies, and had jointly conducted the pre-phase 1 evaluations (see Chapter 2). The output of the MassCode assay was a number, indicating that the sample was positive on both PCR products for a pathogen (positive), near threshold values or positive on one product and near threshold value for the second product (indeterminate), or was not positive on either (negative). No expert interpretation was, therefore, required to read the MassCode outputs, which were also exported electronically from the spectrometer. The cut-offs defining positivity were set by Agilent Technologies, based on their in-house development and were not altered for this evaluation.

Researchers conducting MassCode assays were blinded to results of the reference test and all other clinical information; samples were identified only by their unique study number (assigned by a different laboratory researcher not involved with the MassCode study), i.e. were completely anonymised with regards to patient and microbiological characteristics (what pathogen, if any, had had been isolated).

MassCode results were used only for the research study and were not returned for patient management.

Sample size

In phase 1 testing, 200 positive detections of C. difficile, norovirus and Campylobacter spp. with reference standard culture/toxin detection and new PCR-based tests would produce a 95% CI around sensitivities of 90% and 96% of ± 4.2% and ± 2.7%, respectively. Testing 100 positive isolations of Salmonella spp. would produce a 95% CI around a sensitivity of 90% of approximately ± 5.9%. Samples positive for each pathogen would act as negative controls for other pathogens, together with 300 samples negative for all pathogens (total 1000 samples to be tested in Oxford and Leeds hospitals).

Analysis

Analysis of the MassCode assay results was conducted by an independent member of the team. All microbiological results were extracted from the microbiology database to confirm the reference standard test results, and identify any co-infections with other pathogens in the 13-plex MassCode panel which had not been part of original sample retrieval procedures. Sensitivity and specificity were calculated compared with the microbiological reference standard for each of the main organisms (primary reference standard), counting identification of either Campylobacter species as correct (as reference microbiological testing did not speciate). Primary analysis did not count indeterminate results (either forward or reverse primer above pre-specified threshold, but not both) as a positive: secondary analysis included these indeterminates as positive. Uncertainty was quantified using 95% CIs. All analysis were conducted in Stata 11.2 (StataCorp LP, College Station, TX, USA).

Samples which were either false negative for any of the five key pathogens or false positive for any of the 11 pathogens on MassCode were then retested using single qPCRs. This testing was not done blinded as the researchers needed to know which qPCR to run. The qPCR assays used are described elsewhere (see Chapter 2 and Appendix 1). Each qPCR assay was run in duplicate and any discrepancies confirmed by additional qPCRs.

As several unexpected positives were confirmed by qPCR results (implying the target pathogen was genuinely present in the sample but missed in the original microbiology work-up), and as some missed positives could not be detected in the original sample, even using single qPCR, a secondary analysis was carried out in which the MassCode-positive results were compared with a combined standard of reference microbiological assay plus single qPCR testing. In this secondary reference standard, samples positive on standard reference microbiology but negative on qPCR were considered negative, and samples negative on standard reference microbiology but positive on qPCR were considered positive. Possible explanations for the former are sample degradation or labelling errors. This secondary reference standard does not reflect a gold standard, but rather a standard whereby single qPCR demonstrates that, respectively, either it is unrealistic to expect the multiplex MassCode assay to detect the nucleic acid target (because it cannot be detected even with a single qPCR) or that the target was present and should have been detected by MassCode even if not identified on original microbiological testing. Effectively this standard addresses the impact on diagnostic capability of multiplexing the MassCode assay rather than running multiple single qPCRs.

Phase 2

Analysis of sensitivity/specificity of MassCode compared with standard reference methods was to be conducted at the end of phase 1 in each trust.

A small steering group was set up to review results of phase 1 (in each hospital and overall) in order to determine whether or not phase 2 should proceed. Criteria for not moving to phase 2 were pre-specified as extremely poor estimated sensitivity for detecting any of the key organisms (under 75%) such that the test would not be likely to be adopted in routine NHS practice. The choice of a threshold of 75% was based on estimated sensitivity of the currently widely used ELISAs for detecting C. difficile. The steering group comprised investigators (DWC, TEAP, MHW and ASW) plus two independent researchers – Professor Ajit Lalvani (Imperial College, expertise in development, assessment and implementation of interferon gamma release assays) and (as chairperson) Dr Christine McCartney (Director of the Health Protection Agency Regional Microbiology Network) and one independent member from outside academia, Ms Katherine Innes Ker, who has extensive commercial experience and will also represent the patient community. Inclusion of investigators and independent researchers in one oversight committee was considered appropriate because there was no patient management which could be influenced by knowledge of results to date, as no results were returned for clinical care.

Sequencing Cryptosporidium product

Single SYBR Green PCR results for Cryptosporidium, performed as part of the phase 1 analysis, suggested that the primer pair targeting Cryptosporidium was resulting in amplification of non-specific targets. To investigate, sequencing of the Cryptosporidium PCR product was performed after the primary analysis of sensitivity and specificity had been conducted. Eight samples were chosen for sequencing, including Cryptosporidium Positive Template Control, C. parvum oocysts, samples found to be positive for Cryptosporidium spp. by reference standard testing, and samples negative/not tested by reference standard testing and found to be positive by MassCode. Extracted samples were subject to conventional PCR following the MassCode SOP (see Appendix 1), but with Cryptosporidium spp. primers in singleplex; forward (5′-GAGGTAGTGACAAGAAATAACAATACAGG-3′) and reverse (5′-CTGCTTTAAGCACTCTAATTTTCTCAAAG-3′) primers were both used at 250 nm. Both primers were designed by Agilent, targeted small subunit ribosomal RNA (SSU rRNA) and were included in the MassCode 13-plex panel.

Following amplification, the PCR product was cleaned to remove unused primers and other ingredients. AMPure XP beads (Agencourt, Beckman Coulter, High Wycombe, UK) were added in a ratio of 1.8 : 1 (18 µl of beads added to 10 µl of PCR product), mixed well and incubated for 3 minutes at room temperature. The beads were separated from the supernatant by placing sample tubes on a magnetic stand and the supernatant removed and discarded. The beads were then washed twice with 200 µl of 70% molecular-grade ethanol. All remaining ethanol was removed and the beads left to air dry for 15 minutes. DNA was eluted in 40 µl of 1 × TE [tris-ethylenediaminetetraacetic acid (EDTA)] buffer by adding the buffer to each sample, mixing well, and incubating at room temperature for 2 minutes. Finally, the beads were separated from the eluate by placing the samples on a magnetic stand and the eluate transferred to new tubes.

A sequencing reaction was performed on the cleaned PCR product using the Big Dye Terminator Cycle Sequencing Kit (Life Technologies). The reaction comprised 1.75 µl of 5 × Sequencing Buffer, 0.5 µl of Big Dye, 0.5 µl of either the forward or reverse primer,and 8.75 µl of molecular-grade water per sample. A total volume of 2 µl of each sample was added to the reaction mix, and each sample was subject to a sequencing reaction with both the forward and reverse primer. Cycling conditions were as follows:

96 °C for 2 minutes96 °C for 10 seconds50 °C for 5 seconds60 °C for 3 minutes}× 30

Post-sequencing reaction products were cleaned using CleanSEQ beads (Agencourt). A volume of 10 µl of CleanSEQ beads was added to 10 µl of sequencing reaction product and mixed well. A volume of 42 µl of 85% molecular-grade ethanol was added to each bead sample mixture and, again, mixed well. Sample mixtures were placed on a magnetic stand and incubated for 3 minutes at room temperature to allow the beads to separate from the supernatant. The supernatant was then removed and discarded. The beads were subsequently washed twice with 100 µl of 85% ethanol. Once all ethanol was removed, the beads were air dried for 15 minutes and the DNA eluted from the beads in 80 µl of 0.05 mM EDTA.

The final eluate was sequenced by Applied Biosystems 3730 DNA Analyser (Life Technologies). Forward and reverse sequences for each sample were aligned in Geneious v5.6.6 (Biomatters Ltd, Auckland, New Zealand) and the consensus sequences submitted to Basic Local Alignment Search Tool (BLAST; NCBI) for identification.

Sensitivity and specificity for the main organisms

Sensitivities of each organism varied for the MassCode assay (Figures 6–9 and Table 11), ranging from 43% to 94%. Including indeterminate sample calls (where the MassCode result is near threshold values for a positive call) increased sensitivities only very slightly to 48–96%. Specificities for the MassCode assay were 95–98% or 89–96% including indeterminates. Including qPCR results also led to an increase in sensitivity and specificity, which ranged from 60% to 95% and from 97% to 100%, respectively.

FIGURE 6.

MassCode assay sensitivity and specificity for Campylobacter spp. excluding (a) and including (b) qPCR results. Sensitivity = 100 × called positive/true positive; Specificity = 100 × called negative/true negative.

FIGURE 7.

MassCode assay sensitivity and specificity for C. difficile excluding (a) and including (b) qPCR results. Sensitivity = 100 × called positive/true positive; Specificity = 100 × called negative/true negative.

FIGURE 8.

MassCode assay sensitivity and specificity for norovirus excluding (a) and including (b) qPCR results. Sensitivity = 100 × called positive/true positive; Specificity = 100 × called negative/true negative.

FIGURE 9.

MassCode assay sensitivity and specificity for S. enterica excluding (a) and including (b) qPCR results. Sensitivity = 100 × called positive/true positive; Specificity = 100 × called negative/true negative.

The best-performing organism was C. difficile, although Campylobacter spp. and norovirus also had sensitivities and specificities well above the 75% threshold required by the MassCode protocol in order to proceed to phase 2, with the lower limits of the 95% CI exceeding 83% (sensitivity) and 92% (specificity). However, the sensitivity of S. enterica remained well below this threshold; even including qPCR results the upper limit of the 95% CI around the estimated sensitivity of 60% was just below 75%.

Inspection of the copy numbers returned by qPCR for missed positive samples revealed that 59% of samples positive for Campylobacter spp., 77% of C. difficile-positive samples and 89% of norovirus-positive samples had copy numbers < 100, the previously defined limit of detection. For S. enterica, 61% of the missed positives that were recovered by qPCR had copy numbers < 10, suggesting very low nucleic acid yields.

Tables 12a-c show more detail about the discrepancies between MassCode positives and reference microbiology. C. difficile was the most accurate target, with all but one of the unexpected positives confirmed positive by qPCR (see Table 12c). These samples may reflect isolates that were not producing toxin and so not causing disease at the time of the diarrhoea, i.e. may have been carried toxigenic strains (colonisation). The primer targets the tcdB toxin gene, so would not have identified non-toxigenic strains. The reference laboratory test for C. difficile is an EIA-based test for toxin; however, as this test is known to report false positives and false negatives, the samples chosen for this investigation had been confirmed C. difficile positive by culture.

Higher rates of identification of C. difficile on moving from toxin-based tests to NAAT- or PCR-based tests have been widely reported, and identification of C. difficile on the basis of PCR alone is not recommended, as outcomes are similar in PCR-positive toxin-negative and PCR-negative toxin-negative individuals. 5 All 13 samples testing postive for C. difficile in the laboratory but missed by MassCode were identified by qPCR, suggesting that these errors were simply due to the multiplex assay and the loss of sensitivity seen in large multiplex assays, rather than to the target primers or the extraction method.

For the other species, fewer of the MassCode-positive reference-negative samples were confirmed as containing the target organism using qPCR (50% for Campylobacter spp. and norovirus, 12% for S. enterica). S. enterica had a particularly high rate of false-positive calls by MassCode among the main target organisms, with only two of the unexpected positives being confirmed by qPCR. Most of those positive on PCR had evidence of a test negative on the microbiology system, i.e. not isolated rather than not tested for in error. Although this could, in theory, represent detection of lower levels of organisms than required to cause disease by MassCode, the relatively high limits of detection for the MassCode assay for S. enterica suggest that it is more likely these are genuinely false-positive calls by the multiplex assay.

For Campylobacter spp., most (17/19) of the reference laboratory-positive MassCode-negative samples were positive on qPCR, again, suggesting that these errors were simply because of the multiplex assay, not the target primers or the extraction method. Just over half (19/34) the S. enterica laboratory-positive, MassCode-negative samples were positive using a single PCR, suggesting that both primers or extraction method and multiplexing the assays were playing important roles in the lower sensitivity for this organism. Lower PCR positivity for norovirus may reflect degradation of the single-stranded RNA even with a single assay or possibly sample degradation with storage.

Interestingly, of the 181 samples positive for Campylobacter spp. using both reference laboratory and MassCode assays, MassCode identified 130 C. jejuni infections, 10 C. coli infections and 41 co-infections with both C. jejuni and C. coli. It is not known whether these represent genuine co-infections or cross-reactive primer pairs, as Campylobacter is not routinely speciated in the service laboratory.

Discrepant results for additional organisms

Unexpected positive samples were all retested by the appropriate PCR assay, as for the main organisms. E. coli, S. Typhi, Shigella spp., Cryptosporidium spp., and Giardia spp. all also showed a high rate of false-positive calling (Table 14). However, some unexpected positives were also confirmed by PCR, including four (0.4% of all samples) E. coli O157 positives (three of which had definitely had faecal culture without identifying this pathogen), 13 (1.4%) Giardia spp. (none tested for parasites in the service microbiology laboratory), seven (0.7%) Cryptosporidium (one sample tested for this and no oocysts seen) and three (0.3%) Shigella spp. (two had faecal culture without identifying this pathogen). Nevertheless, these unexpected true positives were identified at the cost of a substantial number of false positives, particularly for rarer species, Cryptosporidium spp. and Giardia spp., with 7% and 13%, respectively, of the total 948 samples testing positive for these species by MassCode.

Co-infections

Co-infections were defined by either expected positives from reference standard testing or unexpected positives by MassCode followed by qPCR to confirm the unexpected positive. C. jejuni and C. coli were included as Campylobacter spp. for this analysis. Of the 948 samples, 32 (3.4%) were confirmed positive for two of the four main organisms (Table 15). Overall, 46 (4.9%) of the 948 samples were confirmed positive for more than one organism, only two of which were confirmed to be positive for more than two organisms. Co-infections with parasites (Cryptosporidium spp. and Giardia spp.) accounted for 11 of the co-infections, whereas co-infections with C. difficile and Campylobacter spp. or norovirus were most common. Of note, only eight Cryptosporidium-positive samples were found in total (one known co-infection, seven unexpected positives), but six out of eight organisms were present with another pathogen, suggesting this organism might be more commonly carried than previously suspected. All of the four unexpected E. coli O157 organisms found were also found with other, more common species, as were five of the unexpected 13 Giardia spp. organisms.

As a consequence of the relatively low sensitivity and specificity of the MassCode assay, far more samples were identified as having co-infections with enteric pathogens using these results alone (Figure 11). A total of 48 (5.1%) were MassCode positive for two of the four main organisms; eight samples were microbiologically negative. Of the 948 tested samples, only 241 (25%) had no organism of any type identified by MassCode, 181/295 (61%) of those microbiologically negative compared with 60/653 (9%) positive for any of the main four pathogens. In one sample up to six organisms were identified by MassCode; 159 samples contained two organisms, 55 samples contained three organisms, 17 contained four organisms and four samples contained five organisms. As expected from the results above, Giardia spp. and Cryptosporidium spp. were commonly incorrectly identified as additional co-infecting organisms.

FIGURE 11.

Number of pathogens in samples by MassCode assay. (a) Including only main pathogens (Campylobacter spp., C. difficile, norovirus and S. enterica); and (b) including all 11 pathogens.

Sequencing Cryptosporidium product

The sections Specificity of MassCode for additional organisms and Discrepant results for additional organisms demonstrated that, although Cryptosporidium spp. were detected in high numbers by MassCode, 94% of these were confirmed to be false positive by singleplex PCR. Results from the sequencing of the Cryptosporidium product confirmed that the primers targeting Cryptosporidium SSU rRNA were also amplifying Candida spp. Of the four clinical samples tested, two that were called positive for Cryptosporidium spp. by MassCode were found to have produced Candida spp. products rather than Cryptosporidium products. This result helps to explain the high false-positive call rate by MassCode, as Candida spp. could be distinguished by melt curve analysis when tested by singleplex SYBR Green PCR. Although not confirmed through sequencing, it is possible that the high false-positive rate seen in Giardia (only 10% of MassCode positives were confirmed positive by singleplex PCR) may also be as a result of unspecific primer pairs being used, as the Giardia primers selected for use in MassCode also target SSU rRNA.

xTAG Gastrointestinal Pathogen Panel

Nucleic acid extracts were screened with the xTAG GPP according to the manufacturers’ protocol (www.luminexcorp.com/Products/Assays/ClinicalDiagnostics/xTAGGPP/).

The GPP is a qualitative multiplex test intended for the simultaneous detection and identification of nucleic acids from multiple gastroenteritis-causing viruses, parasites and bacteria (including toxin gene detection) in human stool samples that are fresh, frozen or in a holding medium, from individuals with signs and symptoms of infectious colitis or gastroenteritis.

The following pathogen types and subtypes are identified using the xTAG GPP:

• adenovirus 40/41

• Campylobacter

• Clostridium difficile toxin A/B

• Cryptosporidium

• Entamoeba histolytica

• Escherichia coli O157

• enterotoxin-producing E. coli (ETEC) – heat-labile toxin/heat-stable toxin

• Giardia

• norovirus GI/GII

• rotavirus A

• Salmonella

• Shiga-like toxin-producing E. coli (STEC) stx 1/stx 2

• Shigella

• Vibrio cholerae

• Yersinia enterocolitica.

The xTAG Gastrointestinal Pathogen Panel (xTAG GPP) incorporates a multiplex reverse transcriptase-polymerase chain reaction (RT-PCR) with Luminex’s proprietary universal tag sorting system on the Luminex platform. Each target or IC in the sample results in PCR amplicons ranging from 58 to 293  base pair (bp) (not including the 24-mer tag). The RT-PCR product is added to a hybridisation/detection reaction containing the universal tag and the streptavidin, R–phycoerythrin conjugate. Each Luminex bead population detects a specific bacterial, viral or parasitic target or assay control through a specific antitag/tag hybridisation.

Briefly, for each sample, 10 µl of extracted nucleic acid was amplified in a single multiplex RT-PCR containing xTAG GPP primer mix, and xTAG^® OneStep buffer and enzyme mix (Luminex). Two positive controls (norovirus GI RNA and S. enterica DNA) and three negative controls (molecular-grade water) were included per PCR run (96 reactions). The bacteriophage MS2 added prior to nucleic acid extraction acted as an internal/inhibition control for the assay. Amplification was carried out on a Mastercycler^® Pro thermal cycler (Eppendorf UK Ltd, Stevenage, UK), using the following cycling conditions:

1 × 53 °C, for 20 minutes

1 × 95 °C for 15 minutes

36 × 95 °C for 30 seconds

58 °C for 30 seconds

72 °C for 30 seconds

1 × 72 °C for 2 minutes

1 × hold at 4 °C.

For the hybridisation/detection reaction, 5 µl of RT-PCR product was added to the appropriate well of a 96-well plate containing xTAG GPP bead mix (20 µl). xTAG reporter solution (75 µl; xTAG 0.22 streptavidin, R–phycoerythrin conjugate and xTAG^® Reporter Buffer) was then added to the wells. Hybridisation was performed on the Mastercycler Pro thermal cycler, programmed for 60 °C for 3 minutes, followed by 45 °C for 45 minutes.

Following hybridisation, the median fluorescence intensity (MFI) was generated for each xTAG bead population using the Magpix^® System (Luminex; pre-heated to 45 °C). Data were analysed with the xTAG Data Analysis Software for the GPP (TDAS; Luminex) to establish the presence or absence of bacterial, viral or parasitic targets and/or controls in each sample.

Blinded investigation

Two sets of Luminex assays were run, blinded to pathogens isolated from the original sample using the same anonymised study numbers (see Chapter 3). The sensitivity and specificity of the Luminex test, based on the same standardised nucleic acid extraction (see Appendix 1) as had been done for MassCode, was investigated first. All remaining DNA/RNA extracts made in Oxford for phase 1 (which had been stored at –20 °C) were shipped to Leeds (delay of 26 weeks between completion of MassCode testing and beginning of Luminex testing) (denoted ‘MassCode’ and/or ‘Oxford’ extracts). Comparison of MassCode with Luminex results on the same extracts removes one source of variability.

However, the MassCode extraction SOP does not follow the manufacturer’s recommendations for the Luminex assay, and, therefore, all original stool samples (stored at 4 °C) were also shipped to Leeds (see Table 10 for dates of collection of samples), and in June–July 2013 all samples with sufficient material remaining were re-extracted using the QIAsymphony^® sample preparation (SP) (Qiagen) automated nucleic acid extraction platform following the manufacturer’s recommendations, and tested in a separate run (denoted ‘fresh’ and/or ‘Leeds’ extracts). Prior to extraction, stool samples (200 µl/pea-sized sample) were emulsified in a pre-treatment solution [900 µl of L6 lysis buffer (Severn Biotech) and 20 µl of isoamyl alcohol], and spiked with 10 µl of MS2 bacteriophage (109 copies/µl; Luminex). Following a bead beating step with Lysing Matrix E (3000 r.p.m. for 1 minute; MP Biomedicals, Santa Ana, CA, USA), samples were centrifuged at 12000 r.p.m. for 15 seconds. The resultant supernatant (250 µl) was transferred to a fresh tube and diluted 1 : 1 with sterile phosphate-buffered saline (PBS). Nucleic acid was extracted using the Virus/Pathogen Mini Kit (Qiagen) and the QIAsymphony^® ‘Pathogen Complex 200 with IC’ programme. The final elution volume was set at 110 µl. Negative controls (sterile PBS) were included in each extraction run. Extracts were stored at –20 °C prior to testing.

All Luminex assays were conducted by one pre-registration clinical scientist, who had received training on the assay from Luminex. The output of the xTAG data analysis software for the GPP was a presence or absence call for each pathogen alongside MFI value. No interpretation was therefore required to read the Luminex outputs, which were also exported electronically.

The pre-registration clinical scientist conducting Luminex assays was blinded to results of the reference microbiology test and the results of previous qPCRs done in Oxford, and all other clinical information; samples were identified only by their unique study number and were completely anonymised with regards to patient and microbiological characteristics (what pathogen, if any, had been isolated). Luminex results were only used for the research study and were not returned for patient management.

As detailed previously (see Chapter 3), samples that were unexpectedly positive or negative for target organisms were retested in duplicate using qPCR assays with either dual-labelled probes or SYBR Green master mix. This testing was not done blinded, as the researcher needed to know which qPCR to run. The qPCR assays used are described elsewhere (see Chapter 2 and Appendix 1).

Analysis

Analysis of the Luminex assay results was conducted by an independent member of the team, using an identical analysis plan to the MassCode assay. The Luminex assay has two targets for C. difficile (tcdA and tcdB gene) and includes two primers against two major groups of noroviruses (GI and GII strains). A positive call on either target was counted as indicating the presence of the organism. However, the single qPCRs targeted only the tcdB gene or the GII strain. As only one A+B– strain of C. difficile has ever been described, missed or unexpected positives on either C. difficile target were retested. However, as GI and GII norovirus strains can co-circulate during an epidemic, only norovirus missed or unexpected positives based on the GII strain were retested with single qPCR.

Sensitivity and specificity were calculated compared with the microbiological reference standard for each of the main organisms (primary reference standard). As for the MassCode assay, several unexpected positives on the Luminex assay were confirmed by qPCR results (implying the target pathogen was genuinely present in the sample, but missed in the original microbiology work-up). In secondary analysis we therefore compared the Luminex-positive results with a combined standard of reference microbiological assay plus single qPCR testing of unexpected and missed positives from the specific Luminex assay. In these analyses, additional qPCR results that had been generated from MassCode missed/unexpected positives were ignored, because they would not have been known had only the Luminex assay been evaluated. Thus, these analyses aimed to consider each assay/extraction method on a like-for-like basis. Thus, samples positive on standard reference microbiology but negative on the Luminex assay and confirmed negative on qPCR were considered negative, and samples negative on standard reference microbiology but positive Luminex and confirmed positive on qPCR were considered positive.

See Chapter 5 for a comparison of the results across MassCode and Luminex assays.

Sensitivity and specificity for the main organisms

For Campylobacter spp., C. difficile and norovirus, high sensitivities (> 92%, most > 97%) and specificities (> 96%) were observed, regardless of extraction method and regardless of whether comparisons were made with microbiology alone or microbiology plus qPCR results (Figures 12–14 and Table 17). Interestingly, the lower sensitivities were observed on the Luminex test of freshly extracted material (Campylobacter spp. 98%, C. difficile 96%, norovirus 92%).

FIGURE 12.

Luminex assay sensitivity and specificity for Campylobacter spp., using original MassCode/Oxford and fresh/Leeds extracts, and excluding (a) and including (b) qPCR results. Sensitivity = 100 × called positive/true positive; specificity = 100 × called negative/true negative.

FIGURE 13.

Luminex assay sensitivity and specificity for C. difficile, using original MassCode/Oxford and fresh/Leeds extracts, and excluding (a) and including (b) qPCR results. Sensitivity = 100 × called positive/true positive; specificity = 100 × called negative/true negative.

FIGURE 14.

Luminex assay sensitivity and specificity for norovirus, using original MassCode/Oxford and fresh/Leeds extracts, and excluding (a) and including (b) qPCR results. Sensitivity = 100 × called positive/true positive; specificity = 100 × called negative/true negative.

However, major differences in results with the same Luminex assay were found for S. enterica (Figure 15). On the original MassCode/Oxford extracts, sensitivity against microbiological testing was 84% (95% CI 73% to 93%). Although this clearly would have some limitations in clinical practice, it was a substantial improvement over the MassCode sensitivity. However, sensitivity when assayed using fresh/Leeds extracts dropped to 46%, very similar to the MassCode sensitivity, with a corresponding increase in specificity from 92% to 99%. Results were similar including qPCR results, with sensitivity for detecting S. enterica from freshly extracted material of only 60% with the upper limit of the 95% CI around the estimated sensitivity just below 79%.

FIGURE 15.

Luminex assay sensitivity and specificity for S. enterica, using original MassCode/Oxford and fresh/Leeds extracts, and excluding (a) and including (b) qPCR results. Sensitivity = 100 × called positive/true positive; specificity = 100 × called negative/true negative.

Discrepant results for the main organisms

Table 17(c) and (d) shows more detail about the discrepancies between Luminex positives and reference microbiology, together with results of confirmatory qPCR testing.

Campylobacter was the most accurate target, with all unexpected positives on Luminex confirmed positive by qPCR on both original MassCode/Oxford extracts and freshly extracted material. All samples had undergone faecal culture in the service laboratory, so this organism had been missed. However, it is still possible that the Campylobacter spp. identified by Luminex were present only as carriage and were not present in sufficient numbers to be the cause of disease. The Luminex assay also missed very low numbers of microbiologically identified Campylobacter on both assay runs (two and seven, respectively); neither call on the first run could be confirmed with qPCR, and four out of seven calls on the second run could not be confirmed, suggesting either sample degradation or laboratory error, but demonstrating that the Luminex multiplex assay was highly specific for Campylobacter spp. For Campylobacter spp., the Luminex assay performed similarly on original MassCode/Oxford extracts and material freshly extracted in Leeds.

For C. difficile and norovirus, the Luminex assay had slightly lower sensitivity, more so with the original MassCode/Oxford extracts than the fresh/Leeds extracts, with a corresponding reduction in specificity with the fresh/Leeds extracts compared with the MassCode/Oxford extracts.

For C. difficile, most of the unexpected positives identified by Luminex on both runs were confirmed by qPCR, even in the minority of samples that were positive based on the tcdA gene only (Table 18). In contrast to the MassCode assay and the single qPCR primers for C. difficile, the Luminex assay targets both the C. difficile tcdA and tcdB toxin genes. The majority of toxin-producing strains are A+B+, but there is one important clade which is A–B+ (identified only using the MassCode/Oxford extracts, not the freshly extracted material). A+B– strains have not been confirmed to occur in vivo, although they have been engineered in vitro in knockout studies. Therefore, where Luminex identified a strain as being A+B–, it was either actually an A+B+ strain (a false negative for the tcdB gene) or the sample was actually negative (a false positive for the tcdA gene and a false positive for the pathogen). Both occurred – there were instances of unexpected A+B+ strains called by Luminex that could not be confirmed by qPCR for the tcdB gene and instances of A+B– samples called by Luminex that could be confirmed by tcdB gene qPCR. Very few samples had C. difficile identified by the laboratory but not by Luminex (based on either gene target).

For norovirus, a lower proportion (just under half) of unexpected positives identified in either Luminex assay were confirmed by qPCR. Lower PCR positivity for norovirus may reflect degradation of the single-stranded RNA even with a single assay, or possibly sample degradation with storage. Proportionately, unexpected positives were more likely to be identified as the GI strain than as the widely circulating GII strain (Table 19). No sample was identified with both GI and GII strains. As qPCR testing was based on GII, it was not possible to assess whether or not these GI strain calls reflect false positives, and difficult to assess whether or not these strains are now more likely to be carried without causing disease.

The most marked mismatch between the Luminex assay run on MassCode DNA/RNA extractions compared with freshly extracted samples was in S. enterica. Of the microbiology-negative MassCode/Oxford extracts, 8.3% (73/879) produced unexpected S. enterica positives, but only 16% (9/58) of microbiology-positives were missed. Of these 73 unexpected and nine missed S. enterica positives, only two and one, respectively, were confirmed on qPCR. Both of the two samples unexpectedly confirmed positive for S. enterica with the MassCode/Oxford extracts run on the Luminex assay had definitely been cultured originally in the service microbiology laboratory without identifying S. enterica; one was also positive for S. enterica using the MassCode assay. These two unconfirmed unexpected positives are particularly concerning for this potentially virulent pathogen. Unconfirmed unexpected positives occurred in similar numbers using the Luminex assay on MassCode/Oxford extracts and using MassCode assay on MassCode/Oxford extracts.

In contrast, running the same assay on freshly extracted samples substantially reduced the proportion of unexpected positives to 1.1% (9/806), but correspondingly far more microbiology positives were missed (55%;18/33). Thus, the sensitivity of the Luminex assay on fresh/Leeds extracts was similar to that of the MassCode assay. None of the nine unexpected positives, but 10 of the 18 missed positives using Luminex, were confirmed by qPCR, suggesting that both the primers/extraction method and multiplexing the assays make important contributions to the reduced sensitivity for this organism; as for MassCode.

Specificity of Luminex for additional organisms

It was possible to calculate the specificity of the Luminex assay for the additional targets. Known co-infections identified by the routine laboratory were excluded from each specificity analysis (one co-infection with Shigella spp. and one with Cryptosporidium spp. as for MassCode, plus one with Entamoeba histolytica). All unexpected positives for organisms in common with the MassCode assay were tested with singleplex assays using the MassCode primers: secondary analyses considered any confirmed positives as a positive standard (i.e. excluded them from analyses of specificity). However, for those targets not included in the MassCode primer set, additional PCRs were not available and so were not carried out. For these pathogens, only known microbiology positives were excluded from analyses of specificity.

The results are summarised in Tables 20 and 21 and Figures 16 and 17. Considering the Luminex assay run on MassCode/Oxford extracts, specificity was lowest for Giardia spp., at 97%, after excluding qPCR positives from the reference standard (compared with 88% for MassCode), and equal to or more than 99% for E. coli O157, Shigella spp. and rotavirus. Specificities were considerably higher than for MassCode, particularly for Cryptosporidium spp. (88% specificity with MassCode excluding qPCR-positives). Specificities were slightly higher for the Luminex assay run on fresh/Leeds extracts.

FIGURE 16.

Specificity of the Luminex assay run on MassCode/Oxford extracts across all organisms. (a) Reference microbiology; and (b) reference microbiology + PCR. Specificity = 100 × called negative/true negative. Light green bars are tests where no qPCR was done to exclude positives missed by original microbiology.

FIGURE 17.

Specificity of the Luminex assay run on fresh/Leeds extracts across all organisms. (a) Reference microbiology; and (b) reference microbiology + PCR. Specificity = 100 × called negative/true negative. Light green bars are tests where no qPCR was done to exclude positives missed by original microbiology.

Discrepant results for additional organisms

Unexpected positive samples were retested using the same PCR primers/probes as used to investigate unexpected positives found by the MassCode assay (see Discrepant results for the main organisms), where these were available (Table 21).

Although false-positive call rates were substantially lower than for the MassCode assay, there were still a reasonable number (45) for Giardia using the MassCode/Oxford extracts, of which only 31% were confirmed on qPCR. In contrast, all but one of the unexpected Cryptosporidium spp. calls using the Luminex assay on the MassCode/Oxford extracts were confirmed by qPCR. We were not able to confirm what proportion of the substantial minority of calls for Entamoeba (n = 22) and ETEC (n = 16) were genuine. However, there were few false positives for Shigella spp., and rotavirus. In contrast to the MassCode assay, using MassCode/Oxford extracts with the Luminex assay, these unexpected true positives were identified without increasing the number of false negatives.

In contrast, performing the identical assay on fresh/Leeds extracts, the number of unexpected positives (false positives vs. microbiology) dropped substantially, most notably for S. enterica and Giardia spp. Correspondingly, a greater proportion of these unexpected positives were confirmed on qPCR. However, this was accompanied by a larger number of false negatives for S. enterica.

Co-infections

Co-infections were defined by either expected positives from reference standard testing or unexpected positives by Luminex followed by qPCR to confirm the unexpected positive. Each assay was considered separately.

Using the original MassCode/Oxford extracts, 47 (5.0%) samples had two organisms identified (none had three or more), with 26 (2.8%) having two of the main four pathogens. The proportion identified with co-infections was smaller using the fresh/Leeds extracts; 26 (3.1%) samples had two organisms identified (none had three or more), with 16 (1.9%) having two of the main four pathogens (Table 22).

Co-infections with parasites (Cryptosporidium spp. and Giardia spp.) accounted for 14 and 6 of the co-infections, respectively, while co-infections with C. difficile and Campylobacter spp. or norovirus were most common overall. Using the MassCode/Oxford extract, there were 10 co-infections with norovirus and Cryptosporidium spp. Of note, only 12 and three Cryptosporidium-positive samples were found in total, but 11 and three were present with another pathogen, suggesting this might be more commonly carried than previously suspected, as suggested by the MassCode results. Many of the unexpected Shigella spp. and Giardia spp. were also found as co-infections with other more common species.

As a consequence of the higher sensitivity and specificity of the Luminex compared with the MassCode assay, fewer samples were identified as having co-infections with enteric pathogens. Nevertheless, because sensitivity/specificity is effectively cumulated across the large number of tests in the multiplex panel, using the MassCode/Oxford extracts (Figure 18), 79 (8.4%) samples were positive for two of the four main organisms (12 microbiologically negative samples) and 6 (0.9%) were positive for three of the four on Luminex assay. Of the 937 tested samples, only 202 (22%) had no organism of any type identified by the Luminex assay on MassCode/Oxford extracts, 191 out of 292 (65%) of those microbiologically negative compared with 11 out of 645 (2%) of those positive for any of the four main pathogens. Up to four organisms were identified by Luminex assay in one sample based on the MassCode/Oxford extracts, with 127 samples having two organisms identified by Luminex, 25 samples having three organisms identified and two samples having four organisms.

FIGURE 18.

Number of pathogens in samples by Luminex assay on MassCode/Oxford extracts. (a) Including only main pathogens (Campylobacter spp., C. difficile, norovirus, S. enterica); and (b) including all 11 pathogens.

Using the fresh/Leeds extracts (Figure 19), only 20 (2.4%) samples were positive for two of the four main organisms (just one microbiologically negative sample), and one (0.2%) sample was positive for all four on Luminex assay. Of the 839 tested samples, only 278 (33%) had no organism of any type identified by the Luminex assay on fresh/Leeds extracts, 243 out of 295 (82%) of those microbiologically negative compared with 35 out of 544 (6%) positive for any of the main four pathogens. Similarly to the MassCode/Oxford extracts, up to four organisms were identified by Luminex assay in one sample based on the fresh/Leeds extracts, but in far fewer cases, with only 42 samples with two organisms identified by Luminex on fresh/Leeds extracts, two samples with three organisms, and one sample with four organisms.

FIGURE 19.

Number of pathogens in samples by Luminex assay on fresh/Leeds extracts. (a) Including only main pathogens (Campylobacter spp., C. difficile, norovirus, S. enterica); and (b) including all 11 pathogens.

Analysis

Results from the three assays (MassCode panel on MassCode/Oxford extracts, Luminex panel on MassCode/Oxford extracts, Luminex panel on fresh/Leeds extracts) were as described above. Pathogens were the four main pathogens common to both assays (Campylobacter spp., C. difficile, norovirus, S. enterica), additional pathogens common to both assays (Giardia spp., Cryptosporidium spp., E. coli O157, rotavirus, Shigella spp.) plus pathogens only in the MassCode (S. Typhi/Paratyphi) or Luminex (adenovirus, E. histolytica, Y. enterocolitica, V. cholerae and ETEC and STEC E. coli).

The results reported previously (see Chapters 3 and 4) considered each assay (MassCode and Luminex) separately, and used single qPCRs to confirm or refute unexpected and missed positives identified on that specific assay. In order to compare results across the different assays/extraction methods, one single common reference standard is needed. For all samples, a microbiological diagnosis of infecting pathogens was available, and so this could have been used. However, each assay found some unexpected positives that were confirmed by qPCR results, implying that the target pathogen was genuinely present in the sample but had been missed in the original microbiology work-up, and it is reasonable to expect that each multiplex assay should have detected this pathogen if it could have been identified using single qPCRs. However, because of time and financial constraints, not every sample was tested with every qPCR, for each assay/extraction method, single PCRs had been performed only for missed and unexpected positives with that assay/extraction method. Further, the comparison with qPCR-confirmed diagnosis on the individual assays was designed to investigate whether there were specific challenges with the primers being used in the multiplex and not to assess the ability of each assay to provide an NHS-relevant diagnostic tool. For example, all samples included with microbiologically confirmed C. difficile had been tested with an EIA and cultured, suggesting that it was highly likely that the sample contained C. difficile even if subsequently this could not be confirmed by single PCR.

Therefore, for this combined analysis, we used a reference standard designed to reflect the best information about whether or not a pathogen was present in the original specimen, defining a sample to be positive for an organism if (i) the organism had been identified microbiologically, regardless of whether or not its subsequent single PCR was positive or negative, or (ii) the organism had not been identified microbiologically, but one or more single PCRs were positive for the organism.

Samples negative for each organism had no microbiological isolation, and had either not been tested by PCR or all PCRs done were consistently negative (note: single PCRs were all done in duplicate, i.e. were confirmed presence of an organism).

Based on the pathogens common to the Luminex and MassCode assays, Table 23 shows the composition of the phase 1 sample panel. Two pathogens were identified in a total of 60 (6.3%) and three pathogens were identified in two samples (0.2%). In only seven of the samples were co-infections identified by microbiological testing (five S. enterica + Campylobacter spp., one Campylobacter spp. + Cryptosporidium spp., one S. enterica + Shigella spp.). Of the 16 samples confirmed to contain Cryptosporidium spp., 13 represented co-infections as did 10 out of 19 confirmed Giardia-positive samples and three out of five confirmed Shigella spp.-positive samples.

Sensitivity and specificity compared with a combined microbiology plus polymerase chain reaction-positive reference standard

Figures 20–23 show sensitivity and specificity across the three assay runs for the four key pathogens. All 948 phase 1 samples were tested with the MassCode assay; because varying amounts of MassCode/Oxford extract and original sample remained, only 937 samples were tested using the Luminex assay on MassCode/Oxford extracts, and 839 samples were tested using the Luminex assay on fresh extracts. In total, 830 samples were tested using all three methods. Pathogens identified varied across the different assays, and even between the Luminex assays based on different extracts, with only moderate to reasonable within-assay agreement for positive calls (42% for S. enterica, 86–94% for other pathogens), and discrepancies within Luminex run for negative calls for S. enterica (91%).

FIGURE 20.

Campylobacter species (including all PCRs). Sensitivity = 100 × called positive/true positive (including microbiology positive plus any PCR positive). Specificity = 100 × called negative/true negative (excluding microbiology negatives with any PCR positive). Not all samples had sufficient volume to be included in all test evaluations, denominators as shown.

FIGURE 21.

C. difficile (including all PCRs). Sensitivity = 100 × called positive/true positive (including microbiology positive plus any PCR positive). Specificity = 100 × called negative/true negative (excluding microbiology negatives with any PCR positive). Not all samples had sufficient volume to be included in all test evaluations, denominators as shown.

FIGURE 22.

Norovirus (including all PCRs). Sensitivity = 100 × called positive/true positive (including microbiology positive plus any PCR positive). Specificity = 100 × called negative/true negative (excluding microbiology negatives with any PCR positive). Not all samples had sufficient volume to be included in all test evaluations, denominators as shown.

FIGURE 23.

S. enterica (including all PCRs). Sensitivity = 100 × called positive/true positive (including microbiology positive plus any PCR positive). Specificity = 100 × called negative/true negative (excluding microbiology negatives with any PCR positive). Not all samples had sufficient volume to be included in all test evaluations, denominators as shown.

Salmonella enterica

Three additional S. enterica positives were identified by PCR testing of unexpected positives from one of the three assays. Overall, therefore, 63 samples had S. enterica identified by microbiology or PCR, and 885 did not have S. enterica identified by either (Figure 23).

• Ten (16%) positive samples did not have S. enterica identified by at least one of the assays [five were negative on all three assays, four on two, and one was only tested (negative) with MassCode]. Conversely 53 out of 63 did have S. enterica identified at least once, 29 in at least two assays. Of 36 positive samples tested with all three assays, 12 (33%) were positive on all three; 15 (42%) were positive on both MassCode/Oxford and fresh/Leeds extracts tested by Luminex.

  • A total of 25 positive samples did not have sufficient material remaining to assay fresh/Leeds extracts using the Luminex assay. Of these ,11 were positive only using Luminex on MassCode/Oxford extracts and 10 were positive on both other assays, a similar ratio to the 14 and 16, respectively, where all three assays had been run.

• Out of 885 negative samples, 794 (90%) never had S. enterica identified on any assay run. Of 794 negative samples tested with all three assays, 712 (90%) were negative on all three; 725 (91%) were negative on both MassCode/Oxford and fresh/Leeds extracts tested by Luminex.

Figures 24–26 show the specificities across all organisms tested, based on all microbiological results and whatever PCR results were available. Overall, specificity of the MassCode assay is clearly suboptimal (see Figure 24), particularly for Giardia and Cryptosporidium spp. In contrast, Luminex achieves better specificity on the MassCode/Oxford extracts (see Figure 25), but would really be suitable only for testing thousands of samples based on results from the fresh/Leeds extracts (see Figure 26).

FIGURE 24.

Specificity: MassCode assay on MassCode (Oxford) extract. Specificity = 100 × called negative/true negative (excluding microbiology negatives with any PCR positive). Light green bars are tests where no qPCR was done to exclude positives missed by original microbiology.

FIGURE 25.

Specificity: Luminex assay on MassCode (Oxford) extract. Specificity = 100 × called negative/true negative (excluding microbiology negatives with any PCR positive). Light green bars are tests where no qPCR was done to exclude positives missed by original microbiology.

FIGURE 26.

Specificity: Luminex assay on fresh (Leeds) extract. Specificity = 100 × called negative/true negative (excluding microbiology negatives with any PCR positive). Light green bars are tests where no qPCR was done to exclude positives missed by original microbiology.

Table 24 shows the trade-off between sensitivity and specificity, which is particularly apparent for Cryptosporidium and Giardia spp. Cryptosporidium spp. were identified in 16 samples in total (one known to be microbiology positive). Four samples were identified as containing Cryptosporidium spp. in all three assays, seven only in MassCode and Luminex assays run on MassCode/Oxford extracts, and four and one in each of these assays singly. However, the increased sensitivity of MassCode came at the cost of 107 false positives, and it is unclear whether these organisms really were causing disease or simply co-carried. Similarly, Giardia spp. were identified in 19 samples (none known from microbiology). Ten samples were identified as containing Giardia spp. in all three assays, another three in two of the three assays, and six in only one assay. Again, the increased sensitivity of MassCode came at the cost of 111 false positives.

TABLE 24 - Identification of samples positive for other organisms based on combined microbiological and PCR testing

N/A, not applicable.

Based on microbiology and any available qPCR results from testing original false positives.

Not all samples tested.

No specific primers, so results vs. microbiology only.

Using E. coli stx1 and stx2 primers.

Organism	Missed positivesa (number missed/number positive)			Additional false positives
	MassCode, MassCode/Oxford extracts	Luminex, MassCode/Oxford extractsb	Luminex, fresh/Leeds extractsb	MassCode, MassCode/Oxford extracts	Luminex, MassCode/Oxford extractsb	Luminex, fresh/Leeds extractsb
Main pathogens
Campylobacter spp.	24/225	13/224	20/141	20	0	0
C. difficile	25/226	4/221	14/224	1	10	4
Norovirus	23/207	6/206	17/207	8	20	9
S. enterica	35/63	10/61	21/36	14	71	9
Additional pathogens in common with MassCode
E. coli O157c	N/A	N/A	N/A	17	5	3
Shigella spp.	2/5	0/5	1/5	59	4	0
Cryptosporidium	1/16	4/16	12/16	107	1	0
Giardia spp.	3/19	5/19	7/19	111	31	5
Rotavirus	0/0	0/0	0/0	0	2	1
Additional pathogens in MassCode only
S. Typhi	0/0	N/A	N/A	21	N/A	N/A
Additional pathogens in Luminex only
E. histolyticac	N/A	1/1	N/A	N/A	22	2
Y. enterocoliticac	N/A	0/0	0/0	N/A	4	4
Vibrio choleraec	N/A	0/0	0/0	N/A	0	0
Adenovirusc	N/A	0/0	0/0	N/A	2	2
ETECc	N/A	0/0	0/0	N/A	16	11
STECd	N/A	4/5	3/3	N/A	3	2

Main four pathogens

Even combining results across the four main pathogens, there was reasonable variety in what would have been concluded about each sample (Table 25). Overall agreement with the combined microbiology and all PCRs performed standard was 85.6% (κ = 0.81), 87.0% (κ = 0.84) and 89.8% (κ = 0.87) for the MassCode assay, Luminex assay/MassCode extract and Luminex assay/fresh extract, respectively. Most discrepancies compared with the conclusion based on microbiology and all PCRs performed were finding no organism in reference-positive samples, or identifying one or more organisms in reference-negative samples (Table 26 and Figures 27–29). Falsely identifying additional organisms in reference-positive samples occurred relatively rarely. Tests of symmetry for the number of main pathogens identified (all p < 0.0001) suggested that the MassCode assay and the Luminex assay on fresh/Leeds extracts (i.e. assays where extracts were recent) tended to systematically underestimate the number of main pathogens in each sample, whereas the Luminex assay on MassCode/Oxford extracts tended to systematically overestimate the number of main pathogens in each sample.

FIGURE 27.

Agreement between MassCode assay and combined microbiology plus PCR for the four main pathogens. Note: size of circle proportional to number of isolates. White indicates either identical pathogen(s) identified, or only additional pathogens were identified in one assay. Black indicates incorrect pathogens identified.

FIGURE 28.

Agreement between Luminex assay using MassCode (Oxford) extracts and combined microbiology plus PCR for the four main pathogens. Note: size of circle proportional to number of isolates. White indicates either identical pathogen(s) identified, or only additional pathogens were identified in one assay. Black indicates incorrect pathogens identified.

FIGURE 29.

Agreement between Luminex assay using fresh (Leeds) extracts and combined microbiology plus PCR for the four main pathogens. Note: size of circle proportional to number of isolates. White indicates either identical pathogen(s) identified, or only additional pathogens were identified in one assay. Black indicates incorrect pathogens identified.

Although the Luminex assay on freshly extracted samples was most accurate overall (89.8%), nevertheless it would have concluded that 5.2% of samples did not represent infectious diarrhoea caused by the four main pathogens, even though these pathogens were genuinely present in the sample.

Common nine pathogens

Of the 948 samples, based on combined microbiology and PCR results, 631 (66.6%), 62 (6.5%) and 2 (0.2%) had one, two or three of the pathogens tested by both the MassCode and the Luminex assays identified. There were 22 different combinations of pathogens (not shown).

Overall, agreement with the combined result from microbiology and all PCRs performed as the standard reference was 64.7% (κ = 0.58), 82.9% (κ = 0.79) and 86.8% (κ = 0.83) for the MassCode assay, Luminex assay/MassCode extract and Luminex assay/fresh extract, respectively. Most discrepancies compared with the conclusion based on microbiology and all PCRs performed were missing any organism in reference-positive samples, or identifying organisms in reference-negative samples (see Table 27 and Figures 30–32). In contrast to the four main pathogens, tests of symmetry for the number of common pathogens identified (all p < 0.0001) suggested that while the Luminex assay on fresh/Leeds extracts still tended to systematically underestimate the number of main pathogens in each sample, the MassCode and Luminex assay on MassCode/Oxford extracts tended to systematically overestimate the number of common pathogens in each sample (as a consequence of their higher false-positive rates).

FIGURE 30.

Agreement between MassCode assay and combined microbiology plus PCR for the nine common pathogens. Note: size of circle proportional to number of isolates. White indicates either identical pathogen(s) identified, or only additional pathogens were identified in one assay. Black indicates incorrect pathogens identified.

FIGURE 31.

Agreement between Luminex assay using MassCode (Oxford) extracts and combined microbiology plus PCR for the nine common pathogens. Note: size of circle proportional to number of isolates. White indicates either identical pathogen(s) identified, or only additional pathogens were identified in one assay. Black indicates incorrect pathogens identified.

FIGURE 32.

Agreement between Luminex assay using fresh (Leeds) extracts and combined microbiology plus PCR for the nine common pathogens. Note: size of circle proportional to number of isolates. White indicates either identical pathogen(s) identified, or only additional pathogens were identified in one assay. Black indicates incorrect pathogens identified.

Although the Luminex assay on freshly extracted samples remained most accurate overall (86.8%, compared with 89.8% for the main four pathogens), nevertheless it would still have concluded that 5.5% of samples did not represent infectious diarrhoea caused by these common pathogens, even though at least one was genuinely present in the sample.

Comparing Luminex assay with different extracts

Overall, agreement between the two Luminex assays on the common nine pathogens was only 81.1% (κ = 0.76), with most discrepancies because of extra organisms being identified using the Luminex assay with the MassCode/Oxford extract compared with no organisms being identified with the fresh/Leeds extract [76 (9.2%) samples] or in addition to other organisms identified in both [68 (8.2%) samples] (Figure 33).

FIGURE 33.

Agreement between Luminex assay using MassCode (Oxford) vs. fresh (Leeds) extracts for the nine common pathogens. Note: size of circle proportional to number of isolates. White indicates either identical pathogen(s) identified, or only additional pathogens were identified in one assay. Black indicates incorrect pathogens identified.

Considering all organisms present in the Luminex assay, with the original MassCode/Oxford extracts 581 (62.0%) samples would have been declared to have a single organism present, and 127 (13.6%), 25 (2.7%) and 2 (0.2%) would have been declared to have two, three or four organisms, respectively. With the fresh (Leeds) extracts, only 516 (61.5%), 42 (5.0%), 2 (0.2%) and 1 (0.1%) would have been declared to have one, two, three or four organisms present, respectively.

Aims and research questions

The main aims of the health economic component of this study were to evaluate MassCode multiplex PCR in terms of net health-care costs and utilisation of isolation resources, and to determine whether or not the use of these tests can improve the hospital management of patients with suspected infectious diarrhoea. The specific objectives were, therefore, to:

1. identify current practice for the management of patients with suspected infectious diarrhoea, in order to determine the likely impact of these new testing practices on factors such as isolation procedures and test turnaround times

2. estimate the cost-effectiveness of MassCode relative to current diagnostic methods and the net cost to the NHS of the new technology.

For the first objective, surveys were planned to map current infection control, and microbiologist and laboratory practice across England, in terms of how patients with suspected infectious diarrhoea are managed. With advances in molecular and genomic testing on the horizon in infection control and microbiology departments, it is important to understand how these departments currently operate, so that the implications of more widespread genetic and genomic testing can be assessed. Three surveys were, therefore, designed to provide data on how typical the study centres were in relation to the rest of the country, making it possible to generalise our findings across the wider NHS. This information would also be useful in terms of developing a base-case scenario within the health economic evaluations.

For the second objective, our planned analytical approach (as specified in the study protocol) was to conduct a cost-effectiveness analysis of the MassCode technology relative to other diagnostic tests, combining information on diagnostic procedures and short-term consequences collected within the phase 2 study with longer-term survival payoffs modelled from the study and a structured literature review. However, as the project was stopped before reaching the phase 2 study, the clinical and laboratory evidence required for the planned health economic analyses was not produced. Therefore, economic evaluations of the technologies under consideration could not be conducted, and we are unable to report any findings for this objective.

This chapter presents the results of the three initial surveys, summarising current infection control, and microbiologist and laboratory practice in terms of how patients with suspected infectious diarrhoea across England are managed. With advances in genetic and genomic testing on the horizon in this context, it is important to understand how infection control departments currently manage this patient group, so that the implications of more widespread genetic and genomic testing can be assessed for both patients and hospital trusts. These surveys provide answers pertaining to this information, which can be incorporated into future health economic evaluations of genetic and genomic interventions in infection control.

Survey design

To inform the design of the surveys, current infection control, and microbiologist and laboratory practice within the Oxford University Hospitals NHS Trust with respect to the management of patients with suspected infectious diarrhoea, was mapped between November 2010 and January 2011. Several approaches were used. First, a number of locations across this NHS trust were visited in order to observe and interview key infection control team members to gain an insight into current practice. These locations included infection control departments, laboratories and multiple wards at both the John Radcliffe Hospital and the Horton General Hospital (including the Emergency Departments, the Medical Assessment Unit, the Medical Short Stay Unit and the Surgical Emergency Unit). To supplement the information gathered during these visits, all relevant SOPs were examined, including infection control (for C. difficile: isolation, management of patients with diarrhoea and vomiting, and standard precautions) and microbiology SOPs (for C. difficile toxins and faeces: culture, see Figure 1 for examples).

The information gathered within this mapping exercise was used to design three surveys to be completed by NHS trusts across England. In all three surveys, it was specified that respondents should only consider infection control practice in relation to adult patients, not children. Copies of the surveys are supplied in Appendices 2–4 of this report.

Survey 1 (‘infection control survey’) was designed to be completed by infection control managers, collecting information on the infection control team, the monitoring of patients with infectious diarrhoea, infection control training and practice, and the management of outbreaks. Participants were also asked to consider how two potential future scenarios might impact on the management of patients with suspected infectious diarrhoea.

Scenario 1 was:

A consolidation of microbiology laboratory services has been proposed. The current model, with smaller microbiology laboratories based in hospitals and serving particular trusts may be replaced by a model which requires samples to be sent for testing to a small number of regional microbiology centres spread at regular intervals throughout the UK, each serving multiple trusts.

Scenario 2 was:

This survey is part of a larger study investigating the feasibility of introducing a new diagnostic test in microbiology laboratories across the UK. Using a stool sample taken from a patient with suspected infectious diarrhoea, this test can accurately detect 30 pathogens in a single reaction and rule in or out an infectious causative agent within 24 hours of the stool sample being taken.

Survey 2 (‘laboratory manager survey’) was designed to be completed by laboratory staff, collecting information on the testing process (e.g. which factors and patient characteristics drive testing decisions) and the cost of current testing practice in this context, for four types of test (PCR, ELISA, microscopy, culture).

Survey 3 (‘microbiologist survey’) was designed to be completed by microbiologists, collecting information on commonly requested tests, standard treatment practice for patients with positive C. difficile, Campylobacter spp. or Salmonella spp. tests, and patient management following discharge. Participants were also asked to consider the same two potential future scenarios as in the infection control survey.

Conducting the survey

The surveys were piloted in three trusts during February 2012: Leeds Teaching Hospitals NHS Trust, Brighton and Sussex University Hospitals NHS Trust and St George’s Healthcare NHS Trust, London. Following some minor formatting changes, the final versions of the surveys were sent to a sample of 51 acute NHS trusts (around one-third of all acute NHS trusts) in May 2012. A sample of all NHS trusts was used because of time and resource restrictions. Trusts were categorised by size (small/medium/large or teaching) and a weighted random sample was chosen to reflect the number of trusts of each type across the country: 10 small trusts, 18 medium trusts and 23 large or teaching trusts were contacted.

Four staff contact points in each trust were identified before the survey began [Director of Infection, Prevention and Control (DIPC), the senior infection control nurse (ICN), the lead microbiologist and the microbiology laboratory manager]. The DIPC was initially contacted by e-mail, which contained an introduction letter and examples of the questionnaires. Staff were offered a £20 Amazon voucher as an incentive for completion.

If approval was declined, the trust was not contacted again. When approval was provided (or no response from the DIPC was received) the remaining three staff were initially e-mailed, followed up by a telephone call. If no response was received, a second e-mail was sent in June 2012, again followed up by a telephone call. If no response was received again, respondents were contacted for a third time in September 2012. A final attempt to recruit respondents was made in January 2013.

Infection control survey results

Tables 28 and 29 report the characteristics of the 17 NHS trusts which completed the infection control survey. NHS trusts of all sizes and regions completed this survey (with the exception of London), although there were proportionally more responses from small and medium trusts. The average number of whole-time equivalent (WTE) infection control staff per trust is around seven (range 2.5–16.0), and most of these are grade 6 staff.

Infection control staff training and practice

Table 32 provides information on the training of infection control staff. A ‘standard precautions policy’ operates in 94% of trusts and median compliance with this policy is 80%. All trusts undertake terminal cleans of side rooms or bed spaces in bays when these are vacated by patients with suspected or confirmed infectious diarrhoea. Two-thirds of trusts would undertake a terminal clean of a whole bay in the same situation. All trusts would change the curtains in a side room in this situation, while 88% would change the curtains in a bed space in a bay and and 63% would change the curtains in a whole bay. Glove use increases in 88% of trusts in cases of suspected or confirmed infectious diarrhoea, and in 29% of trusts the cleaning policy varies depending on whether a diagnosis of infectious diarrhoea is suspected or confirmed. In 94% of trusts, the cleaning policy extends to cover other locations that an affected patient may visit, but only 6% of trusts carry out routine environmental testing.

TABLE 32 - Infection control staff training

n = 17 unless otherwise indicated.

n = 13.

n = 16.

Multiple answers possible.

Relevant reasons for prioritisation were ranked from 1 (top priority) to 5 (lowest priority).

Variable	Category	Valuea
Percentage of trusts with a ‘standard precautions policy’		94.12
Median compliance with standard precautions: percentage (width of the interquartile range)		80.00 (15.00)b
Percentage of trusts in which a terminal clean is undertaken when a patient with suspected or confirmed infectious diarrhoea vacates a space within a ward, by type of spaced	Side room	100.00
	Bed space in bay	100.00
	Whole bay	68.75c
Percentage of trusts in which curtains are changed as part of this terminal clean, by type of spaced	Side room	100.00c
	Bed space in bay	88.24
	Whole bay	62.50c
Percentage of trusts in which glove use increases in cases of suspected or confirmed infectious diarrhoea		88.24
Percentage of trusts in which the cleaning policy varies depending on whether a diagnosis of infectious diarrhoea is suspected or confirmed		29.41
Percentage of trusts in which the cleaning policy is extended to cover other locations in the hospital which the affected patient may visit		94.12
Percentage of trusts that carry out routine environmental testing		5.88
Percentage of trusts with a policy for the ‘management of patients with diarrhoea and vomiting’		94.12
Percentage of trusts with an ‘isolation policy’		94.12
Median percentage of patients with suspected infectious diarrhoea isolated in a side room (width of the interquartile range)c		99.00 (10.00)
Percentage of trusts in which all patients (with suspected infectious diarrhoea) are isolated in a side roomc		43.75
When insufficient side rooms are available to manage multiple patients with suspected infectious diarrhoea, how are patients prioritised? (Mean rankinge)	The most severely ill patients are prioritised	2.33
	Older patients are prioritised	3.50
	Particular pathogens or strains are prioritised	1.00
	Patients who have been sick for longer are prioritised	4.00
	Other	1.60
Median length of time (hours) for symptomatic patients to be isolated in a side room or cohorted in a closed bay width of the (interquartile range)		2.00 (2.60)
Mean number of bed moves a typical patient with suspected or confirmed infectious diarrhoea will make across the entire duration of their inpatient stay (SD)		2.12 (0.86)
Percentage of trusts with a ‘Clostridium difficile policy’		100.00
Percentage of trusts in which staff receive training on the management of patients with potentially infectious diarrhoea, over and above that which is provided in standard operating procedures and policies		94.12

All trusts have a ‘Clostridium difficile policy’, while 94% of trusts have a policy for the ‘management of patients with diarrhoea and vomiting’. The same percentage have an ‘isolation policy’. A total of 16 trusts provided information on the circumstances that would lead to the isolation of a patient with suspected infectious diarrhoea. In 10 trusts, respondents said that all patients with diarrhoea were isolated. Three trusts had criteria based on the frequency and type of stool as classified by the Bristol Stool Chart (‘type 5 stool or above on more than one occasion’; ‘A risk assessment is completed for all cases of type 5–7 stool and a pathway assigned following assessment’; ‘After three or more type 5, 6 or 7 stools in 24 hours’). The other three trusts provided differing criteria (‘prioritised for single room isolation in all cases’; ‘if we have an outbreak or clinically suspect norovirus or C. difficile infection’; ‘if infective diarrhoea suspected, i.e. no other cause’).

The median percentage of patients with suspected infectious diarrhoea who are isolated in a side room is 99%, with 44% of all trusts isolating all such patients in a side room (see Table 32). When insufficient side rooms are available to manage multiple patients with suspected infectious diarrhoea, most trusts prioritise patients by isolating those with particular pathogens/strains, or the most severely ill (see Table 32).

The median length of time it takes to isolate a patient is 2 hours, with patients making two bed moves, on average, during their inpatient stay. In 94% of trusts, staff receive training on the management of patients with potentially infectious diarrhoea over and above that which is provided in SOPs and policies. This extra training can take a number of different forms: Table 33 provides further details.

TABLE 33 - Extra training for staff on the management of patients with potentially infectious diarrhoea: what is provided?

Trust	Form of extra training
1	Included in mandatory clinical updates
2	Induction and mandatory training face-to-face and e-learning, study days, drop-in sessions
3	Clostridium difficile drop in sessions/ad-hoc training on wards
4	Currently face-to-face as part of annual infection control update. E-learning being developed
5	Ward training based on audit findings and following issues of concern or cases of infection
6	Via Infection Prevention and Control link practitioner scheme and ward ad-hoc training
7	Specifically part of clinical mandatory training. There is also a commode cleaning objectively structured clinical examination (OSCE) and additional study days and ‘road shows’
8	‘Getting Ready for Winter – Think Norovirus’ campaign, each winter
9	Covered in annual updates specifically addressing the need for isolation and personal protective equipment
10	Covered in mandatory training and induction. Also ad-hoc by ICN if situation arises where knowledge needs updating
11	Mandatory training and link networker sessions pre-winter season
12	Ad-hoc ward-based training. ICN attendance at ward training days. Infection control link days
13	Ward-based training sessions

See Figure 34 for information on how infection control policy documents are distributed to staff. These documents are predominantly made available via local intranet and targeted training sessions.

FIGURE 34.

Methods by which policy documents are made available to staff.

Microbiologist survey results

Table 36 reports the characteristics of the six NHS trusts who completed the microbiologist survey. No small NHS trusts completed this survey and only a partial geographic spread was achieved.

Trusts were asked a number of questions about stool sample testing, focusing on tests for seven pathogens (C. difficile, Shigella spp., E. coli O157, Campylobacter spp., norovirus, Salmonella spp. and Cryptosporidium spp.). Table 37 and Figures 35 and 36 detail trust responses to these questions. Pathogens are specified on stool test requests only once every five requests, usually either norovirus or C. difficile. The patient characteristics that influence testing decisions vary by pathogen. Patient age is said to be important for C. difficile testing (probably reflecting Department of Health guidance on mandatory testing), while length of stay is noted as being important for Shigella spp., E. coli, Campylobacter spp. and Salmonella spp. testing (probably reflecting the fact that these pathogens are typically acquired in the community and bought into hospital, rather than being acquired in the hospital). Symptoms/clinical details are important for E. coli, norovirus and Cryptosporidium spp. testing, other patient diagnoses in same ward/hospital are important for Salmonella spp. testing, and the immunocompromised status of a patient is important for Cryptosporidium spp. testing. As expected, the average length of time between taking a stool sample and receiving test results was around 1 day for C. difficile and norovirus, but around 2 days for Cryptosporidium spp. and Campylobacter spp., and 2.5 days for Shigella spp., E. coli and Salmonella spp.

FIGURE 35.

Which patient characteristics influence the decision to test for particular pathogens?

FIGURE 36.

What actions are taken if specific pathogens are not identified in a patient sample?

Trusts were also asked about the predicted sensitivity of tests for the seven pathogens, but most stated that predicted sensitivities were unknown. The exceptions were C. difficile (predicted sensitivity 90.50%, SD 0.00, n = 6), norovirus (predicted sensitivity 90.50%, SD 0.00, n = 5) and Cryptosporidium spp. (predicted sensitivity 80.50%, SD 20.00, n = 4). For each of these three pathogens, most trusts (four, five and three, respectively) stated that, because sensitivity was known (to be reasonably high), a negative result was sufficient to rule out an infection. [A reviewer queried the wording of this question: if the sensitivity of a test is low, then practitioners will be concerned that just having a negative result does not rule out infection. This is more than the negative predictive value, because if there are many more negatives than positive, then the negative predictive value may be high. But where the consequences of a missed positive are very severe (e.g. death), practitioners will still be extremely worried that, in an individual patient, a negative could still be an important missed positive because of the low sensitivity.] Most infection control staff remove patients from isolation in the event of a negative result. Antibiotics are stopped in most cases of C. difficile and norovirus, whereas only one trust noted that additional cleaning was stopped following negative results from all seven pathogens.

Around 5% of C. difficile cases were estimated to fall into a potential outbreak situation (Table 38). Strain typing information is requested in 96% of such cases, falling to 66% in cases that fall outside a potential outbreak situation. Strain typing information is used to inform clinical management in a number of ways, including:

• to identify potentially linked cases and map the extent of outbreaks

• to detect evidence of cross-transmission over longer periods

• to demonstrate to commissioners absence of outbreak.

Table 39 summarises treatment practice for suspected or confirmed infectious diarrhoea. Two-thirds of trusts routinely empirically treat patients following a positive C. difficile test. In the majority of cases, first-line treatment is metronidazole and second-line treatment is vancomycin. Standard practice differs on whether treatment is started before samples are sent for testing or once a causative pathogen has been identified. No trusts routinely empirically treat patients with positive Campylobacter spp. or Salmonella spp. tests, focusing any treatment on only the most severe cases. About 47% of all patients complete antibiotic therapy in hospital. Of those discharged before treatment has been completed, 10% will be readmitted within 14 days of discharge. In 27% of cases, causative pathogens are identified after discharge. In these circumstances, all trusts inform the clinical team, advising them to contact the patients’s GP or the patient directly. However, only 40% of trusts ever follow-up patients with diagnoses of infectious diarrhoea in primary care.

Laboratory survey results

Only three trusts responded to the laboratory survey. Furthermore, these three surveys all contained incomplete information: only three questions were answered by all three trusts. The level of missing information was such that we determined that no insight would be gained from presenting this incomplete information in this report.

Questions common to multiple surveys

Two questions were common to both the infection control and microbiologist surveys: trusts were asked to consider how two potential future scenarios might impact on the management of patients with suspected infectious diarrhoea. Table 40 presents the responses to scenario 1, which asked respondents to consider the impact of a consolidation of microbiology laboratory services. This scenario is being discussed, but has not currently been implemented. Typically, at present, each large NHS trust runs its own service microbiology laboratory; smaller NHS trusts would send specimens to the closest large trust microbiology laboratory. Centralising services further might provide economies of scale, enabling more rapid uptake of new technologies such as the MassCode assay, and also making them more cost-efficient, particularly if such ‘super-laboratories’ ran 24 hours a day. The responses from trusts indicated a number of concerns. These included:

• increased length of time to receive test results (because of greater transportation costs, loss of specimens);

• greater transmission and more frequent outbreaks;

• slower decision-making (leading to delayed treatment, bed-blocking, and delayed discharges);

• loss of local epidemiology data and responsiveness to local needs.

Successful implementation was said to require well-defined processes and good communication systems, and could lead to increased consistency in infection control advice offered.

Table 41 presents the responses to scenario 2, which asked respondents to consider the impact of a new diagnostic test which could detect 30 pathogens in a single reaction. Most trusts were positive about the consequences of this scenario. A number of potential benefits were identified, which included:

• more informed and faster decisions regarding the need for isolation and de-isolation;

• more effective use of limited side room space and reduced bed-blocking;

• improved patient treatment outcomes;

• earlier identification of outbreaks and implementation of cohorting.

Concerns were, however, raised about the need for such tests to be accurate and the requirement for samples to be taken as simply as possible.

Sample prep and bead bashing

1. For each extraction, add 1 ml of Roche stool transport and recovery (STAR) buffer to a tube of MPBiomedicals Lysing matrix E.

2. Add 1 µl of MS2 ssRNA (10⁵ copies) to each sample.

3. For each sample add 200 µl or one 10 µl loop full of stool sample to a tube.

4. Add 100 µl of chloroform to each tube.

5. Bead bash for 40 s at 6.0 m/s on the MPBiomedicals Fastprep 24 then store sample at 4 °C for 5 minutes.

6. Repeat bead bash for 40 s at 6.0 m/s on the MPBBio Fastprep 24.

7. Centrifuge samples for 1 minute at 1000 × g

8. Transfer at least 500 µl of supernatant to a labelled sterile 2 ml microcentrifuge tube (suitable for use in Qiasymphony).

Symphony loading

Waste Draw:

1. Affix waste bag to waste draw undercarriage.

2. Insert tip station, tip chute and waste vessels into waste draw ensuring that the tip station is fully inserted then close the waste draw.

3. On touchscreen select ‘scan now’ or ‘scan later’.

   Elution Draw:

4. Place black elution rack onto stainless carrier and place on slot 1 (cooled) in the elution draw.

5. On touchscreen select the rack and press ‘rack id’ button.

6. Scan the barcode on the Qiagen elution plate and place in elution rack in correct orientation and close the eluate draw.

7. On touchscreen press next and the Qiasymphony will briefly lock the draws and hood while scanning the elution plate. Once scanning is complete and the machine is unlocked, proceed to the next step.

   Reagents and Consumables Draw:

8. Remove the magnetic beads container from the reagents cassette and vortex for 3 minutes or until beads are evenly suspended and reinsert into reagent cassette.

9. Lift the reagent cartridges out of the white plastic stand and place into a grey cassette rack.

10. If the reagent cassette is being used for the first time then proceed to steps 11–15. If the cassette has been used before then go to steps 16–19.

11. Place a piercing lid on top of the cassette with the curved side of the lid on the opposite side of the cassette to the magnetic beads cartridge. Do not press down on the piercing lid or otherwise pierce the top of the reagent cartridges as the QiaSymphony will do this automatically.

12. Remove the seal from the magnetic beads trough ensuring that all parts of the seal are fully detached.

13. Using some clean tissue paper wipe away all droplets of the magnetic bead solution from around the rim of the bead container.

14. Slide the white enzyme rack onto the side of the reagents cassette and unscrew the proteinase K lids (the lids can be placed in the lid holders between the proteinase K rack and the cassette rack).

15. Proceed to step 20.

16. If the reagents cassette has been used before then remove the lids from the individual reagent cartridges and from the magnetic beads cartridge.

17. Using some clean tissue paper wipe away all droplets of the magnetic bead solution around the rim of the bead container.

18. Unscrew the proteinase K lids (the lids can be placed in the lid holders between the proteinase K rack and the cassette rack).

19. Proceed to step 20.

20. Load black 1500 µl and blue 200 µl tips, 8 rod covers and sample prep cartridges into reagents and consumables draw. For one run consisting of 24 samples the QIAsymphony will require 85 1500 µl tips, 24 200 µl tips, 3 8-rod covers and 18 sample prep cartridges. These quantities do not have to be exact but are the minimum amounts required.

21. On touchscreen press ‘bottle id’ button and scan ATL buffer bottle barcode and insert bottle into rear of reagents and consumables draw.

22. Load the reagents cassette that has been setup as described in steps 10–17.

23. Close the reagents and consumables draw and on touchscreen press ‘scan now’.

24. While the Qiasymphony is scanning, proceed to step 25.

    Samples Draw:

25. Make up carrier RNA in 2 ml microcentrifuge tubes suitable for the Qiasymphony using AVE buffer and carrier RNA stock. The amount required for any number of samples (up to 24) can be found in Table 47.

26. Unscrew the lids of the carrier RNA microcentrifuge tubes and place into a sample rack ensuring that the insert type in the rack matches the tube type. For 2 ml microcentrifuge tubes the insert has a red number above the barcode.

27. Unscrew the lids of the sample tubes and place them into a sample rack ensuring that the insert type in the rack matches the tube type. For 2 ml microcentrifuge tubes the insert has a red number above the barcode.

28. Only when the Qiasymphony has finished the scan will the sample drawer be unlocked. Once open, slide the samples rack up to the line in one of the first four slots and wait for the barcode scanner to align. Wait for the slot light to flash and then slide the rack past the barcode scanner until it clicks into place at the back of the draw. If correctly inserted the light will change from green to yellow. Do not place samples in the slot marked ‘A’ as this is for carrier RNA or internal controls only.

29. Slide the carrier RNA rack up to the line in the fourth slot and wait for the barcode scanner to align. Wait for the slot light to flash and then slide the rack past the barcode scanner until it clicks into place at the back of the draw. If correctly inserted the light will change to yellow.

30. Gently close the samples draw.

Setting up the run:

1. On touchscreen select the samples display, press the ‘batch’ button and then press the ID button. As barcodes are not used for these samples, a unique identifier must be manually assigned to each tube in the samples rack (e.g. 1, 2, 3, 4, etc.). To do this click on the first sample and press the ‘edit ID’ button and then click ‘1’ and then ‘enter’. Repeat this for all samples in the run and then press ‘next’.

2. To assign the correct assay to the samples press ‘select all’ then find the ‘Pathogen’ folder and select the program ‘virus pathogen complex 400 DSP’ from the list. Press ‘next’.

3. To select the elution volume press on the elution rack in slot 1 and then press ‘60’.

4. Press the ‘Queue’ button.

5. Go back to the samples menu and press the carrier button. Select all carrier tubes and then assign the carrier status to them (‘virus pathogen complex 400 DSP’).

6. Press the ‘Run’ button.

7. If prompted to scan any draw again press ‘scan’ and the run will start automatically after completion.