ABSTRACT
The incidence of tick-borne infections in the United States has risen significantly in the past decade. Ticks can transmit a variety of pathogens, including bacteria, protozoa, and viruses, that can cause serious illnesses. Therefore, the use of rapid, sensitive, and specific multiplex tests is important to identify the pathogen(s) in the acute phase and determine appropriate treatment to minimize the severity of the disease. The purpose of this study was to evaluate ChromaCode’s research use only (RUO) nine-target high-definition PCR (HDPCR) tick-borne pathogen (TBP) panel using 379 retrospective, remnant whole-blood and synovial fluid specimens previously submitted to Associated Regional and University Pathologists (ARUP) Laboratories and tested by clinically validated real-time PCR assays for Ehrlichia spp., Anaplasma phagocytophilum, Babesia spp., or Lyme Borrelia spp. The performance characteristics evaluated included positive percent agreement (PPA) and negative percent agreement (NPA) with the ARUP laboratory-developed tests (LDTs). All tested targets had an initial PPA greater than 97.0%, except Ehrlichia ewingii, with a PPA of 88.9%. The NPAs for all targets were between 98.8% and 100%. The TBP panel detected three coinfections, with two of Babesia microti and A. phagocytophilum and one of B. microti and E. chaffeensis, which were confirmed by the LDTs. There were 16 samples with discordant results compared to the LDT results, five of which were resolved by repeat testing on the TBP panel and bidirectional sequencing. Following discrepant resolution, the final PPA and NPA for the TBP panel were 97.7% (95% confidence interval [CI], 95.2% to 99.0%) and 99.6% (95% CI, 99.3% to 99.8%), respectively, with an overall agreement of 99.5% (95% CI, 99.2% to 99.7%) with the LDTs.
INTRODUCTION
Tick-borne illnesses, including Lyme disease, human granulocytic anaplasmosis, babesiosis, human monocytic ehrlichiosis, and relapsing fever, are the most common tick-borne diseases in the United States and have continued to rise over the last decade (1–3). Clinical presentations of tick-borne infections can range from mild to life-threatening, with symptoms including fever, headaches, myalgia, arthralgia, nausea, and vomiting, often overlapping in the early stages of disease. Most tick-borne pathogens are difficult to culture in the laboratory; thus, diagnosis has been based primarily on clinical presentation, history of exposure in areas of endemicity, microscopic examination of blood smears, and serological tests (https://www.cdc.gov/ticks/tickbornediseases/index.html) (1, 4). Though serologic testing may support laboratory evidence of tick-borne disease, it is limited by decreased sensitivity in the acute phases of disease and poor clinical specificity (4, 5). Nucleic acid amplification tests (NAAT) offer the advantages of directly detecting these pathogens during early infection. Real-time PCR tests for tick-borne diseases are available through the Centers of Disease Control and Prevention (CDC), state health laboratories, and certain reference laboratories, but these vary in sensitivity and specificity and are limited to singleplex assays or those that detect three or four targets only (6–10). This highlights an unmet need for a multiplex syndromic panel for accurate identification of these tick-borne disease agents. A comprehensive multiplex panel that targets a broader array of tick-borne pathogens will be necessary for the early detection and effective management of disease.
The purpose of this study was to evaluate ChromaCode’s research use only (RUO) high-definition PCR (HDPCR) tick-borne pathogen (TBP) panel (Carlsbad, CA) using whole-blood and synovial fluid specimens compared to the Associated Regional and University Pathologists (ARUP) laboratory-developed tests (LDTs) currently used for clinical testing. The TBP panel is a multiplex, 4-color channel PCR assay which allows for the simultaneous detection of nine tick-borne pathogens in a single well by endpoint signal intensity. The TBP panel detects Anaplasma phagocytophilum, Ehrlichia chaffeensis, Ehrlichia ewingii, Ehrlichia muris eauclarensis, Borrelia miyamotoi, Borrelia group 1 (B. burgdorferi and B. mayonii), Borrelia group 2 (B. hermsii, B. parkeri, and B. turicatae), Babesia microti, and Rickettsia spp. A recent study by Buchan et al. describes a preliminary evaluation of the TBP panel for the identification of tick-borne pathogens in human clinical and simulated specimens (11). The study findings describe high specificity (>98%) and sensitivity (100%) for A. phagocytophilum, B. miyamotoi, and Rickettsia spp. among clinical specimens, in addition to 100% analytical sensitivity for all targets and a combined analytical specificity of 99.5% in simulated samples. The conclusions of this study focused on the potential utility and clinical impact of implementing the TBP panel; however, because it was a prospective study, a minimal number of positive clinical samples were evaluated. For a broader understanding of the performance of the assay, we tested a large set of well-characterized, clinical specimens archived at ARUP Laboratories that were positive for six of the nine targets in the TBP panel. Our retrospective study design evaluated the TBP panel to detect tick-borne pathogens of low incidence in a standard quantitative PCR (qPCR) instrument and compared the performance characteristics to those of LDTs. The results of this study demonstrate the potential value of the TBP panel in detecting common tick-borne pathogens in a simple, high-throughput, scalable assay that may be easily adopted in clinical laboratories.
MATERIALS AND METHODS
Clinical samples.A total of 371 retrospective, whole-blood samples that were archived at ARUP Laboratories and previously tested via laboratory-developed PCR tests for the detection of Ehrlichia spp., Anaplasma phagocytophilum, Babesia spp., and Lyme Borrelia spp. were enrolled in this study. Eight synovial fluid samples were included to evaluate the analytical performance of the Lyme Borrelia sp. target in the TBP panel. Specimens were deidentified under a study protocol approved by the University of Utah institutional review board (IRB; protocol 00042995). The results of the reference method were blinded prior to testing with the TBP panel.
DNA extraction.Nucleic acids were extracted from 200 μl of whole blood or synovial fluid using the chemagic magnetic separation module (MSM) I automated extraction platform (PerkinElmer, Waltham, MA), according to standard laboratory procedures. Ten microliters of internal control provided by ChromaCode was added to each of the samples prior to extraction at a concentration of approximately 103 copies/reaction. The internal control served as a control for both extraction efficiency and presence of PCR inhibitors. The samples were eluted in 50 μl of elution buffer.
Instrument characterization.All testing for this study was performed at ARUP Laboratories on a QuantStudio 12K Flex system (Thermo Fisher Scientific, Waltham, MA) using the fast 96-well block. Prior to TBP testing, an instrument characterization step was performed to equalize the instrument-specific noise profile on the QuantStudio 12K system using synthetic DNA provided in the TBP panel equalization kit, according to the manufacturer’s instructions for use (IFU). Briefly, four individual MicroAmp fast optical 96-well reaction plates of synthetic DNA template corresponding to the four individual fluorophore channels at known concentrations were mixed with HDPCR master mix in every well of a 96-well plate and run according to the manufacturer’s IFU. The results from each of these four runs were uploaded into ChromaCode Cloud, and a noise correction mask specific to the QuantStudio 12K instrument used in the study was generated by ChromaCode’s proprietary signal processing software analysis.
TBP panel design and testing.The TBP panel is a single-well, 4-channel assay that detects nine common tick-borne pathogens and also includes an internal control. The TBP panel has the following design: 6-carboxyfluorescein (FAM) channel, Borrelia group 1 (B. burgdorferi and B. mayonii), Ehrlichia chaffeensis, and Borrelia miyamotoi; ATTO532 channel, Rickettsia spp., Ehrlichia muris eauclarensis, and Anaplasma phagocytophilum; ROX channel, internal control; and ATT0647N channel, Borrelia group 2 (B. hermsii, B. parkeri, and B. turicatae), Babesia microti, and Ehrlichia ewingii. The specific genes targeted by the TBP panel are described by Buchan et al. (11). The TBP assay thermocycling parameters were as described in the manufacturer’s IFU, as follows: stage 1, initial denaturation for 1 min at 95°C; and stage 2, denaturation for 10 s 95°C and annealing for 60.0°C for 2 min for 65 cycles.
For the TBP testing in the study, 5 μl of extracted DNA from whole blood or synovial fluid was added to 15 μl of master mix containing primers, probes, and enzyme (all provided in TBP test kit) in 96-well fast plate. Four-plate calibrators provided in the TBP test kit run with each plate to set the levels for target classification. The results for each TBP test were analyzed in ChromaCode Cloud by uploading the raw data file (.xls file) from the study instrument to the study account in ChromaCode Cloud. A report of positive for a target, negative, or invalid result for each sample is generated. Positive percent agreement (PPA) and negative percent (NPA) agreement compared with the ARUP LDTs were calculated.
ARUP Laboratories real-time PCR assays for tick-borne pathogens.The comparator methods for the study were ARUP’s real-time PCR LDTs for Ehrlichia spp. and A. phagocytophilum, Babesia sp., and Lyme Borrelia sp. performed on the QuantStudio 12K Flex instrument (Thermo Fisher Scientific, Waltham, MA). The assay for Ehrlichia and Anaplasma spp. detects E. chaffeensis, E. muris-like pathogen, E. ewingii, and Ehrlichia canis (without differentiating E. ewingii and E. canis), as described by Harris et al. (7). The Babesia assay amplifies a 190-bp segment of the 18S rRNA of Babesia spp. with a probe specific for B. microti and a probe to detect other Babesia spp. (B. duncani, B. divergens, Babesia sp. strain MO-1, and Babesia sp. strain EU1), as described by Couturier et al. (12). For the Lyme Borrelia assay, primers and probes designed to amplify a 68-bp segment of the ospA gene were used. The primer sequences were GA*AAAAATATTTATTGGGA*ATAGGTCT for BOR-L3 and GGCTGCTAACATTTTGCTTACAT for BOR-E3, and the Borrelia probe sequence BOR-FAM1 was MGB-FAM-G*AGCCTTA*A*TA*GCA*TG-EDQ (G*indicates super-G modified base, A* indicates super-A modified base; MGB, minor groove binder; FAM, 6-carboxyfluorescein; and EDQ, Eclipse dark quencher [ELITech Group, Bothell, WA, USA]). The reaction was prepared by using a 5× GoTaq probe qPCR master mix and 4.5 mmol/liter MgCl2 (Promega, Madison, WI, USA) with the following amplification parameters: 50.0°C for 10 min, denaturation at 95.0°C for 2 min, and 50 cycles of 95.0°C for 5 s, 56.0°C for 20 s, and 76.0°C for 20 s. The ospA gene is conserved among the Lyme Borrelia species and can also detect Borrelia afzelii and Borrelia garinii.
Discrepant analysis.Samples with discrepant results initially underwent repeat testing on the TBP panel. Only dual-positive samples that repeated as dual positive with the TBP panel were tested on the LDT for Ehrlichia spp., Anaplasma spp., and Babesia spp. to determine whether the TBP panel detected a coinfection not originally detected by the LDT. The final call for discordant samples was made based on the results of a repeat TBP panel result and repeat LDT result. Those samples that could not be resolved by these two methods were further tested by PCR and bidirectional sequencing.
Discrepant resolution by PCR and bidirectional sequencing.Discrepant sample resolution was executed by PCR and bidirectional sequencing. The primer sequences used for amplification and bidirectional sequencing are proprietary and not included in the manuscript. Samples were amplified using AmpliTaq Gold 360 DNA polymerase (catalog no. 4l398823; Applied Biosystems). The amplification was performed for 40 cycles with initial denaturation for 10 min at 95°C, denaturation for 30 s at 95°C, annealing for 30 s at 50°C, extension for 1 min at 72°C, and a final extension for 7 min at 72°C. The PCR was performed using the Bio-Rad T100 thermal cycler. The amplification products were analyzed by 2% agarose gel electrophoresis, and DNA was sequenced by the Sanger method at Retrogen, Inc. (San Diego, CA). Sequencing analysis was performed using the KB Basecaller algorithm with a Phred Q20 score.
RESULTS
A total of 371 archived whole-blood samples and eight synovial fluid clinical samples that were submitted to ARUP Laboratories for PCR between 2014 and 2018 for the detection of Ehrlichia spp., Anaplasma spp., Babesia spp., or Borrelia spp. were tested using the TBP panel. These samples included 325 samples positive by PCR for any of E. chaffeensis, E. ewingii, E. muris-like, A. phagocytophilum, B. microti, or Lyme Borrelia spp. Fifty-three negative whole-blood samples were also included. Figure 1 shows the distribution of positive specimens included in the study across various U.S. states. The case incidence correlates with the areas where cases of A. phagocytophilum, E. chaffeensis, and B. microti have been previously reported (13). However, these may not necessarily be the state where the patient was infected. The majority of the A. phagocytophilum-positive samples tested (n = 78) were from Massachusetts (38%) and New Hampshire (27%), followed by Maine and Wisconsin (9%) (Fig. 1a). The cases of positive E. chaffeensis samples (n = 70) were distributed across 20 states, including Tennessee (17%), Indiana (14%), Missouri, and Kentucky (8.5%) (Fig. 1b). The majority of B. microti-positive samples (n = 124) were from New York (26%), Massachusetts (17%), Minnesota (15%), Maine (10%), and New Jersey (8%) (Fig. 1c).
(a to c) Map of the number of positive cases of A. phagocytophilum (a), E. chaffeensis (b), and B. microti (c) in the United States sent to ARUP laboratories for reference testing.
Table 1 shows the initial performance of the TBP panel in comparison to LDTs. The TBP panel call rate was 99.7% (378/379). One sample was excluded from the overall analysis due to an internal control failure causing an invalid result. All tested targets had a positive percent agreement (PPA) greater than 97.0%, except for E. ewingii (88.9%). All eight synovial fluid specimens tested positive for Borrelia group 1 (PPA, 100%; 95% confidence interval [CI], 59.8% to 100%). The PPA values for A. phagocytophilum, B. microti, and E. chaffeensis were 98.8% (95% CI, 92.6 to 99.9%), 97.7% (95% CI, 92.8 to 99.4%), and 97.4% (95% CI, 90.2 to 99.6%), respectively. None of the samples tested were positive for spotted fever Rickettsia spp., Borrelia group 2 (relapsing fever Borrelia), or B. miyamotoi. The negative percent agreement (NPA) values for all targets were between 99.3% and 100%, except for B. microti (98.8%) and Borrelia group 2 (98.9%). The total PPA and NPA values for the HDPCR TBP panel were 97.7% (301/308) and 99.5% (3,082/3,095), respectively, with an overall accuracy of 99.4% (95% CI, 99.1% to 99.6%) compared to the LDTs.
Performance of the TBP RUO assay using whole-blood or synovial fluid clinical samples (n = 378)a
There were 16 samples with 20 discrepant results compared to the LDTs for tick-borne infections in the initial analysis (Table 2). All 16 samples were retested on the TBP panel to confirm the initial TBP panel result. Of the 16 samples, seven samples had dual-positive results by the TBP panel or were positive for a second pathogen not originally detected by LDT. These samples were tested by the LDT for B. microti, E. chaffeensis, or A. phagocytophilum. Of the 7 samples tested for dual positivity, 2 samples (TBP_144 and TBP_179) were dual positive for B. microti and A. phagocytophilum, and one sample (TBP_032) was dual positive for B. microti and E. chaffeensis, which confirmed these codetections.
Discrepant analysis of dual-positive samples by the TBP assay
The remaining four dual-positive samples by the TBP panel (TBP_226, TBP_358, TBP_202, and TBP_367) were tested by PCR and bidirectional sequencing. Samples TBP_226 and TBP_358 were determined to be false positive for the second target, E. chaffeensis and Borrelia group 2, respectively, based on negative codetections by PCR and lack of amplification with bidirectional sequencing. However, samples TBP_202, positive for B. microti and A. phagocytophilum, and TBP_367, positive for E. chaffeensis and B. microti, were unresolved, as repeat TBP panel testing and LDT were negative for the second target, but bidirectional sequencing was positive.
Discrepant analysis of the remaining nine samples was performed by repeat testing on the TBP panel or with bidirectional sequence analysis alone (Table 3). Four samples were false negative for B. microti, and in all four, either Borrelia group 1 or Borrelia group 2 was detected in the initial TBP test. Two of these discrepant samples (TBP_029 and TBP_043) repeat tested as B. microti by the TBP panel, while the other two (TBP_218 and TBP_264) were negative for B. microti both by the TBP panel and bidirectional sequencing. These two samples were low positives for B. microti by LDT, suggesting differences in the limits of detection between the LDT and TBP panel.
Discrepant analysis for samples with incorrect amplification and step down
In the initial analyses, samples TBP_205 and TBP_363 were false negative for E. chaffeensis, with TBP_205 testing false positive for Borrelia group 1. Both samples tested as E. chaffeensis upon TBP panel repeat testing, suggesting PCR inhibition in the initial TBP panel run and/or incorrect assembly of the signal in channel 1 by the data analysis software. Sample TBP_193, which was positive for E. ewingii/E. canis by LDT, was not detected in the TBP assay nor by bidirectional sequencing. This suggested that the assay design is specific to E. ewingii and does not detect E. canis as demonstrated by the manufacturer in their exclusivity studies (14). Sample TBP_059 was determined to be a false positive for A. phagocytophilum on the initial TBP panel run and was not detected upon repeat testing. Last, sample TBP_176 was false negative for A. phagocytophilum and could not be resolved by repeat testing on the LDT or further analyzed due to sample depletion.
Following discrepant analyses and resolution, the PPA and NPA for the TBP panel were 97.7% (95% CI, 95.2% to 99.0%) and 99.6% (95% CI, 99.3% to 99.8%), respectively, compared to LDTs, with an overall agreement of 99.5% (95% CI, 99.2% to 99.7%).
DISCUSSION
In this study, we evaluated the performance of a novel HDPCR TBP panel for the detection of tick-borne pathogens in whole-blood and synovial fluid specimens. Our results show that the TBP panel shows good concordance with validated LDTs and is capable of simultaneous detection of common tick-borne pathogens in a single-well, multiplex panel. The scalable throughput of the system allows for testing of up to 92 samples in less than 3 h. Moreover, the user-friendly cloud-based ChromaCode software allows for an easy and rapid analysis of the results efficiently within 2 to 3 min. The HDPCR technology can be readily adopted on other standard qPCR instruments, enhancing their ability to multiplex with 4 to 6 channels. Our evaluation was performed using the 96-well fast block on the QuantStudio 12K system, while other groups have evaluated this assay on the ABI 7500 FastDx instrument (Thermo Fisher Scientific, Waltham, MA) (11), highlighting the ease of adopting this assay on existing qPCR platforms.
The discrepancies in the results between the TBP panel and LDTs may be attributed to sample degradation of the frozen whole-blood samples, well-to-well contamination, differences in the assay limit of detection, variant sequences of the targets being amplified, or to the inclusivity of strains used in the TBP panel design. Additionally, incorrect assembly of the signal amplification curves or weak signal amplification due to low-positive samples, PCR inhibition, or probe mismatch can create false-negative and false-positive results. In our study, we observed the majority of false positives with B. microti-positive samples that tested incorrectly either as Borrelia group 1 (n = 2) or Borrelia group 2 (n = 3). These samples likely were low-positive B. microti or samples with a PCR inhibition resulting in a lower amplification signal intensity level, thus classifying the PCR curves incorrectly with the software, creating a false-positive Borrelia group 1 or group 2 result. This is a potential issue in the diagnosis of B. microti, especially in patients with mild infection/low-level parasitemia, for whom additional testing may be required. Further evaluation of the TBP panel B. microti target and software analysis algorithm may be warranted for improved accuracy and specificity. Furthermore, the Babesia target in the TBP panel is inclusive to the B. microti species and does not cross-react with other species that cause human infections, including B. duncani in the western regions of the United States, B. divergens, and unnamed strains designated MO-1 and EU-1. Fourteen samples positive for Babesia species other than B. microti by LDT tested negative by the TBP panel (data not shown). Though B. microti is the most common species in the United States, the TBP panel will miss these less-common Babesia spp., and diagnosis by microscopic examination of blood smears will still be necessary. The clinical implications of the false-positive or false-negative results are important to consider. Treatment with doxycycline or tetracycline as first-line treatment is recommended for Lyme disease, ehrlichiosis, anaplasmosis, tick-borne relapsing fever, and Rocky Mountain spotted fever. Though a false-positive Lyme Borrelia result for an E. chaffeensis infection highlights the analytical discrepancies of the assay, it may not result in a change in treatment or have a low impact on clinical care. In contrast, a false-positive relapsing fever Borrelia result for a B. microti infection may have severe implications, as B. microti requires treatment with atovaquone plus azithromycin or with clindamycin plus quinine. This limitation of the analytical performance could result in missed diagnoses and lack of appropriate directed therapy for babesiosis.
Our study has several limitations. First, with a retrospective study design, we tested deidentified samples known to be previously positive for Ehrlichia spp., A. phagocytophilum, B. microti, or Lyme Borrelia spp. in a reference laboratory. The positivity rates of these targets are higher than what may be observed in a prospective study due to a sampling bias. This study set was enriched for these positive specimens to better evaluate the analytical performance of the assay. Since the detection of Lyme Borrelia DNA in blood is exceedingly rare and has limited diagnostic utility (15), we included a limited number of positive synovial specimens to evaluate the analytical performance of the Borrelia group 1 (Lyme Borrelia) target in the TBP panel. Second, the whole-blood and synovial fluid specimens underwent at least 1 to 2 freeze/thaw cycles before extraction and TBP testing, which could result in false negatives due to sample degradation. Third, no cases of B. miyamotoi, relapsing fever Borrelia spp. (B. hermsii, B. parkeri, and B. turicatae), and Rickettsia spp. were identified in our study, limiting the evaluation of these targets.
Despite these limitations, our study is able to provide useful preliminary data on the analytical performance of this novel multiplex tick-borne panel using clinical specimens at a reference laboratory. Overall, the TBP panel assay is a novel, user-friendly method for the detection of common tick-borne pathogens in clinical specimens. This assay when used in areas of high incidence of tick-borne illnesses could impact the early detection of tick-borne pathogens and the early administration of treatment, which may contribute to better outcomes.
ACKNOWLEDGMENTS
We thank ChromaCode, Inc. for providing the funding and reagents for the study.
D.R.H. is a consultant for ChromaCode.
FOOTNOTES
- Received 2 October 2019.
- Returned for modification 24 October 2019.
- Accepted 12 December 2019.
- Accepted manuscript posted online 18 December 2019.
- Copyright © 2020 Shakir et al.
This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license.