Molecular Rapid Diagnostic Testing for Bloodstream Infections: Nanopore Targeted Sequencing with Pathogen-Specific Primers

doi:10.21203/rs.3.rs-3840855/v1

Download PDF

Research Article

Molecular Rapid Diagnostic Testing for Bloodstream Infections: Nanopore Targeted Sequencing with Pathogen-Specific Primers

https://doi.org/10.21203/rs.3.rs-3840855/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 01 Apr, 2024

Read the published version in Journal of Infection →

Version 1

posted

You are reading this latest preprint version

Background

Nanopore sequencing, known for real-time analysis, shows promise for rapid clinical infection diagnosis but lacks effective assays for bloodstream infections (BSIs).

Methods

We prospectively assessed a novel nanopore targeted sequencing (NTS) assay's performance in identifying pathogens and predicting antibiotic resistance in BSIs, analyzing 387 blood samples from December 2021 to April 2023.

Results

The positivity rate for NTS (69.5%, 269/387) nearly matches that of metagenomic next-generation sequencing (mNGS) (74.7%, 289/387; p = 0.128) and surpasses the positivity rate of conventional blood culture (BC) (33.9%, 131/387; p < 0.01). Frequent pathogens detected by NTS included Klebsiella pneumoniae (n = 54), Pseudomonas aeruginosa (n = 36), Escherichia coli (n = 36), Enterococcus faecium (n = 30), Acinetobacter baumannii(n = 26), Staphylococcus aureus(n = 23), and Human cytomegalovirus (n = 37). Against a composite BSI diagnostic standard, NTS demonstrated a sensitivity and specificity of 84.0% (95% CI 79.5%-87.7%) and 90.1% (95% CI 81.7%-88.5%), respectively. The concordance between NTS and mNGS results (the percentage of total cases where both either detected BSI-related pathogens or were both negative) was 90.2% (359/387), whereas the consistency between NTS and BC was only 60.2% (233/387). In 80.6% (50/62) of the samples with identical pathogens identified by both NTS tests and BCs, the genotypic resistance identified by NTS correlated with culture-confirmed phenotypic resistance. Using NTS, 95% of samples can be tested and analyzed in approximately 7 hours, allowing for early patient diagnosis.

Conclusions

NTS is rapid, sensitive, and efficient for detecting BSIs and drug-resistant genes, making it a potential preferred diagnostic tool for early infection identification in critically ill patients.

Bloodstream infections

Targeted nanopore sequencing

Metagenomic next-generation sequencing

Blood culture

Antibiotic Susceptibility testing

Bloodstream infections (BSIs) remain a major health concern, causing significant morbidity and mortality worldwide, especially in critically ill patients [1]. They often escalate to sepsis or septic shock, leading to multiple organ failure. For patients with sepsis, a delay in administering effective antibiotic therapy within the first 24 hours is likely associated with an 8% reduction in survival rate [2]. Prompt identification of pathogens and their drug resistance patterns is crucial for improving BSIs prognosis.

To date, blood culture (BC) remains the gold standard for diagnosing BSI, but it exhibits a notably low positive rate (20%-30%) and an indeterminate microbial growth time, ranging from 1 day to several months [3–5]. The polymerase chain reaction (PCR), including the Digital Droplet PCR (ddPCR) technique, which operates independently of microbial culture, has been proven to reduce the turnaround time for pathogen detection and significantly enhance the analytical sensitivity in identifying pathogens[6, 7]. However, it comes with unavoidable limitations, including a limited target range, susceptibility to nucleic acid contamination, and the potential for amplification bias, all of which significantly impact its accuracy in pathogen identification [8].

The detection of various pathogens is significantly enhanced by metagenomic next-generation sequencing (mNGS), offering the potential to identify nearly every microbe in a sample with a single test. However, the majority of these tests currently employ short-read sequencing workflows based on Illumina or BGI sequencing, resulting in a turnaround time of 24 hours or even longer from sample processing to result reporting [9]. Nanopore sequencing, functioning through the detection of electrical signals when DNA/RNA passes through a nanopore protein, allows the inclusion of bacteria and fungi with marker genes of different sizes in the same sequencing library [10]. Despite the Nanopore metagenomic sequencing based on a real-time analytical pathway reducing the detection time to less than 6 hours, it still faces challenges such as insufficient sensitivity, high sequencing errors, and elevated testing costs. Several studies have utilized the principle of amplicon sequencing to develop nanopore-targeted sequencing (NTS) methods for the rapid detection of dozens or even hundreds of pathogens, enhancing detection sensitivity [11, 12]. However, NTS methods commonly identify targeted bacteria and fungi by detecting 16S rRNA genes or internal transcribed spacer (ITS) gene sequences, with few designed for the detection of pathogens such as viruses and parasites [11, 13, 14].

We established an NTS assay pipeline utilizing pathogen-specific primer sets and multiplex PCR reaction systems (Fig. 1). The assay aims to identify 114 pathogens (57 bacteria, 21 viruses, 17 fungi, 12 special pathogens, and 7 parasites) associated with BSIs, along with 27 antibiotic resistance genes (ARGs) in blood samples within a few hours (the list of pathogens and ARGs in the Table S1). Based on the characteristics of pathogen genomes (e.g., genome length, gene mutation, sequence similarity to other organisms), each pathogen possesses 5–25 pairs of specific primers located at different positions in the genome.

In this study, we assessed the diagnostic performance of this NTS assay by comparing it to conventional BC and a plasma mNGS assay conducted in our clinical laboratory (see reference [15, 16])for pathogen identification and antibiotic resistance prediction in patients diagnosed with BSIs. Pathogen-specific PCR, multiplex ddPCR, were used, along with retrospective infectious disease consultations based on electronic chart reviews, to interpret discordant results.

Study Design

This was a prospective study to evaluate the performance of the novel NTS assay in diagnosing BSI. Participants were enrolled at the First Affiliated Hospital, Zhejiang University School of Medicine (FAHZU), a 5000-bed tertiary university hospital with a State Key Laboratory for Diagnosis and Treatment of Infectious Diseases located in southeastern China. The study was approved by the Ethics Committee of the Institutional Review Board of FAHZU (study no. IIT20220714A). Written informed consent was waived because of the non-interventional study design. Patients and samples were enrolled from December 2021 and April 2023.

Study Population

Eligible participants for this study comprised intensive care unit(ICU) patients (older than 18 years) of FAHZU, for whom diagnostic BCs and plasma mNGS tests were routinely ordered as part of suspected BSI or sepsis assessment (Fig. 2 ). Consistent with recent studies[17, 18], no specific criteria were employed to define suspected BSI; the decision to conduct BC and mNGS tests was at the discretion of the attending physician. Only patients with plasma mNGS tests within one week of the BC were included in this study. The remaining EDTA-anticoagulated whole blood samples after mNGS testing were reserved for NTS testing. NTS results were not available to physicians caring for study patients and did not affect clinical decision making. Electronic chart review was conducted to obtain clinical data for evaluating the significance of conflicting test results. Participants were excluded if any of the BC or NTS results were unavailable, or if electronic medical records were inaccessible (Fig. 2).

Blood culture and antimicrobial susceptibility testing

Blood samples (both aerobic and anaerobic bottles, 10–20 mL per bottle) were collected and incubated for a maximum of 5 days (BD BACTEC FX; BD Biosciences). The pathogens in positive cultures were identified using matrix-assisted laser desorption ionization-time of flight mass spectrometry (bioMérieux, Marcy-l’Étoile, France). Antibiotic susceptibility testing (AST) were carried out using an automated Vitek 2 Compact system (bioMérieux, France), as well as the disk diffusion method and/or gradient diffusion method, in accordance with the guidelines established by the Clinical and Laboratory Standards Institute (M100-ED30)[19].

Plasma mNGS assay

The plasma mNGS tests were performed on illumina sequencing platform in our clinical laboratory as described in our recent studies[15, 16]. Pathogens and their reads were sent to the clinicians.

NTS assay

Workflow

As depicted in Fig. 1, the NTS workflow encompassed sample processing, library preparation, sequencing, and real-time analysis. Initially, glass beads were introduced to 500 uL whole blood samples for cell disruption via vortex mixing. After prompt centrifugation, 200 uL of the supernatant was employed for automated extraction.

Subsequently, amplicon libraries were created for nanopore sequencing using the Bloodstream Infection Pathogen Nucleic Acid Detection Kit (MD EasyDiagnosis, Wuhan, China). This involved multiplex PCR amplification (13 cycles) of pathogen-specific target genes using 10 µL of extracted DNA, followed by purification with magnetic beads. The resulting supernatant was used in another PCR reaction (18 cycles) to introduce sample-specific indexes. The library concentration was measured with Qubit after PCR amplification and bead purification. The purified library, now incorporating individual-specific indexes (ranging from 10–30 ng), was then incubated for 10 minutes at 25°C with the ligation buffer and sequencing adapter (Oxford Nanopore Technologies).

Following bead purification, the ligation products were combined with Sequencing Buffer II (SBII) and Loading Beads (LBII) (Oxford Nanopore Technologies) to form the sequencing library. Flush Tether (FLT) was added to Flush Buffer (FB) to prepare the loading solution (Oxford Nanopore Technologies). Subsequently, the loading solution and the sequencing library were introduced into the R9 Spot-ON Flow Cell (Oxford Nanopore Technologies) loading and sample ports, respectively, for sequencing on a GridION X5 instrument (Oxford Nanopore Technologies).

Finally, the Nanopore sequencing results were processed using the Guppy software, which automatically categorized reads into "pass" and "fail" folders. Reads in the "pass" folder exhibited a mean Q-score exceeding a specific threshold (Q8), while reads shorter than 200 base pairs were excluded. The remaining reads were aligned to an in-house reference database using Minimap2 with the map-ont parameter. Reads with less than a 90% similarity to the target area were discarded.

Detection threshold: The following formula was utilized to calculate the confidence rating for microorganisms: c = (1- a×0.1^b)×100, where 'c' represents the confidence rating, 'a' is the number of target sequences designed for the respective microorganism in the panel, and 'b' is the number of target sequences detected. If 'c' exceeds 90%, the microorganism is considered to be detected. To minimize the risk of contamination, a true positive was defined as having reads per million (RPM) for specific microorganism sequences detected in clinical samples at least ten times higher than the RPM of the same microorganism's sequences in the negative quality control.

After the removal of human sequences, the data was aligned against the Comprehensive Antibiotic Research Database using the Minimap2 software (CARD). Sequences that didn't align, had multiple alignments, or had alignment error rates greater than 0.2 were eliminated. The remaining sequences represented antibiotic resistance genes found in the sample under investigation. The correlation between antibiotic resistance genes and the detected bacteria was assessed using the CARD database [20].

Definitions of Bloodstream Infection.

Similar to existing studies evaluating broad-spectrum sequencing technologies[21, 22], we employed a more comprehensive diagnostic standard combining pathogen detection with clinical adjudication (see Table 1), categorizing patients into definite BSI, probable BSI, possible BSI, or unlikely BSI. This approach allows for a more comprehensive evaluation of microbiological results, identifying a wider range of pathogens correlated with clinical symptoms of BSI, identifying BSIs that BCs fail to diagnose, such as EBV-induced hemophagocytic syndrome and Scrub typhus caused by Orientia tsutsugamushi.

For each patient, clinical data including demographics, complications, underlying diseases, sequential organ failure assessment (SOFA) scores, therapeutic interventions, clinical outcomes, and microbiological tests from all suspected infection sites collected within 7 days before and after enrollment are compiled for clinical adjudication. These definitions were applied by one ICU doctor (J.H) and two infectious disease doctors (X.Z, D.X), along with two microbiologists (D.H, and Y.C).

Table 1

Definitions of BSI in this study.
Category	Definitions	Examples	BC/mNGS/NTS Result Interpretation
Definite BSI	Pathogen detected by physician-ordered microbiologic tests within 7 days before and after enrollment is identified as responsible for causing BSIs based on clinical adjudication.	E.g., E. coli found in BC after an abdominal surgery, or T. marneffei detected through mNGS and/or NTS in the blood of a HIV patient.	True Positive (TP): the responsible pathogen is detected; False negative (FN): The result is negative; False Positive (FP): the detected pathogen is not related to BSI.
Probable BSI	Pathogen detected by physician-ordered microbiologic tests within 7 days before and after enrollment is the likely cause of BSI syndrome based on clinical, radiologic or laboratory findings.	This category primarily includes common organisms detected solely through molecular methods, where the possibility of colonization or DNA fragment contamination cannot be completely ruled out. E.g., in a patient clearly diagnosed with septic shock, if both mNGS and NTS detect P. aeruginosa, a common pathogen for septic shock. However, if all BCs taken within 7 days of enrollment do not detect the presence of P. aeruginosa, this finding requires careful interpretation.	True Positive (TP): the result includes pathogens that may cause BSI symptoms; False negative (FN): The result is negative; False Positive (FP): the detected pathogen is not related to patient clinical presentation.
Possible BSI	Pathogen detected by physician-ordered microbiologic tests within 7 days before and after enrollment has potential for pathogenicity and is consistent with clinical presentation, but alternate explanation is more likely.	This primarily includes clinically ambiguous viral positive events that require close integration with patient condition changes. E.g., in patients with hematopoietic stem cell transplants or solid organ transplants, the detection of HCMV or EBV in blood by mNGS or NTS necessitates treatment decisions closely based on the patient's condition and long-term monitoring.
Unlikely BSI	a. Pathogen detected by physician-ordered microbiologic tests within 7 days before and after enrollment has potential for pathogenicity but is not consistent with clinical presentation of BSI; b. Pathogen detected by physician-ordered microbiologic tests within 7 days before and after enrollment is negative and alternate explanation for clinical presentation is more likely.	E.g., after comprehensive medical history analysis, patients with unexplained fever are more likely to have tumor progression, autoimmune diseases, or reactions to treatment drugs (drug fever, etc.) rather than infections.	False Positive (FP): pathogen detected; True Negative (TN): Pathogen not detected.

Statistical analysis

The primary outcome was to validate the clinical performance (e.g., sensitivity and specificity) of NTS results by comparing them with a composite BSI diagnostic standard that integrates microbiological testing and clinical judgment for diagnosing BSIs. Secondary outcomes involve comparing NTS, mNGS, and BC for BSI diagnosis consistency and pathogen profiles, as well as predicting antibiotic resistance in common clinically significant microorganisms compared to BC-positive strains. Statistical analysis was conducted using IBM SPSS Statistics software (SPSS 25, SPSS Inc., Chicago, IL, USA). Continuous variables were expressed as the median and interquartile range (IQR), while categorical variables were reported as frequencies and percentages. The Chi-square test was employed to explore differences in positivity rates between NST, mNGS and BC, with differences considered statistically significant if P values were ≤ 0.05.

Patient Enrollment, Diagnostic Category and Microbiologic Results.

Samples and clinical data of 451 patients were initially collected. 14.2% (64/451) were excluded because of mNGS test or culture testing failures (n = 21), incomplete medical histories (n = 13), or because of insufficient remaining samples for NTS testing (n = 30) (Fig. 2). Samples from 387 patients were tested in the study. The median patient age was 58 years, 66.4% (257/387) were men. Among these patients, 55.0% (213/387) suffered from sepsis. The 28-day cumulative mortality rate was 17.6% (68/387) (Table 2).

Table 2

Demographic and clinical characteristics of patients included in this study.
Clinical characteristics	Enrolled patients (N = 387)
Age, years, [median (IQR)]	58 (44–69)
Gender, n (%)
Male	257 (66.4)
Female	130 (33.6)
Underlying Disease, n (%)
Solid-organ transplant	41 (10.6%)
hematopoietic stem cell transplant	22 (5.7%)
Leukemia without HSCT	34 (8.8%)
Lymphoma without HSCT	13 (3.4%)
Cancer	39 (10.1%)
Cardiovascular diseases	129 (33. 3%)
Trauma	79 (20.4%)
Autoimmune disease	65 (16.8%)
HIV	20 (5.2%)
Diabetes	9 (2.3%)
Surgery performed before 14 days of inclusion	21 (5.4%)
Clinical manifestations (n, %)
Sepsis	213 (55.0%)
Neurological (Headache, consciousness disorders, etc.)	49 (12.7%)
Gastrointestinal (Abdominal pain, diarrhea, etc.)	183 (47.3%)
Respiratory (Pneumonia, respiratory failure, etc.)	229 (59.2%)
Cardiovascular (chest pain, endocarditis, etc.)	142 (36.7%)
SOFA score, [median (IQR)]	6 (3, 8)
APACHE II score, [median (IQR)]	18 (13, 23)
Length of hospital stay, years, [median (IQR)]	19 (10–33)
28-day mortality, n (%)	68(17.6)

After diagnostic categorization, definite BSI accounted for 54% (209/387). Probable BSI, possible BSI, and unlikely BSI constituted 12.4% (48/387), 12.7% (49/387), and 20.9% (81/387), respectively (Fig. 2). A total of 70 organisms (39 bacteria, 17 fungi, and 14 viruses) were detected by NTS, mNGS, and BC for these patients (Table S2-S5). Among these, the most frequently detected pathogens were K. pneumoniae (n = 60), P. aeruginosa (n = 39), E. coli (n = 37), E. faecium (n = 34), S. aureus (n = 29), and Acinetobacter baumannii (n = 28) and Human cytomegalovirus (HCMV, n = 48) and Epstein-Barr virus (EBV, n = 31).

Comparison of NTS Results with mNGS and Blood Cultures

We found that NTS was positive in 69.5% (269/387) of samples, with identification of 352 organisms (42 species) [including 235 bacteria (24 species), 26 fungi (9 species) and 91 virus (9 species)] (Table S3, Fig. 2). This positivity rate was slightly lower compared to the mNGS results (74.7%, 289/387; p = 0.128), but significantly higher than the positivity rate achieved with BC (33.9%, 131/387; p < 0.01) (Fig. 2). The top five strains detected by NTS were K. pneumoniae (n = 54), P. aeruginosa (n = 36), E. coli (n = 36), HCMV (n = 37) and E. faecium (n = 30) (Table S3). In total, 88.9% (313 organisms, 40 species) of the 352 organisms in NTS were also detected by mNGS (Table S6), while only 41.8% (109 organisms, 16 species) of bacteria(n = 235) and fungi (n = 26) were grown in BCs (Table S7).

When compared with mNGS results, 90.2% (212/235) of bacteria, 88.5% (23/26) of fungi, and 85.7% (78/91) of viruses detected by NTS also found by mNGS (Fig. 3A). This included 87% (47/54) of K. pneumoniae, 100% (36/36) of P. aeruginosa, 88.9% (32/36) of E. coli, 93.3% (28/30) of E. faecium, 96.2% (25/26) of A. baumannii, 91.3% (21/23) of S. aureus, 90.9% (10/11) of T. marneffei, 86.5% (32/37) of HCMV and others. mNGS additionally detected 21 pathogen species (27 organisms) not found by NTS, of which 13 species including Enterobacter hormaechei, Aeromonas veronii, Enterobacter asburiae, Gemella morbillorum, Aspergillus oryzae, Rhizomucor miehei, Rhizopus oryzae, Rhizopus microsporus, Cunninghamella bertholletiae, Severe fever with thrombocytopenia syndrome virus (SFTSV), Hepatitis B virus (HBV), BK virus (BKV) and Human Immunodeficiency Virus 1 (HIV-1) were outside the detection range of NTS(Table S6, Fig. 3A). These differential results were validated through specific PCR in our clinical laboratory (data was not shown). Figure 3B illustrates that 39.9% (101/235) of bacteria and 25.0% (8/26) of fungi identified by NTS were confirmed by BCs. All the viruses detected by NTS were validated either by mNGS or laboratory PCR methods.

We observed that NTS detected multiple pathogens in 16.5% (64/387) of the samples (Table S3, Fig. 3C). Among these, 49 samples had two pathogens, and 13 samples had three and 2 had four pathogens. The proportion of samples with multiple pathogens detected by mNGS was similar, at 19.9% (77/387) (p = 0.264). In contrast, BCs only identified multiple microbial infections in four samples, three of which involved dual-pathogen infections: (1) a 25-year-old female with acute myeloid leukemia (M1) in the marrow suppression stage developed sepsis from Candida tropicalis and P. aeruginosa (Patient P169); (2) a 58-year-old female post kidney transplantation developed an intra-abdominal infection from E. coli and S. aureus (Patient P11321); (3) a 57-year-old male developed a mixed infection of K. pneumoniae and E. faecium post-acute hemorrhagic necrotizing pancreatitis surgery (Patient P10128). Additionally, a 70-year-old male with acute leukemia in the marrow suppression stage post chemotherapy developed septic shock from a mixed infection of K. pneumoniae, P. aeruginosa, and A. baumannii (Patient P6612). All the pathogens in these four cases, except for the E. coli strain in Patient P11321, were confirmed by both NTS and mNGS (Table S3, Table S4).

Performance of NTS testing for BSI Diagnosis

We first assessed the diagnostic consistency among NTS, mNGS, and BC in suspected BSI patients, that is, whether they provided the same test results for the diagnosis of the patients. During the analysis, we considered the sum of patients with definite, probable, and possible BSIs as the diagnosed BSI patients, and the unlikely BSIs as non-BSI patients. A true positive (TN) was defined as any method that detected a pathogen related to the clinical presentation of BSI; otherwise, it was considered negative, which includes no detected pathogen (FN) or only detection of organisms unrelated to BSI (FP) (as defined in Table 1 and Fig. 2). Based on this, we found that the diagnostic consistency between NTS and mNGS reached 90.2% [(248 + 101)/387], whereas the consistency between NTS and BC was only 60.2% [(113 + 120)/387] (Fig. 4A). Overall, NTS detected pathogens related to BSI in 257 patients, of which 111 were confirmed by both mNGS and BC, 137 cases (35.4%, 137/387) were confirmed by mNGS alone, and 2 by BC alone (Figs. 2 and 4B, Table S8). Only 7 cases (1 definite, 3 probable BSI, and 3 possible) were uniquely detected by NTS (detailed information provided in Table 3). We confirmed the pathogens in 6 of these cases using PCR methods (Table 3).

Subsequently, we applied the same criteria to define BSI-positive patients as those with a sum of definite, probable, and possible BSI, and identified non-BSI patients as those unlikely to have BSI, for comparison. We discovered that the accuracy of NTS (85.3%, 95% CI 81.4%-88.5%) was slightly lower than that of mNGS (90.4%, 95% CI 87.1%-93.0%) (p = 0.028), but significantly higher than BC (53.2%, 95% CI 48.3% − 58.1%) (p < 0.01). The sensitivity of NTS was 84.0% (95% CI 79.5%-87.7%), with a specificity of 90.1% (95% CI 81.7%-88.5%) (Fig. 4C).

Table 3

Clinical information on the 7 cases in which the suspected pathogen was identified by NTS alone.
ID	Patient	Disease Description	NTS Results	Verification Method and Results	Diagnosis Category
1	P603	A 61-year-old patient with Type 2 diabetes was admitted to the ICU for a 10-day period due to sepsis. During his stay, he required mechanical ventilation, exhibited pulmonary bullae, and had an elevated white blood cell count	A.baumannii	Plasma multiplex ddPCR: A.baumannii,2.3×10⁴ copy/ml; Sputum culture positive for A.baumannii and P.aeruginosa	Probable BSI
2	P2626	A 52-year-old man developed sepsis, accompanied by acute myocarditis and acute liver injury. His chest CT scan revealed the presence of pleural effusion and signs of a pulmonary infection	K.pneumoniae	Plasma multiplex ddPCR: K.pneumoniae, 1.7×10² copy/ml; Sputum culture positive for A.baumannii and P.aeruginosa.	Probable BSI
3	P4088	A 66-year-old man with open traumatic brain injury. He was in comatose and mechanical ventilation status	S.aureus	Plasma multiplex ddPCR: S.aureus, 5.4×10² copy/ml	Probable BSI
4	P5240	A 15-year-old girl, who was undergoing a hematopoietic stem cell transplantation, presented with an unexplained fever and was found to have severe anemia	HCMV	Whole blood HCMV quantitative PCR:1.3×10³ copy/ml	Possible BSI
5	P6041	A 73-year-old man with severe pneumonia following lung transplantation surgery	EB, HCMV	Whole blood HCMV quantitative PCR: 3.7×10⁴ copy/ml; Whole blood HCMV quantitative EBV:1.4×10² copy/ml	Possible BSI
6	P8022	A 68-year-old woman, who had been receiving long-term corticosteroid treatment for optic neuritis, presented with multiple lung shadows on imaging, a high fever, and signs of respiratory failure	T. marneffei	Bronchoalveolar lavage fluid culture showed growth of T. marneffei	Definite BSI
7	P9085	A 57-year-old man presenting with suspected infective endocarditis following cardiac valve disease surgery, accompanied by heart failure, acute pulmonary edema, and respiratory failure	K.pneumoniae	Plasma multiplex ddPCR: negative; multiple BCs were negative	Possible BSI

Performance of NTS testing for resistance gene detection

In our study, we correlated genotypic resistance detected by Next-Generation Sequencing (NTS) with phenotypic drug-susceptibility outcomes for 62 samples, revealing an 80.6% (50/62) concordance rate (Table S9). Our analysis included 52 Gram-negative strains, with NTS successfully identifying all Extended-spectrum β-lactamase (ESBL) genes in 15 phenotypically confirmed ESBL-producing bacteria (8 E. coli, 5 K. pneumoniae, 1 Klebsiella aerogenes, 1 P. aeruginosa), encompassing 9 bla_TEM, 3 bla_SHV, and 3 bla_CTX−M genes. However, among the carbapenem-resistant (CR) bacteria, only 54.2% (13/24) (7 K. pneumoniae, 3 A.baumannii, 2 P. aeruginosa and 1 E. coli) showed corresponding CR genes in NTS findings, comprising 8 bla_KPC, 3 bla_NDM and 3 bla_OXA−48 genes. This reduced detection rate in 11 CR Gram-negative bacteria could be attributed to lower sequencing depths (average reads: 1,052) compared to other CR isolates (average reads: 117,152). Notably, no resistance genes were observed in 10 drug-sensitive Gram-negative bacteria. Among the 13 Gram-positive bacteria studied, two multidrug-resistant isolates were identified: one Methicillin-resistant S. aureus (MRSA) and one Vancomycin-resistant E. faecium (VREF). NTS successfully detected mecA and vanA genes in these, while the remaining 11 drug-susceptible Gram-positive bacteria showed no resistance genes (Table S9).

Turnaround time (TAT) of the sequencing and analysis

Nanopore sequencing offers real-time analysis. We calculated the analysis time required for all samples in this study, starting from the initiation of sequencing to the delivery of results. The results revealed that 50% of the samples could generate results within 13.5 minutes, and 95% of the samples could generate results within 209 minutes (Fig. 5). In other words, when factoring in approximately 0.5 hours for sample processing and nucleic acid extraction (showed in Fig. 1), as well as 3 hours for library preparation, 95% of the samples could be tested and reported within 418 minutes, which is approximately 7 hours.

Due to the generally nonspecific signs and symptoms, the clinical management of BSIs in critically ill patients presents several challenges. The inability to promptly identify pathogens through BCs is a significant factor contributing to higher mortality rates in these patients[23]. This study introduced an innovative NTS approach, designed based on the epidemiological characteristics of global and local BSI pathogens. It demonstrated exceptional capability in diagnosing BSIs and detecting pathogens and their antibiotic resistance genes in nearly 400 patient samples. 95% of the samples produced results within 7 hours, highlighting the potential significance of this NTS assay in providing prompt diagnosis for critically ill patients.

The developed NTS approach significantly improved the detection rate of samples when compared to BCs (66.9 vs 33.1%, p < 0.01), approaching the unbiased mNGS assay (75.2%). However, a high-throughput sequencing technology with a broad detection range naturally detects organisms that are clinically insignificant (potentially contaminants or non-pathogenic organisms residing in the human body) [18, 24, 25]. By implementing reliable quality control measures and developing intelligent algorithms for contamination filtering, we can effectively reduce the interference of contaminants on the results [26–28]. In this study, we developed specific primers targeting different genomic locations for each pathogen (5–25 primer pairs per pathogen), rather than focusing on a single region, such as the 16S rRNA gene. This not only enhances the specificity of microbial identification at the species level but also allows for the analysis of the distribution of detected reads across the genome, aiding in the exclusion of potential contamination from microbial nucleic acid fragments introduced during reagent handling or other experimental steps. This greatly increased the specificity of the NTS approach in clinical applications, with only 15 cases (3.9%) of detected organisms unrelated to clinical symptoms considered false positives (Fig. 2).

Recent studies indicate that multiplex molecular detection can uncover 10–40% more positive cases than conventional BCs, but with no consensus yet on interpreting multiplex molecular method positive(+)/BC negative(−) results[15, 29]. We specifically focused on the clinical interpretation of inconsistent NTS positive results, a question that, to our knowledge, has not been addressed in the literature. Therefore, future wise decision-making must consider positive NTS results in the context of other factors, including clinical, epidemiological, and additional laboratory data, to determine whether they represent a true BSI event[6, 17, 22]. Molecular method+/BC − outcomes might be due to non-viable, non-proliferating bacteria, transient bacteremia/viremia, intracellular microbes, antibiotic effects, or contamination[17]. In our study, NTS+/BC- results accounted for 37.2% (144/387) of cases. Among these, 47.2% (68/144) were confirmed as definite BSI based on consistency with mNGS, other pathogen tests, and clinical symptoms, like rare Chlamydia psittaci and T. marneffei infections that usually be identified rapidly mainly through molecular methods. The rest were split into probable BSI (25%, 36/144) and possible BSI (27.8%, 40/144), verified only by molecular methods (mNGS, species specific PCR, ddPCR etc.), without culture confirmation or highly specific clinical signs. Especially for viral positive cases, for which there are no current clinical diagnostic guidelines, we often rely on the patient's immune status and clinical presentation at the onset to assess the likelihood of them being the causative pathogens. This strategy is commonly employed in current research literature[6, 17, 18]. Taken together, we found that NTS's sensitivity (84.0%) and specificity (90.1%) were both higher than BC, making it effective for early detection of suspected BSI cases missed by BCs.

Both our published research and studies from other teams have confirmed that existing mNGS testing procedures can aid in predicting the antibiotic resistance of common microorganisms in respiratory samples[30–32]. The NTS panel in this study includes 27 pathogenic genes related to common clinically significant pathogens. We observed that the genotypic resistance predicted by NTS was consistent with phenotypic resistance confirmed in 80.6%(50/62) of the samples with bacterial culture results. This suggests that NTS results can also rapidly provide potentially useful evidence in clinical scenarios where culturing and phenotypic resistance information cannot be obtained. However, it's worth noting that, in comparison to pathogen identification, the accuracy of high-throughput sequencing techniques in detecting and predicting antibiotic resistance genes is more challenging, involving several ongoing issues: (1) the comprehensiveness, of functional annotations, and the timeliness of updates of antibiotic resistance gene mutations in resistance databases[20]; (2) accurately associating genotypic resistance with phenotypic resistance in multi-microbial samples[33]; (3) establishing standardized bioinformatics analysis pipelines[34]; (4) improving sequencing depth to enhance sensitivity[35]. Currently, clinicians need to consider a patient's prior antibiotic usage history and their response to treatment to determine whether adjusting the treatment plan based on the resistance information provided by NTS tests.

This NTS testing, incorporating automated nucleic acid extraction, rapid library preparation, and real-time analysis systems, provides results for 95% of clinical patients within 7 hours. Moreover, in urgent situations, one clinical sample can be independently tested together with quality control samples. This endows the technology with the characteristics of Point-Of-Care Testing (POCT), but further cost optimization is still required. Without considering the potential effects of batch variations, this NTS assay is significantly faster than most of BC and mNGS tests used in clinical practice [9, 15, 36]. NTS offers real-time analysis that doesn't require waiting for the entire sequencing process to complete before beginning the analysis. For samples with a higher microbial load, NTS can even detect and report pathogens within the first 5 minutes of sequencing, thus providing crucial microbiological evidence for the timely and accurate treatment of critically ill patients.

Despite these insights, our research has limitations. Firstly, although the sensitivity and specificity of NTS were reasonable in our study, variations may occur in different clinical practices. Future research should be conducted to verify its performance stability across samples from multiple centers and diverse populations. Secondly, this version of NTS panel is unable to differentiate certain important subtypes of ARGs, such as the KPC gene. It would be more valuable if NTS could accurately distinguish between subtypes like KPC-2 and KPC-33, as they have different resistance profiles [37]. We will continue to optimize its detection capabilities of drug resistance genes in subsequent studies.

In summary, our research underscored the clinical value of the newly developed NTS method in the rapid and accurate detection of BSI-related pathogens and their resistance genes, providing an alternative option for the early diagnosis of infectious diseases in critically ill patients. Before its application in clinical diagnosis, it is necessary to conduct multi-center clinical utility evaluation studies to identify the optimal testing timing and the most suitable patient groups, thereby maximizing its potential clinical significance.

Ethical Approval

The study was approved by the Ethics Committee of the Institutional Review Board of FAHZU (study no. IIT20220714A). Written informed consent was waived because of the non-interventional study design.

Data availability

Data supporting the conclusions of this article are included within the article and its additional files. The datasets used and/or analyzed in the current study are available from the corresponding author on reasonable request. The pathogen reads of NTS tests were deposited in the National Center for Biotechnology Information under project PRJNA1061589, which are publicly accessible at http://www.ncbi.nlm.nih.gov/bioproject/1061589.

Acknowledgements

We thank all clinicians who provided detailed diagnostic and treatment data of patients for our study, as well as all infectious disease (ID) physicians and clinical microbiologists who received our infectious disease consultations. We also thank Dr. Ying Wang, Yan Zhang and Hairong Guo from Wuhan Easy Diagnosis Biomedicine Co., Ltd, who helped us to solve technical issues encountered in the experiments.

Funding

This study was supported by “Leading Geese” Research and Development Plan of Zhejiang Province (No. 2024C03218), Zhejiang Provincial Natural Science Foundation (grant number LY23H200001), Zhejiang Provincial Natural Science Foundation for Distinguished Young Scholar (grant number LR23H200002).

Author contributions

Study design, D.H., and Y.C.; Sample detection, F.Y., D.Z., J.W and S.Z; Data collection, J.Z., B.L., and S.Z.; Data analysis, D.H., J.H., X.Z., D.X. and Y.C.; Wrote the paper: D.H. All authors have read and approved the final version of the manuscript.

Competing interests

The authors declare no competing interests.

Costa SP, Carvalho CM. Burden of bacterial bloodstream infections and recent advances for diagnosis. PATHOG DIS 2022 2022-9-27;80(1).
Castillo DJ, Rifkin RF, Cowan DA, Potgieter M. The Healthy Human Blood Microbiome: Fact or Fiction? FRONT CELL INFECT MI 2019 2019-1-1;9:148.
Falconer K, Hammond R, Gillespie SH. Improving the recovery and detection of bloodstream pathogens from blood culture. J MED MICROBIOL 2020 2020-6-1;69(6):806–11.
Lamy B, Sundqvist M, Idelevich EA. Bloodstream infections – Standard and progress in pathogen diagnostics. CLIN MICROBIOL INFEC 2020 2020-1-1;26(2):142 – 50.
Tjandra KC, Ram-Mohan N, Abe R, et al. Diagnosis of Bloodstream Infections: An Evolution of Technologies towards Accurate and Rapid Identification and Antibiotic Susceptibility Testing. Antibiotics (Basel) 2022 2022-4-12;11(4).
Wu J, Tang B, Qiu Y, et al. Clinical validation of a multiplex droplet digital PCR for diagnosing suspected bloodstream infections in ICU practice: a promising diagnostic tool. CRIT CARE 2022 2022-8-8;26(1).
Pliakos EE, Andreatos N, Shehadeh F, Ziakas PD, Mylonakis E. The Cost-Effectiveness of Rapid Diagnostic Testing for the Diagnosis of Bloodstream Infections with or without Antimicrobial Stewardship. CLIN MICROBIOL REV 2018 2018-1-1;31(3).
Vardakas KZ, Anifantaki FI, Trigkidis KK, Falagas ME. Rapid molecular diagnostic tests in patients with bacteremia: evaluation of their impact on decision making and clinical outcomes. Eur J Clin Microbiol Infect Dis 2015 2015-11-1;34(11):2149–60.
Gu W, Deng X, Lee M, et al. Rapid pathogen detection by metagenomic next-generation sequencing of infected body fluids. NAT MED 2021 2020-11-9;27(1):115–24.
Petersen LM, Martin IW, Moschetti WE, Kershaw CM, Tsongalis GJ. Third-Generation Sequencing in the Clinical Laboratory: Exploring the Advantages and Challenges of Nanopore Sequencing. J CLIN MICROBIOL 2019 2019-12-23;58(1).
Fu Y, Chen Q, Xiong M, et al. Clinical Performance of Nanopore Targeted Sequencing for Diagnosing Infectious Diseases. Microbiol Spectr 2022 2022-4-27;10(2):e27022.
Wang M, Fu A, Hu B, et al. Nanopore Targeted Sequencing for the Accurate and Comprehensive Detection of SARS-CoV-2 and Other Respiratory Viruses. SMALL 2021 2021-8-1;17(32):e2104078.
D'Andreano S, Cusco A, Francino O. Rapid and real-time identification of fungi up to species level with long amplicon nanopore sequencing from clinical samples. Biol Methods Protoc 2021 2021-1-22;6(1):bpaa26.
Chen J, Xu F. Application of Nanopore Sequencing in the Diagnosis and Treatment of Pulmonary Infections. MOL DIAGN THER 2023;27(6):685–701.
Han D, Yu F, Zhang D, et al. The Real-World Clinical Impact of Plasma mNGS Testing: an Observational Study. Microbiology spectrum 2023 2023-1-1:e398322.
Han D, Yu F, Zhang D, et al. Applicability of Bronchoalveolar Lavage Fluid and Plasma mNGS Assays in the Diagnosis of Pneumonia. OPEN FORUM INFECT DI 2023 2023-12-13.
Nguyen MH, Clancy CJ, Pasculle AW, et al. Performance of the T2Bacteria Panel for Diagnosing Bloodstream Infections: A Diagnostic Accuracy Study. ANN INTERN MED 2019 2019-6-18;170(12):845–52.
Lee RA, Al Dhaheri F, Pollock NR, Sharma TS. Assessment of the clinical utility of plasma metagenomic next-generation sequencing in a pediatric hospital population. J CLIN MICROBIOL 2020 2020-1-1;58(7):e419-20.
CLSI M100-ED30:2020 Performance Standards for Antimicrobial Susceptibility Testing, 30th edn..
Alcock BP, Huynh W, Chalil R, et al. CARD 2023: expanded curation, support for machine learning, and resistome prediction at the Comprehensive Antibiotic Resistance Database. NUCLEIC ACIDS RES 2023 2023-1-6;51(D1):D690-9.
Jing C, Chen H, Liang Y, et al. Clinical Evaluation of an Improved Metagenomic Next-Generation Sequencing Test for the Diagnosis of Bloodstream Infections. CLIN CHEM 2021 2021-8-5;67(8):1133–43.
Blauwkamp TA, Thair S, Rosen MJ, et al. Analytical and clinical validation of a microbial cell-free DNA sequencing test for infectious disease. NAT MICROBIOL 2019 2019-2-11;4(4):663–74.
Gupta S, Sakhuja A, Kumar G, McGrath E, Nanchal RS, Kashani KB. Culture-Negative Severe Sepsis: Nationwide Trends and Outcomes. CHEST 2016 2016-12-1;150(6):1251–9.
Rodino KG, Toledano M, Norgan AP, et al. Retrospective Review of Clinical Utility of Shotgun Metagenomic Sequencing Testing of Cerebrospinal Fluid from a U.S. Tertiary Care Medical Center. J CLIN MICROBIOL 2020 2020-11-18;58(12).
Charalampous T, Kay GL, Richardson H, et al. Nanopore metagenomics enables rapid clinical diagnosis of bacterial lower respiratory infection. NAT BIOTECHNOL 2019;37(7):783–92.
Du J, Zhang J, Zhang D, et al. Background Filtering of Clinical Metagenomic Sequencing with a Library Concentration-Normalized Model. Microbiol Spectr 2022 2022-10-26;10(5):e177922.
Gu W, Miller S, Chiu CY. Clinical Metagenomic Next-Generation Sequencing for Pathogen Detection. Annual Review of Pathology: Mechanisms of Disease 2019 2019-1-24;14(1):319 – 38.
Han D, Li R, Shi J, Tan P, Zhang R, Li J. Liquid biopsy for infectious diseases: a focus on microbial cell-free DNA sequencing. THERANOSTICS 2020 2020-1-1;10(12):5501–13.
Peker N, Couto N, Sinha B, Rossen JW. Diagnosis of bloodstream infections from positive blood cultures and directly from blood samples: recent developments in molecular approaches. Clin Microbiol Infect 2018 2018-9-1;24(9):944–55.
Chen R, Xie M, Wang S, et al. Secondary Infection Surveillance with Metagenomic Next-Generation Sequencing in COVID-19 Patients: A Cross-Sectional Study. INFECT DRUG RESIST 2023.
Serpa PH, Deng X, Abdelghany M, et al. Metagenomic prediction of antimicrobial resistance in critically ill patients with lower respiratory tract infections. GENOME MED 2022;14(1).
Hu X, Zhao Y, Han P, et al. Novel Clinical mNGS-Based Machine Learning Model for Rapid Antimicrobial Susceptibility Testing of Acinetobacter baumannii. J CLIN MICROBIOL 2023 2023-5-23;61(5).
Břinda K, Callendrello A, Ma KC, et al. Rapid inference of antibiotic resistance and susceptibility by genomic neighbour typing. NAT MICROBIOL 2020 2020-2-10;5(3):455 – 64.
Boolchandani M, D Souza AW, Dantas G. Sequencing-based methods and resources to study antimicrobial resistance. NAT REV GENET 2019 2019-3-18.
Rooney AM, Raphenya AR, Melano RG, et al. Performance Characteristics of Next-Generation Sequencing for the Detection of Antimicrobial Resistance Determinants in Escherichia coli Genomes and Metagenomes. MSYSTEMS 2022 2022-6-28;7(3).
Han D, Li Z, Li R, Tan P, Zhang R, Li J. mNGS in clinical microbiology laboratories: on the road to maturity. CRIT REV MICROBIOL 2019 2019-11-6:1–18.
Camargo CH, Yamada AY, de Souza AR, et al. Genomic analysis and antimicrobial activity of β-lactam/β-lactamase inhibitors and other agents against KPC-producing Klebsiella pneumoniae clinical isolates from Brazilian hospitals. SCI REP-UK 2023 2023-9-5;13(1).

No competing interests reported.

Supplementarymaterialfinial.xlsx

Download PDF

Journal Publication

published 01 Apr, 2024

Read the published version in Journal of Infection →

Version 1

posted

You are reading this latest preprint version

Molecular Rapid Diagnostic Testing for Bloodstream Infections: Nanopore Targeted Sequencing with Pathogen-Specific Primers

Status:

Journal Publication

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Introduction

Materials and Methods

Study Design

Study Population

Blood culture and antimicrobial susceptibility testing

Plasma mNGS assay

NTS assay

Statistical analysis

Results

Comparison of NTS Results with mNGS and Blood Cultures

Performance of NTS testing for BSI Diagnosis

Performance of NTS testing for resistance gene detection

Turnaround time (TAT) of the sequencing and analysis

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1