gms | German Medical Science

For the growing number of international participants we provide a brief discussion of the current results in an English version.

Despite the relatively low amounts of C. trachomatis and N. gonorrhoeae target organisms in the current set of QC samples, the availability of well-established commercial or in-house NAT-assays has led to a high portion of correct results.

The current set of QC samples contained two samples with almost identical amounts of C. trachomatis (~1x10⁴ IFU/mL; sample # 1515301 and sample # 1515304), one sample with ~1x10³ IFU/mL of C. trachomatis (# 1515302) and two samples with various amounts of N. gonorrhoeae organisms (~10⁵ CFU/mL in sample # 1515302 and ~10⁶ CFU/mL in sample # 1515303).

Despite relatively low amounts of C. trachomatis target cells in the positive sample 1515302, only 4 false-negative results were observed among the Chlamydia trachomatis-specific results, reported by the 209 participants. Among the N. gonorrhoeae-specific results, false-negative results were reported by only 3 of the 210 participants for samples # 1515302 and # 1515303, which contained N. gonorrhoeae target organisms in an amount of 1x10⁵ CFU/mL and 1x10⁶ CFU/mL respectively. Also false-positive results for the two GO-negative samples were reported by 10 participants.

Since the amount of target organisms in CT-positive samples # 1515301, # 1515302, # 1515304, and NG-positive samples # 1515302 and # 1515303 could not be considered as “extremely low”, false negative results should encourage the participants to review and optimize their CT- and GO specific NAT-based assays.

Inhibition controls were included by all of the 210 participants and no inhibitory events were reported.

Tables 4 to 7 (Attachment 1 [Attach. 1], p. 2-3) were included this time to enable a detailed evaluation of the C. trachomatis – and GO-specific NAT components of combined GO/CT test systems. In Tables 4 and 5 (Attachment 1 [Attach. 1], p. 2) only the C. trachomatis (CT) specific results and in the Tables 6 and 7 (Attachment 1 [Attach. 1], p. 3) only the Neisseria gonorrhoeae (GO) specific results are presented and evaluated statistically.

The current set of QC samples contained two positive samples: # 1515314 with ~1x10³ IFU/mL of C. trachomatis target organisms and sample # 1515312 with ~1x10⁴ IFU/mL of C. trachomatis target organisms. Samples # 1515311 and # 1515313 contained no target organisms but only human cells and E. coli cells.

As depicted in Tab. 2 (Attachment 1 [Attach. 1], p. 4), the reported results were generally correct for the two positive samples.

For the C. trachomatis-negative samples # 1515311 and # 1515313 containing only non-infectious human cells and E. coli, 4 false-positive results were observed among the 121 participants.

Assuming a sequential processing of the 4 individual samples of the current set, a contamination event of the “negative” sample 3 by target organism or PCR products of the positive sample “2” is obvious. So false positive results should encourage the affected participants to review and optimize their DNA extraction procedure and their CT-specific NAT-based test system. For sample # 1515312, results were classified as “questionable” by one participant. For questionable results, certificates are only issued when correct results are reported by the participant for the remaining 3 samples of RV 531.

This striking match of the current results with observations and accuracy rates in the last years can be considered as an evidence for a high reliability and consistency of the applied assays and overall sample processing.

Run controls were performed by all of the 121 participants and inhibition events were not observed this time. In this context, it should be noted, that we have not added putative inhibitory substances into the samples of the current distribution.

Overall, a very good diagnostic performance and no noticeable issues regarding sensitivity and specificity were observed for the C. trachomatis-specific NAT assays used by the 121 participants.

The current set of QC samples contained one sample with a relatively high amount of Bordetella pertussis (# 1515321; 1x10⁵ CFU/mL), one sample with an approximately tenfold lower number of Bordetella pertussis (# 1515323; 1x10⁴ CFU/mL), one sample with an approximately hundredfold lower number of Bordetella pertussis (# 1515322 with ~1x10³ CFU/mL), as well as one sample containing only non-infected human cells and Escherichia coli (# 1515324).

The availability of well-established commercial or in-house NAT-assays has led to a high portion of correct results. All of the 152 participants reported correct positive results for the sample # 1515321 (B. pertussis, 1x10⁵ CFU/mL). Sample # # 1515323, which contained ~1x10⁴ CFU/mL of Bordetella pertussis, was correctly tested by 145 of the 152 participants, but 7 of the participating laboratories observed negative results for B. pertussis DNA. The amount of 10⁴ CFU/mL of B. pertussis target organisms is significantly above the previously observed lower limit of detection for the corresponding PCR assays or test systems. False-negative or questionable results should therefore lead to re-evaluations of the assay sensitivity. Sample # 1515324 contained only E. coli. All but two participants correctly reported this sample as negative for Bordetella pertussis. The false-positivity issue is probably due to contamination events in the course of sample preparation or low specifity of the used PCR/NAT test system. For sample # 1515322 (10³ CFU/mL of B. pertussis) 26 false-negative results were observed. With an amount of 10³ CFU/mL of B. pertussis target organisms we obviously touched the lower limit of detection of appropriate test systems. Given the relatively small amount of target organisms in the sample # 1515322 reported results were not included in the assessment for the certificates. This is also characterized by the two gray shaded boxes in Tab. 2 (Attachment 1 [Attach. 1], p. 5).

For the detection of B. pertussis, most participants used self-developed (in-house) test systems with inhibition and/or positive controls. Therefore, 53 participating laboratories used IS481 insertion sequence, 9 the pertussis toxin coding gene and 2 ribosomal genes. Run controls were performed by 150 of 152 participants and no inhibition events were observed with the samples of the current distribution.

The current set of QC samples contained two samples with a relatively high amount of target organisms. Sample # 1515331 contained approximately 1x10⁵ CFU/ml of a Clarithromycin-susceptible Helicobacter pylori patient strain, and sample # 1515333 contained approximately 1x10⁴ CFU/ml of a Clarithromycin-resistant Helicobacter pylori.

The availability of well evaluated NAT-based assays and the relatively high amount of target organisms in the two Helicobacter pylori-positive samples (#1515331: ~1x10⁵ CFU/mL and # 1515333: ~1x10⁴ CFU/mL) led to positive predictive values of 100%. Also for the Helicobacter pylori-negative sample #1515332 correctly negative PCR/NAT-results were reported by all the participating laboratories.

Sample # 1515334 of the current distribution contained a culture suspension of Campylobacter jejuni (~1x10⁵ CFU/ml), which was correctly reported "negative" by all of the 45 participants. This indicates an overall good analytical specificity of the used PCR test systems.

As noted in the description of RV 533, clarithromycin resistance testing in the examined H. pylori isolates could be performed by participants on a voluntary basis. This molecular resistance testing is usually based on amplification and sequencing of characteristic regions within the H. pylori 23 S rDNA or the use of hybridization probes based qPCR assays. Results for clarithromycin resistance were reported by 39 of the 45 participants. With one exception, the results were correct.

As discussed previously, the challenge in NAT-based detection of EHEC/STEC is not the detection of small amounts of target organisms, but the sophisticated analysis and typing of different Shiga toxin genes and other putative pathogenic factors (such as the eae gene encoding intimin or the hlyA gene encoding enterohemolysin).

The current set of QC samples contained two samples positive for EHEC: # 1515342 (E. coli, 1x10⁴ CFU/mL, clinical isolate, stx₁-positive, stx₂-positive, eae-positive and hlyA-positive) and # 1515343 (E. coli, 1x10⁵ CFU/mL, clinical isolate, stx₁-negative, stx₂-positive, eae-positive and hlyA-positive). The other two EHEC-negative samples contained an ETEC strain (sample # 1515344; 1x10⁵ CFU/mL) and an eae- and hlyA-negative E. coli K12 strain (# 1515341).

All participants correctly reported negative results for sample # 1515341, containing only E. coli K12. The second EHEC/STEC-“negative” sample (#1515344), containing a significant amount of an ST-positive ETEC isolate was also reported PCR-negative by all but one participant. For the EHEC/STEC-positive samples # 1515342 and 1515343, the availability of well-established NAT-based assays and strategies for molecular differentiation resulted in consistently high accuracy rates. Sample # 1515342 was correctly reported positive by 128 of the 130 participants and 129 of the 130 participants detected the target organism in the EHEC/STEC-positive sample # 1515343 correctly.

As in most of the participating laboratories, a NAT-based detection of shiga toxin coding genes is used primarily as a culture confirmation test most future positive samples will contain relatively high amounts of target organisms. The focus will remain more on the analytical specificity of the used test systems and less on the lower detection limit obtained. Partial or complete shiga-toxin subtyping, eae-, and hlyA-detection techniques were performed by 112 of the 130 participating laboratories. With one exception, the reported results were correct. None of the participants observed significant inhibition of the NAT reaction.

Due to numerous requests, here a short note for our participants outside Europe: as this proficiency testing panel is designed for a specific and sensitive detection of B. burgdorferi sensu lato DNA, the positive samples do not necessarily contain suspensions of “prototype” isolates of B. burgdorferi sensu stricto and in many of the bi-annual rounds of our external quality assessment scheme (EQAS) also other B. burgdorferi genotypes or genospecies will be present in individual samples.

Short recapitulation: So far 21 different species belonging to the B. burgdorferi sensu lato complex were described, that naturally present genetic differences in commonly used target genes. To further address this heterogeneity, B. kurtenbachii, an only recently delineated B. burgdorferi s.l.species which so far was only described from ticks and hosts in the USA was included in the actual panel. For specificity testing, B. miyamotoi was included. This species was first described in 1994 from Japan, belongs to the relapsing fever group spirochetes but is transmitted by the same hard ticks as B. burgdorferi s.l.. B. miyamotoi meanwhile was detected from Ixodes ticks from the USA, Asia and Europe. In 2011 first human cases were described from Russia, later on from the USA, Japan and Europe. Symptoms comprise fever, chills, headache, muscleache, arthralgias and nausea. Diagnosis relies on stained blood smears and PCR. For therapy doxycyclin is recommended, in more severe cases ceftriaxone or penicillin G.

The current distribution of QC samples contained one sample with Borrelia bavariensis (# 1515352; ~1x10⁵ organisms/mL), one sample with Borrelia garinii OspA type 8 (sample # 1515353; ~1x10⁵ organisms/mL), one sample with Borrelia kurtenbachii (sample # 1515354; ~1x10⁴ organisms/mL) and one sample with a very low amount of Borrelia miyamotoi (sample # 1515351; ~1x10³ organisms/mL).

With the exception of seven false-negative results for sample # 1515354 (containing the lowest amount of Borrelia target organisms), one false-negative result for sample # 1515353 (1x10⁵ organisms/mL) and two false-negative and one questionable results for sample # 1515352 (1x10⁵ organisms/mL) all participants reported correct results for the three positive samples. The false-negative results should prompt re-evaluation of the assay sensitivity.

The B. burgdorferi sensu lato complex “negative” sample # 1515351 containing ~1x10³ organisms/mL of Borrelia miyamotoi was classified false-positive by 36 and “questionable” by four laboratories. Potentially, this is either due to cross-reactivity with this genetically very closely related spirochete or due to a contamination during sample preparation or analysis.

Approximately half of the participating laboratories used self-developed (in-house) tests with inhibition and/or positive controls. None of the participants noted significant inhibition of the NAT-reaction. There were also no significant differences in test performance between commercially available kits and in-house assays for the diagnostic detection of Borrelia burgdorferi by PCR/NAT techniques.

Due to numerous requests: this test is designed exclusively for the testing of NAT-based methods and protocols for direct detection of low amounts of Legionella pneumophila from appropriate clinical specimen (such as respiratory specimens for example). Individual samples may contain relatively small amounts of the corresponding target organism. For this reason, participation is promising only for diagnostic laboratories, which have established a highly sensitive and specific PCR-/NAT-based method for the detection of L. pneumophila DNA or who want to evaluate their method with the help of an external quality control.

In order to assess the analytical sensitivity of certain Legionella pneumophila-specific PCR assays, the current set of QC samples contained a kind of dilution series of Legionella pneumophila serogroup 2: sample # 1515363 (1x10⁵ CFU/ml), sample # 1515362 (1x10⁴ CFU/ml) and sample # 1515361 (1x10³ CFU/ml). Sample # 1515364 contained no target organisms but only human cells and E. coli cells.

The L. pneumophila-positive (~1x10⁵ and ~1x10⁴ CFU/mL) samples # 1515363 and #1515362 were correctly tested positive by 101 and 87 of the 105 participating laboratories, respectively. Sample # 1515364, which contained only E. coli, was classified as false-positive by 2 laboratories. As cross reaction is unlikely, accidental contamination during the process of sample preparation and analysis is most likely to be causative. Therefore, the workflow during the process should be re-checked. All but one participant included inhibition controls in their test systems. No significant inhibitions of the PCR-/NAT reactions were reported.

The current set of QC samples contained a kind of dilution series of Salmonella enterica serovar Typhi: sample # 1515374 contained ~5x10⁴ CFU/mL, sample # 1515373 contained ~5x10³ CFU/mL and sample # 1515372 contained ~5x10² CFU/mL. Sample # 1515371 contained no target organisms but only human cells and E. coli cells. All participants reported correct results for samples #1515371, #1515373 and #1515374. Sample #1515372, containing ~5x10² CFU/mL Salmonella enterica serovar Typhi was correctly identified as “positive” by 17 of the 22 participants. Reporting a false-negative result for this sample should prompt a thorough re-evaluation of the performance of the test system. Inhibitory events in the PCR-/NAT reaction were not detected by any of the participants.

The current set of QC samples contained a sample without the corresponding target organisms (# 1515382; only E. coli cells), two samples positive for L. monocytogenes (# 1515384 with ~5x10⁶ CFU/mL and # 1515381 with ~5x10⁵ CFU/mL) and one sample with Listeria ivanovii (# 1515383) as Listeria species other than L. monocytogenes. The Listeria monocytogenes containing samples (# 1515381 and # 1515384) were correctly reported positive by all participants. In addition, the “negative” E. coli containing sample # 1515382 was also identified as negative by all laboratories. Most of the participants used Listeria monocytogenes-specific assays, which is reflected by the high number of “false-negative” results for sample # 1515383, containing 1x10⁵ CFU/mL L. ivanovii. However, as noted in the report form, participants using L. monocytogenes-specific PCR/NAT assays may indicate the corresponding results by the accessory code number 71. In this case, (false) negative results for non-Listeria monocytogenes species do not negatively affect issuing the corresponding QC certificates. In sum, the current results indicate a remarkably high analytical sensitivity of the current L. monocytogenes-specific PCR assays.

The concept of this proficiency testing series is designed to determine the analytical sensitivity and specificity of NAT-based assays for the direct detection of MRSA DNA in typical clinical sample material. With the development and composition of the corresponding sample materials we want to mimic the situation of processing clinical samples like wound or nasal swabs, so the lyophilized samples usually contain low amounts of target organisms in a background of human cells and other components. It is therefore important to note that NAT assays designed mainly for MRSA culture confirmation purposes may fail due to the low number of MRSA organisms in individual samples of the QC set.

Sample # 1515394 of the current distribution contained a mixture of S. aureus (MSSA, PVL-negative, ~5x10³ CFU/mL) and a CoNS strain (S. epidermidis; mecA-positive, ~5x10³ CFU/mL).

Correct (negative) results were reported by 292 of the 318 participating laboratories. Most of the 8 participants who reported “questionable” for sample # 1515393 indicated the use of assay concepts for the independent detection of the mecA gene and a S. aureus species marker gene (where “questionable” is the expected and correct classification for this mixed sample). Some of the 18 participants who reported (false-) positive MRSA PCR-results listed the use of in-house or commercial assay concepts relying on the quantitative detection of the mecA and S. aureus target genes.

One sample of the current set (# 1515393) contained an oxacillin-sensible CoNS strain (S. epidermidis; mecA-negative, ~1x10³ CFU/mL). Correct (negative) results were reported by 317 of the 318 participating laboratories. Assuming a sequential processing of the 4 individual samples of the current set, a contamination event of the "negative" sample 3 by target organisms or PCR products of the positive sample “2” is obvious. The false positive result should encourage the affected participant to review and optimize his DNA extraction procedure and/or the MRSA specific NAT-based test system.

Sample # 1515392 contained a typical cMRSA or CA-MRSA isolate (MRSA, PVL-positive, spa:t 310; ~1x10⁴ CFU/mL) which tested positive with the MRSA-specific assays in 313 participating laboratories.

One sample of the current set (# 1515391) contained a relatively high number of an “atypical” methicillin-resistant S. aureus SCC mec TypV isolate (MRSA, PVL-negative, ~1x10⁵ CFU/mL). As expected, the latter organisms were not reliably detected by a number of in-house SCCmec-based assay concepts and they were also missed by some of the current commercial tests. Such isolates are admittedly rare and hence false-negative results were not counted in the course of issuing the certificates.

Overall, it should be noted that a pleasingly large proportion of participants reported correct PCR/NAT results for MRSA. This indicates excellent sample workup functioning of laboratory-specific prevention measures to avoid the risk of contamination and carry-over events.

Also, an optional molecular detection of putative pathogenicity factor PVL (Panton-Valentine Leukocidin) or its coding gene lukF/S-PV was inquired. Corresponding results were reported by 78 of the 318 participating laboratories and within the current distribution, the results for the molecular PVL testing were correct in all but one case. Additional information can be found at Linde and Lehn [2] or Witte et al. [3] A well evaluated protocol for the detection of PVL-positive PVL isolate can be found at Reischl et al. [4].

In addition, commercial real-time PCR assays reliably targeting PVL-genes in MRSA and MSSA isolates are available in the meantime (for example: r-biopharm and TIB Molbiol).

The concept of this proficiency testing series is designed to determine the analytical sensitivity and specificity of NAT-based assays for the direct detection of C. pneumoniae in typical (clinical) sample material. With the development and composition of the corresponding sample materials we intended to mimic the situation of processing typical clinical samples like BAL or other respiratory specimens. So the lyophilized samples usually contain low amounts of target organisms in a natural background of human cells and other components. As a consequence, diagnostic assays designed for C. pneumoniae antigen detection in clinical specimens or other serological assays will fail due to the low number of C. pneumoniae infected cells in individual samples of the QC set.

The current set of QC samples contained two samples positive for C. pneumoniae. Sample # 1515401 was spiked with ~5x10⁵ IFU/ml of C. pneumoniae, whereas sample # 1515402 contained an approximately ten fold lower amount of C. pneumoniae (~5x10⁴ IFU/ml). Only E. coli and human cells were present in sample # 1515403 and sample # 1515404.

As depicted in Tab. 2 (Attachment 1 , p. 13), with one exception all participants reported correct results for the positive sample # 1515401. 123 of the 127 participants also reported correct positive results for sample # 1515402. For the samples #1515403 and # 1515404, both containing only E. coli, only one participant reported a false-positive result. Overall, there were no noticeable problems with the current set of QC samples and a good overall correlation with the expected results was observed.

General note to our participants: the concept of this proficiency testing series is designed to determine the analytical sensitivity and specificity of NAT-based assays for the direct detection of M. pneumoniae in typical sample material. With the development and composition of the corresponding sample materials we aim to mimic the situation of processing typical clinical specimens like BAL or other respiratory materials. Therefore, the lyophilized samples may contain low amounts of target organisms in a natural background of human cells and other components typically present in patient specimens. As a consequence, diagnostic assays designed for M. pneumoniae antigen detection in clinical specimens or other serological assays will fail due to the low number of M. pneumoniae infected cells in individual samples of the RV 541 distributions.

The current set of QC samples contained two positive samples. A relatively high amount of M. pneumoniae (~5x10⁶ genome copies/mL) was present in sample # 1515413 and an approximately fivefold lower amount of M. pneumoniae (~1x10⁶ genome copies/mL) was present in sample # 1515411. Sample # 1515414 was designed to monitor assay specificity: it contained a considerable amount of M. genitalium (~10⁴ genome copies/mL) as a related species to the target organism. The set was completed by sample # 1515412, which contained only human cells and a considerable amount of E. coli organisms.

Similar to the result constellations observed with past distributions of our external quality assessment schemes for Mycoplasma pneumoniae PCR/NAT detection, the availability of well-established commercial or in-house PCR/NAT-assays has led to a high percentage of correct results. With the exception of one laboratory, all 139 participants correctly reported sample # 1515412 as negative. The Mycoplasma pneumoniae containing samples (#1515411, ~10⁶ genome copies/mL and # 1415413, ~5x10⁶ genome copies/mL) were correctly reported by 136 and 137 of the 139 participants, respectively. Sample # 1515414 contained M. genitalium (~10⁴ genome copies/mL), and was erroneously reported positive by five laboratories. This may indicate lacking species-specificity of the test systems and trigger further investigations.

General note to our participants: the concept of this proficiency testing series is designed to determine the analytical sensitivity and specificity of NAT-based assays for the direct detection of C. burnetii DNA and/or Bacillus anthracis DNA in typical sample material. With the development and composition of the corresponding sample materials we want to mimic the situation of processing typical clinical samples. So the lyophilized samples may contain low amounts of target organisms in a natural background of human cells and other components typically present in patient specimens.

The current set of QC samples contained two samples with different amounts of Coxiella burnetii organisms (~5x10⁴ genome copies/mL in sample # 1515423 and ~1x10⁶ genome copies/mL in sample # 1515422), one sample with ~1x10⁵ genome copies/mL of Bacillus anthracis (sample # 1515423) and one sample with ~1x10⁵ genome copies/mL of a Bacillus anthracis Pasteur Strain (sample # 1515421). Sample # 1515424 contained only human cells and a considerable amount of E. coli organisms.

For convenient data presentation and analysis, we decided to depict the PCR/NAT results for each target organisms within this combined EQAS scheme in two separate tables: please see Tables 2 and 3 (Attachment 1 [Attach. 1], p. 16) for the Coxiella burnetii-specific results and Tables 4 and 5 (Attachment 1 [Attach. 1], p. 17) for the Bacillus anthracis-specific results.

Coxiella burnetii: The relatively high amount (~10⁶ genome copies/mL) of C. burnetii organisms in sample # 1515422 was correctly reported by all participants, as well as the twenty-fold lower concentration of the pathogen in sample #1515423. The two “negative” samples (#1515424 contained only E. coli and #1515421 contained only B. anthracis) were correctly reported negative by all participants. Overall, there were no noticeable problems with the current set of QC samples and a good correlation with the expected results was observed.

Bacillus anthracis: The results for this newly introduced EQAS scheme are easily discussed. All 17 participants correctly reported a positive result for sample # 1515423 (~1x10⁵ genome copies/mL). The second “positive” sample # 1515421 contained ~1x10⁵ genome copies/mL of B. anthracis strain “Pasteur”. This particular strain is positive for the virulence plasmid pXO2 and the B. anthracis-specific markers rpoB and dhp61, but does not harbor “lethal and edema factor” encoding plasmid pXO1 and is therefore also negative for the commonly used pathogenicity marker pagA. In addition, all participants correctly reported negative results for the two “negative” samples # 1515424 (containing E. coli and human cells) and # 1515422 (containing ~10⁶ genome copies of C. burnetii in a suspension of human cells).

After this very successful round of external quality assessment, “standardized samples” are now available for colleagues who are interested in obtaining B. anthracis DNA positive material for assay validation purposes. Requests for backup samples should be addressed to the EQAS coordinator (U. Reischl).

General note to our participants: the concept of this proficiency testing series is designed to determine the analytical sensitivity and specificity of NAT-based assays for the direct detection of F. tularensis DNA in typical sample material. With the development and composition of the corresponding sample materials we want to mimic the situation of processing typical clinical samples. So the lyophilized samples may contain low amounts of target organisms in a natural background of human cells and other components typically present in patient specimens.

The current set of QC samples contained three positive samples: a high amount of Francisella tularensis holarctica (~1x10⁵ CFU/mL) was present in sample # 1515434, an approximately tenfold lower amount (~1x10⁴ CFU/mL) was present in sample # 1515432, and an approximately hundred fold lower amount (1x10³ CFU/mL) was present in sample # 1515433.

Similar to QC samples from past distributions, the positive sample # 1515434 (~1x10⁵ CFU/mL of Francisella tularensis holarctica) was correctly tested positive by all of the 17 participating laboratories. Even with pathogen amounts of ~1x10⁴ CFU/mL and ~1x10³ CFU/mL (samples #1515432 and #1515433), 16 and 15 out of 17 labs were able to detect Francisella DNA, respectively. As no false-positive result was observed for the "negative" sample # 1515431, it seems that the participating laboratories have implemented functional precautions measures to prevent deleterious contamination events. Overall, these results corroborate the lower limits of detection observed in our previous EQAS distributions. Although the number of participating laboratories is still not very high, the results of the present distribution indicate that the lower limit of detection is about or slightly below 10⁴ organisms/mL when using currently employed and well evaluated PCR/NAT-based assay concepts for the detection of F. tularensis DNA.

The concept of this novel EQA-panel for the detection of carbapenemase genes is designed exclusively for the testing of NAT-based methods and protocols for molecular resistance testing or the direct detection of carbapenemase genes from DNA preparations of Enterobacteriaceae culture isolates. Because of the methodologically challenging design of EQAs for the molecular resistance testing of the wide range of known carbapenemase coding genes in different bacteria, the panel is narrowed down to a small selection of the currently most common carbapenemase genes in Enterobacteriaceae: KPC, VIM, OXA-48 like genes, GES carbapenemases, NDM, IMP, and GIM.

As shown in Tab. 1 (Attachment 1 [Attach. 1], p. 19), the current set contained three samples with carbapenem-resistant Enterobacteriaceae: sample # 1515441 contained Klebsiella pneumoniae with NDM-1 and OXA-232 gene (~1x10⁷ genome copies/mL), sample # 1515442 contained an Klebsiella pneumoniae isolate with a KPC-3 gene (~1x10⁷ genome copies/mL) and sample # 1515444 contained a Serratia marcescens strain harbouring a VIM-1 gene (~1x10⁷ genome copies/mL). The fourth sample # 1515443 was designed as negative control – it contained only E. coli without carbapenemase genes.

All participants detected a carbapenemase gene in the NDM-1 and OXA-232 positive K. pneumoniae sample (# 1515441) as well as the KPC-3 positive K. pneumoniae sample (# 1515442) and thus all results were evaluated as correct. However, a considerable proportion of participants did not detect the presence of two carbapenemase genes in the sample # 1515441. Two participants reported false-negative results for bla_NDM and as much as 25 participants (46.3%) did not detect the OXA48 like gene (OXA232). It is known that some OXA48 like genes such as OXA181 and OXA232 are missed in certain commercial test systems. As these OXA48 variants are inreasingly imported into Europe any user should be fully aware if the molecular tests used in their laboratory is able to detect such OXA48 variants. One participant reported a false-positive result for VIM in addition to the correct KPC result in the KPC3 positive K. pneumoniae sample (# 1515442).

For sample # 1515444, containing a carbapenem-resistant S. marcescens, two false-negative results were reported. This definitively demands intensive trouble-shooting. It might be possible that the DNA yield after preparation from lyophilised EQA-samples is lower than from bacterial colonies. Additionally, two false-positive results were observed in the sample without target organisms (# 1515443), containing only human cell material and E. coli. In this case, the workflow in the laboratory should be re-assessed and effective mechanisms to prevent cross-contamination during sample preparation and analysis should be implemented. In-house NAT assays were used for the detection of carbapenemase coding genes by 14 of the 54 participating laboratories, while all others quoted the use of commercial test systems or kits on the result form. As these commercial test system were not specified by all of the participants, a detailed comparisons between commercial kits and the heterogeneous group of proprietary (in-house) test systems with respect to sensitivity, specificity or susceptibility to contamination events is not yet possible. Overall, a very good diagnostic performance for KPC and NDM was observed. However, the OXA48 variant OXA232 was missed by a high proportion of participants.

General note to our participants: the concept of this proficiency testing series is designed to determine the analytical sensitivity and specificity of NAT-based assays for the direct detection of C. difficile DNA in typical sample material. With the development and composition of the corresponding sample materials we want to mimic the situation of processing typical clinical samples. So the lyophilized samples may contain low amounts of target organisms in a natural background of human cells and other components typically present in patient specimens.

The current set of QC samples contained two samples with similar amounts (~1x10⁴ CFU/mL) of Clostridium difficile organisms (sample # 1515451 and in sample # 1515454). Samples # 1515452 and # 1515453 contained only human cells and a considerable amount of E. coli organisms.

The C. difficile containing samples # 1515451 and # 1515454 were correctly reported as “positive” by 87 and 85 of the 88 participating laboratories, respectively. False-negative results should prompt a thorough evaluation of the test system and the workflow. The latter is definitely warranted for the two participants reporting false-positive results for sample # 1515453, containing only E. coli, but no target organism. Cross-reaction of the applied test system with E. coli DNA is unlikely, as the second “negative” sample # 1515452, which also contained E. coli DNA, was correctly reported “negative” by all participants. Therefore, cross-contamination during the process of sample preparation and analysis is most likely to be causative. All participants included controls to detect inhibitions of the PCR reaction. Significant inhibitory events were not reported.

General note to our participants: the concept of this proficiency testing series is designed to determine the analytical sensitivity and specificity of NAT-based assays for the direct detection of vancomycin-resistant enterococci DNA in typical sample material. With the development and composition of the corresponding sample materials we want to mimic the situation of processing typical clinical samples. So the lyophilized samples may contain low amounts of target organisms in a natural background of human cells and other components typically present in patient specimens.

Sample # 1515461 of the current set contained a relatively high amount of Enterococcus faecalis vanA (~1x10⁵ CFU/mL) and sample # 1515463 contained a similar amount of Enterococcus faecium vanB (~1x10⁵ CFU/mL). Sample # 1515464 contained Enterococcus faecalis (~1x10⁴ CFU/mL) and sample # 1515462 contained no target organisms but only human cells and E. coli cells. Of the 30 participating laboratories, 30 and 29 correctly reported positive results for the samples # 1515461 and 1515463, respectively. Of note, 29 participants reported dedicated vanA/vanB identification for these two samples, which were all correct. We were pleased to see that also for the “negative” samples #1515462 and # 1515464, all but one participant reported correct results. A false-positive result in the analysis of these samples should once again prompt evaluation of the test system and/or the workflow in the laboratory. This is especially important when considering the impact of molecular VRE detection on the clinical management of patient. All participants included controls to detect inhibitions of the PCR reaction. Significant inhibitory events were not reported.

General note to our participants: the concept of this proficiency testing series, which was started in 2013, is designed to determine the analytical sensitivity and specificity of NAT-based assays for the direct detection of P. jirovecii DNA in typical sample material. With the development and composition of the corresponding sample materials we want to mimic the situation of processing typical clinical samples. So the lyophilized samples may contain low amounts of target organisms in a natural background of human cells and other components typically present in patient specimens.

The latest set of QC samples contained two positive specimens (Tab. 1; see Attachment 1 [Attach. 1], p. 22). A relatively high amount of Pneumocystis jirovecii (~1x10⁶ organisms/mL) was present in sample # 1515602 and an approximately tenfold lower amount of Pneumocystis jirovecii (~5x10⁴ organisms/mL) was present in sample # 1515603. The set was completed by sample # 1515601 and sample # 1515604, which contained only human cells and a considerable amount of E. coli organisms.

Sample # 1515602, which contained the highest amount of P. jirovecii target organisms (~1x10⁶ organisms/mL) and sample # 1515604 with a twenty-fold lower concentration of P. jirovecii, were reported "positive" by 87 and 77 of the 85 participating laboratories, respectively. Although this could be due to a loss of template DNA during pre-analytical sample preparation procedures or other reasons, observation of false-negative results should give reason to check the diagnostic workflow, consider improving the sensitivity and/or analysing the species coverage of the individual assay concept. Interestingly, four laboratories reported a questionable result for sample # 1515601, containing only E. coli. Additionally, two laboratories also reported questionable results for the second “negative” sample # 1515604. This should definitely prompt investigations regarding all processes involved in sample preparation and analysis and the participants with false-negative results should be encouraged to optimize their NAT assays.

The yet limited number of participants in the trials RV 542 to RV 546 and RV 560, together with an incomplete reporting of the assays and manufacturers applied, do not allow a serious evaluation of the quality of either commercial tests or the very heterogeneous in-house PCR/NAT assays in regard to analytical sensitivity, analytical specificity, susceptibility to contamination, or simply the “overall performance”. In this regard we would like to encourage the participants to complete the protocol as good as possible.