The Comparison of Diagnostic Performance Between Next-Generation Sequencing of Blood and Tissues for Primary Spinal Infections

Abstract

Study design

prospective study.

Objectives

The hematogenous spread of pathogens from a distant infected area is the main route of primary spinal infections. It is expected that blood metagenomic next-generation sequencing (mNGS) has potential in the pathogen detection of primary spinal infections. The aim of this study is to compare the diagnostic performance of blood and tissue mNGS in primary spinal infections.

Methods

A total of 21 patients with primary spinal infections were analyzed. The results of mNGS and culture of blood and spinal specimens were used to calculate the diagnostic efficiency-related parameters.

Results

The positive rate, sensitivity and specificity of blood mNGS were significantly lower than those of tissue mNGS (42.86% vs 90.48%, 9.52% vs 95%, 12.5% vs 100%). The positive rate and sensitivity of blood mNGS were higher (42.86% vs 4.76, 9.52% vs 5%) than those of blood pathogen culture. Also, the sensitivity and specificity of blood mNGS were lower than tissue pathogen culture (9.52% vs 45%, 12.5% vs 100%). Moreover, the specificity of blood mNGS was the lowest among the 4 pathogen identification techniques.

Conclusions

The diagnostic performance of blood mNGS is worse than tissue mNGS in primary spinal infections. The application prospects of blood mNGS in pathogen identification of primary spinal infections are limited. Further studies will be required to investigate the diagnostic values of blood mNGS in other types of spinal infections or in subpopulations of spinal infections.

Keywords

primary spinal infections next-generation sequencing diagnosis blood

Introduction

Spinal infections (SIs) are pathologies that affect the vertebral body, intervertebral discs, paravertebral soft tissues, epidural space, dura mater, and spinal cord and account for 2%-16.7% of all cases of osteomyelitis.^1,2 The age distribution of patients with spinal infections is bimodal, with peaks occurring in those under 20 years old and those between 50 and 70 years old.³ Established risk factors include prior spinal surgery history, distant infection foci, diabetes mellitus, ageing, intravenous drug use, HIV infection, immunosuppression, history of tumours, renal failure, rheumatic disease and liver cirrhosis.^4,5 In recent years, the global annual incidence of SIs has reached approximately 2.2/100,000.⁶ In developed countries, the mortality rate from SIs ranges from 2% to 20%.⁷ Demographic shifts towards ageing populations and expanding immunocompromised cohorts contribute to rising incidence trends.^3,8 According to the source and background of infection, spinal infections are classified into primary infection and iatrogenic infection. In clinical practice, primary spinal infection is the most common type of spinal infection.⁶

These infections present with insidious onset, nonspecific manifestations, protracted disease courses, and elevated mortality/disability rates.^7,9 This is a demanding problem for spine surgeons. Rapid and accurate aetiological diagnosis is crucial for the management of spinal infections. Conventionally, the aetiological diagnostic methods for spinal infections include blood, pus or spinal tissue culture,^6,10 which have critical limitations: low sensitivity/positivity rates, narrow pathogen spectra, prolonged turnaround times, and antibiotic interference. As a result, SIs are sometimes difficult to be distinguish from degenerative diseases, noninfectious inflammation and spinal tumours,¹ leading to a 3-6 month delay from the onset of symptoms to the diagnosis. Diagnostic delays precipitate suboptimal management, potentially progressing to spinal instability, deformity, epidural abscesses, neurological deficits, cachexia, or mortality. This is a great challenge for the management of spinal infections.^4,11-13 In addition, owing to the indefinite aetiological diagnosis, the antibiotic regimen is not accurate, which can lead to fostering resistance and adverse effects.^14-16 Collectively, these diagnostic inadequacies necessitate advanced rapid precision techniques for pathogen detection in SI management.

Metagenomic next-generation sequencing (mNGS) represents an emerging technology for pathogen detection.^17,18 Unlike traditional microbial culture, it rapidly identifies microorganisms independent of culture via high-throughput sequencing of clinical samples and enables a hypothesis-free and unbiased aetiological diagnosis. The initial application of mNGS mainly focused mainly on the diagnosis of central nervous system infection, and demonstrated efficacy in detecting rare, new and atypical infectious encephalitis.^19-21 Unlike traditional microbial culture, mNGS directly performs high-throughput sequencing of nucleic acids in clinical samples, aligns the obtained sequences with databases, and performs pathogen detection without assumption and culture dependence.^22,23 Therefore, it can theoretically identify all known pathogens in samples, including viruses, bacteria, fungi and parasites.^22,23 Previous studies on the clinical application of mNGS in infectious diseases confirmed that the detection rates of microorganisms in blood, cerebrospinal fluid and biopsy samples were significantly improved.^24,25 In addition, mNGS has been shown to guide treatment regimens and improve antibiotic management by sequencing antibiotic resistance genes.²⁶ At present, mNGS has been widely used in the diagnosis of clinically atypical, rare, difficult-to-culture and coinfected pathogens because of its advantages of rapidity, broad spectrum, high sensitivity and high specificity.^22,26

Our previous study revealed that spinal tissue mNGS has greater sensitivity than traditional pathogen culture of tissues in the identification of spinal pathogens, with good specificity as well. Moreover, mNGS substantially reduces the time cost of pathogen identification.²⁷ Similarly, more studies have demonstrated a better diagnostic performance of spinal tissue mNGS than the conventional aetiological diagnostic techniques for detecting spinal infections.^28,29 The application of mNGS which is based on spinal tissue samples, has advanced of the rapid and accurate diagnosis of spinal infections. However, obtaining spinal tissues through percutaneous biopsy or surgery is invasive and technically demanding. Compared with spinal tissue samples, blood samples have distinct advantages in terms of accessibility and ability to safety obtain. Additionally, the exceptional sensitivity of mNGS allows it to identify microorganisms from specimens with few bacteria (such as blood).³⁰ Previous studies have reported that promising applications of blood mNGS for pathogen identification of surgical site infections after spine surgery.^31,32 Notably, the haematogenous spread of pathogens from a distant infection source is the main route of primary spinal infection. The slow blood flow in endplate vessels increases the susceptibility of endplates to pathogen colonization and eventually causes spinal infections.² Thus, mNGS of blood samples are a potential method for the aetiological diagnosis of primary spinal infections. However, the diagnostic performance of blood mNGS in primary spinal infections remains unknown.

The current study systematically compared the performance of blood mNGS and spinal tissue mNGS in determining the aetiological diagnosis of primary spinal infections. In this study, we analyzed a cohort in which both mNGS of spinal tissues and mNGS of blood samples were applied to identify the pathogens of primary spinal infections. The conventional blood culture and tissue culture results of these patients were also reviewed. The diagnostic efficiency of the different methods was determined based on the basis of the criteria for the clinical diagnosis of primary spinal infections.

Materials and Methods

Study Design

Data from Xinqiao Hospital from December 2023 to March 2025 were reviewed by using the electronic medical records system. The inclusion criteria were described as follows: (1) at least 18 years old; (2) diagnosis was performed in our study and integrated three components: clinical symptoms (refractory back pain with or without neurological deficits), laboratory tests (CRP, ESR, WBC), and imaging findings (MRI and CT); a definitive diagnosis of infection required a comprehensive assessment of all the criteria.^1,2,6; (3) the patients whose spinal tissue samples were obtained by surgery or percutaneous biopsy and were subjected to bacterial culturing, histopathological analysis and mNGS; and (4) the patients whose blood samples were subjected to bacterial culture and mNGS. The exclusion criteria were described as follows: (1) the patients with a history of spinal surgeries or invasive spinal procedures, (2) patients with a history of spinal tumours and spinal trauma, and (3) the patients with the incomplete medical history records. The demographic data of the enrolled patients and the results of conventional examinations and tests were also reviewed (Figure 1).

Figure 1.

Flow Diagram of This Diagnostic Accuracy Study

Specimen Harvesting

Spinal tissue samples were harvested through fluoroscopy-guided percutaneous needle biopsy or debridement surgery. These tissue samples were subjected to mNGS, pathogen culture and histopathological analysis as previously reported.^31,32 The storage and transportation of the tissue samples were performed under strict aseptic conditions to minimize the risk of contamination. Blood samples were harvested on admission and were subjected to pathogen culture and mNGS.

mNGS

mNGS was performed as previously described.^27,32,33 In this study, either a blood sample or a tissue sample was collected from the patient and centrifuged through centrifugation at 3000 rpm for 20 minutes at room temperature to separate the supernatant from the cellular debris. Following centrifugation, the DNA present in the supernatant was extracted using a TIANamp Micro DNA Kit (DP316, TIANGEN BIOTECH, Beijing, China). To accurately quantify the extracted DNA, a Qubit 2.0 fluorometer (Invitrogen, USA) was utilized, which provides precise and sensitive measurements of the DNA concentration, ensuring sufficient input material for library preparation. The extracted DNA was subsequently subjected to a series of library preparation steps. These steps included DNA fragmentation to generate appropriately sized DNA fragments, end repair to produce blunt ends, adapter ligation to enable the binding of sequencing adapters, and polymerase chain reaction (PCR) amplification to increase the yield of DNA fragments with adapters. After these steps, the quality and quantity of the libraries were assessed to ensure their suitability for sequencing. For sequencing, the prepared libraries were converted into DNA nanoballs (DNBs) using the rolling circle replication method. DNB preparation increases sequencing accuracy by increasing the signal-to-noise ratio. The DNBs were then loaded onto the sequencing platform, either the BGISEQ-50 or MGISEQ-2000 platform, for high-throughput sequencing. After sequencing, the raw data generated were subjected to rigorous bioinformatics analysis. The first step involved removing human host sequences by aligning the sequencing reads to the human reference genome (hg19) using the Burrows-Wheeler Alignment (BWA) tool. The remaining sequencing data, which are free from human DNA, were then classified by simultaneous alignment to comprehensive microbial genome databases, such as those available through the NCBI Microbial Genome Databases (https://ftp.ncbi.nlm.nih.gov/genomes/). This alignment enabled the identification of microbial species, including bacteria, viruses, fungi, and parasites (Figure 2).

Figure 2.

Flowchart of mNGS

Diagnostic Parameters

The positive rate, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were used to evaluate the diagnostic efficacy of the different techniques. The reference test was the clinical diagnosis of primary spinal infections based on the basis of a comprehensive analysis of clinical manifestations, imaging features, blood tests, medical history and spinal tissue histopathology. The results of the reference test were determined by three surgeons with at least 5 years of experience in spine surgery. Furthermore, the investigators interpreting the blood and spinal tissue mNGS results, the investigators interpreting the blood and spinal tissue culture results and the investigators determining the results of the reference test were mutually blinded. Moreover, the time costs of different diagnostic techniques were compared.

Statistical Analysis

Continuous variables and categorical variables are described as the means ± standard deviations and frequencies respectively. Sensitivity, specificity, Youden’s index, positive predictive value (PPV) and negative predictive value (NPV) were used to evaluate the diagnostic value of the different aetiological detection methods. Differences between groups were assessed by chi-square tests and McNemar’s test. P value <0.05 indicated statistical significance. Statistical analyses were performed with SPSS 22.0 statistical software (IBM, Armonk, NY).

Results

General Characteristics of the Included Patients

A total of 21 patients with primary spinal infections were enrolled in this study, of whom 57.14% were male and 42.86% were female. A total of 23.81% of the patients had received antibiotics within 1 month. A total of 47.61% of the patients presented with comorbidities, including diabetes mellitus, chronic hepatitis and leukemia. Lumbar involvement was present in 80.95% of the patients. Only 2 patients had a fever before admission. The mean course of disease was 63 (32.5-93) days. Their C-reactive protein and erythrocyte sedimentation rates were increased to 29.99 ± 31.77 and 53.09 ± 31.93 respectively. However, the peripheral blood analysis results were normal. The urinalysis results of 5 patients (23.81%) demonstrated increased leucocyte counts.

mNGS and Pathogen Culture of Blood and Spinal Tissues

Pathogen identification revealed that the pathogens identified in the 20 patients with confirmed diagnosis of infection were tuberculosis (23.81%), gram-positive bacteria (28.57%, of which 14.29% was Staphylococcus epidermidis), gram-negative bacteria (33.33%, of which 14.29% was Escherichia coli), and fungi (4.76%, Candida albicans). Only one patient’s pathological diagnosis was chronic inflammation with no positive test result, which yielded insufficient diagnostic evidence to diagnose infection. (Table 1 and Table 2).

Table 1.

Demographic Characteristics and Clinical Information of the Patients

Characteristics		Included cases (n = 21)
Ages (years)		56.81 ± 10.83
Body mass index		23.01 ± 2.72
Sex (n)	Male	12 (57.14%)
Sex (n)	Female	9 (42.86%)
Smoking (n)	Yes	2 (9.53%)
Course of disease (days)		63 (32.5-93)
Antibiotic usage within one month (n)	Yes	5 (23.81%)
Comorbidity	Total	10 (47.61%)
	Diabetes mellitus	2 (9.53%)
	Chronic hepatitis	1 (4.76%)
	Essential hypertension	4 (19.05%)
	Gastrohelcosis	1 (4.76%)
	Leukemia	1 (4.76%)
	Pneumosilicosis	1 (4.76%)
Involved segment (n)	Cervical spine	1 (4.76%)
	Thoracic spine	3 (14.29%)
	Lumber spine	17 (80.95%)
Fever (>38.5°C)	Yes	2 (9.53%)
Urinalysis	leukocyte+	5 (23.81%)
C-reactive protein (mg/L)		29.99 ± 31.77
White blood cell count (10⁹/L)		7.69 ± 3.39
Erythrocyte sedimentation rate (mm/h)		53.09 ± 31.93
Neutrophils (%)		67.08 ± 12.62

Table 2.

Detailed Information of Each Case

Case	Gender (Male/Female)	Age (years)	Weight (kg)	Height (m)	BMI (kg/m²)	Course of disease (days)	Blood		Tissue		Pathology type	Sample method
Case	Gender (Male/Female)	Age (years)	Weight (kg)	Height (m)	BMI (kg/m²)	Course of disease (days)	Culture	mNGS	Culture	mNGS	Pathology type	Sample method
1	Female	51	59	1.62	22.48	32	Negative	Negative	Streptococcus oralis	Streptococcus oralis	Acute inflammation	Fluoroscopy-guided biopsy
2	Female	52	65	1.63	24.46	63	Negative	Cutibacterium acnes	Negative	Cutibacterium acnes	Chronic inflammation	Fluoroscopy-guided biopsy
3	Male	67	67.5	1.65	24.79	187	Negative	Negative	Negative	Mycobacterium tuberculosis	Granulomatous inflammation	Fluoroscopy-guided biopsy
4	Male	52	55.6	1.72	18.79	67	Negative	Negative	Brucella	Brucella	Chronic inflammation	Surgical
5	Male	58	70	1.62	26.67	93	Negative	Negative	Negative	Staphylococcus epidermidis	Chronic inflammation	Surgical
6	Male	58	55	1.70	19.03	7	Staphylococcus aureus	Torque teno virus	Staphylococcus aureus	Staphylococcus aureus	Acute inflammation	Fluoroscopy-guided biopsy
7	Female	72	50	1.55	20.81	182	Negative	Epstein-barr virus	Negative	Proteus mirabilis	Chronic inflammation	Surgical
8	Female	57	63	1.57	25.56	38	Negative	Staphylococcus epidermidis	Negative	Staphylococcus epidermidis	Acute inflammation	Fluoroscopy-guided biopsy
9	Male	68	70	1.70	24.22	31	Negative	Negative	Negative	Mycobacterium tuberculosis	Granulomatous inflammation	Surgical
10	Male	55	64	1.68	22.68	36	Negative	Negative	Escherichia coli	escherichia coli, acinetobacter pittii	Acute inflammation	Surgical
11	Female	61	60	1.58	24.03	10	Negative	Negative	Escherichia coli	Escherichia coli	Acute inflammation	Surgical
12	Female	54	69	1.58	27.64	30	Negative	Rickettsia bellinis	Negative	Escherichia coli	Acute inflammation	Surgical
13	Male	50	68	1.55	28.30	67	Negative	Negative	Negative	Negative	Acute inflammation	Surgical
14	Male	69	60	1.65	22.04	63	Negative	Negative	Negative	Negative	Chronic inflammation	Surgical
15	Female	73	52.5	1.49	23.65	122	Negative	Negative	Staphylococcus epidermidis	Staphylococcus epidermidis	Chronic inflammation	Surgical
16	Male	27	63	1.76	20.34	29	Negative	Hepatitis B virus	Staphylococcus aureus	Staphylococcus aureus	Acute inflammation	Surgical
17	Female	61	48.8	1.48	22.28	34	Negative	Epstein-barr virus	Negative	Mycobacterium tuberculosis	Granulomatous inflammation	Fluoroscopy-guided biopsy
18	Male	66	52.4	1.60	20.47	63	Negative	Torque teno virus	Brucella	Brucella	Chronic inflammation	Surgical
19	Female	50	45.3	1.58	18.15	94	Negative	Negative	Negative	Mycobacterium tuberculosis	Granulomatous inflammation	Surgical
20	Male	55	60	1.58	24.03	121	Negative	Candida albicans	Candida albicans	Candida albicans	Chronic inflammation	Surgical
21	Male	37	63.9	1.67	22.91	32	Negative	Negative	Negative	Mycobacterium tuberculosis	Granulomatous inflammation	Surgical

In the blood samples, the positive rate from the pathogen cultures was 4.76%, which was significantly lower than that of blood mNGS 42.86% (Table 3). The sole pathogen isolated by blood culture was Staphylococcus aureus (Figure 3A). Multiple viruses such as the Epstein-barr virus, torque teno virus and Rickettsia, were detected by blood mNGS (Figure 3B). However, these microorganisms identified via blood mNGS have no clinical or microbiological correlation with spinal infections (Table 2). In the subsequent analysis, these blood mNGS results were classified as false positives. Conversely, the positive rate from the tissue pathogen cultures was 42.86%, and the positive rate from the tissue mNGS was 90.48% (Table 3). More pathogens including E. coli, Brucella and C. albicans were identified by tissue culture and mNGS (Figure 3C and D). Notably, 10 patients had both positive mNGS results for both blood and tissue. Only 3 patients had concordant positive mNGS results between blood and tissue, including samples from Cutibacterium acnes, Candida albicans and Staphylococcus epidermidis. Six patients presented discordant positive mNGS results between blood and tissue (Table 4). Overall, lesion tissue-derived tests demonstrated superior diagnostic performance. (Table 3)

Figure 3.

Pathogenic Microorganism Detected by Blood Culture, Blood mNGS, Tissue Culture and Tissue mNGS

Table 3.

Results of Next-Generation Sequencing and Conventional Pathogen Culture in Patients

Methods	Culture			mNGS
Source	Blood	Tissue	P Value	Blood	Tissue	P Value
Positive rate	4.76%	42.86%	0.0089	42.86%	90.48%	0.0025
PPV	100%	100%	1.0000	22.22%	100%	<0.0001
NPV	5%	8.33%	1.0000	5%	50%	0.1775
Sensitivity	5%	45%	0.0084	9.52%	95%	<0.0001
Specificity	100%	100%	1.0000	12.5%	100%	0.2222
Time cost (days)	4.95 ± 0.21	4.67 ± 0.94	0.1937	1.05 ± 0.05	1.09 ± 0.29	0.5604

Source	Blood			Tissue
Methods	Culture	mNGS	P Value	Culture	mNGS	P Value
Positive rate	4.76%	42.86%	0.0089	42.86%	90.48%	0.0025
PPV	100%	22.22%	0.3000	100%	100%	1.0000
NPV	5%	5%	1.0000	8.33%	50%	0.2747
Sensitivity	5%	9.52%	1.0000	45%	95%	0.0012
Specificity	100%	12.5%	0.2222	100%	100%	1.0000
Time cost (days)	4.95 ± 0.21	1.05 ± 0.05	<0.0001	4.67 ± 0.94	1.09 ± 0.29	<0.0001

Source	Blood	Tissue	P value	Tissue	Blood	P value
Methods	Culture	mNGS	P value	Culture	mNGS	P value
Positive rate	4.76%	90.48%	<0.0001	42.86%	42.86%	1.0000
PPV	100%	100%	1.0000	100%	22.22%	<0.0001
NPV	5%	50%	0.1775	8.33%	5%	1.0000
Sensitivity	5%	95%	<0.0001	45%	9.52%	0.0012
Specificity	100%	100%	1.0000	100%	12.5%	0.2222
Time cost (days)	4.95 ± 0.21	1.09 ± 0.29	<0.0001	4.67 ± 0.94	1.05 ± 0.05	<0.0001

Table 4.

The Consistency of the Tissue and Blood mNGS

	Proportion	Patient	Pathogens
Consistent positive results of blood and tissue mNGS	33.3%	Case 2	Cutibacterium acnes
		Case 8	Staphylococcus epidermidis
		Case 20	Candida albicans
			Blood NGS	Tissue NGS
Inconsistent positive results of blood and tissue mNGS	66.7%	Case 6	Torque teno virus	Staphylococcus aureus
		Case 7	Epstein-barr virus	Proteus mirabilis
		Case 12	Rickettsia bellinis	Escherichia coli
		Case 16	Hepatitis B virus	Staphylococcus aureus
		Case 17	Epstein-barr virus	Mycobacterium tuberculosis
		Case 18	Torque teno virus	Brucella

Diagnostic Performance of Different Pathogen Identification Techniques

Compared with blood culture, tissue culture demonstrated superior sensitivity in this study. Likewise, compared with blood mNGS, tissue mNGS exhibited a significantly greater sensitivity and positive predictive value (PPV) (Table 3). Although tissue mNGS also showed a greater specificity and negative predictive value (NPV) than did blood mNGS, these differences lacked statistical significance (Table 3). In terms of turnaround time, mNGS requires significantly less processing time than conventional culture dose, with no significant difference observed between blood and tissue mNGS workflows.

Discussion

Previous studies have reported the potential diagnostic potential value of blood mNGS for infectious diseases. Qian et al. (2023) analyzed 1046 surgical and ICU cases and demonstrated the superiority of blood mNGS over conventional methods in detecting fastidious pathogens and polymicrobial infections.³⁴ Moreover, Zhang et al. further reported acceptable diagnostic accuracy of blood mNGS for acute osteomyelitis across 66 patients.³⁵ Given the haematogenous dissemination underlying most primary spinal infections, blood mNGS represents a theoretically promising diagnostic approach. However, our findings revealed that the positive rate from blood mNGS was 42.86%. The sensitivity, PPV and specificity from blood mNGS were significantly lower than those of tissue mNGS vs culture (Table 3).

Both blood mNGS and culture displayed a low sensitivity and NPV, suggesting that blood samples have limited diagnostic value for the aetiological diagnosis of primary spinal infections. In the course of primary spinal infections caused by the haematogenous spread of distant infections, transient bacteraemia is inevitable and is crucial to the occurrence of spinal infections.² Harvesting blood samples from patients with bacteraemia is important for identifying pathogens in the blood. However, the onset of primary spinal infections is insidious, and the clinical manifestations are nonspecific. These patients delay health care-seeking behaviour when they have bacteraemia. As a result, it is difficult to harvest blood samples at the proper time. In the present study, only two patients in our cohort presented with fever. Although fever does not necessarily indicate bacteraemia,³⁶ only 2/21 (9.5%) patients who presented with fever displayed potential bacteraemia indicators, which decreases the decreased sensitivity of blood tests for pathogen identification of primary spina infections to some extent.

Notably, blood mNGS demonstrated the lowest specificity and PPV among the 4 diagnostic modalities, necessitating rigorous clinical correlation to prevent misdiagnosis of primary spinal infections. This limitation is partly attributable to the interpretation of the results. Specifically, viruses are frequently detected via blood mNGS and are classified as positive findings. However, these viral signals lack consistent corroboration in corresponding tissue mNGS samples, preventing their classification as true positives. This discrepancy significantly reduced the specificity of blood mNGS. Furthermore, the low NPV indicates that a negative blood mNGS result cannot reliably exclude a spinal infection. Although its diagnostic efficiency is inferior to that of tissue mNGS and conventional bacterial culture, blood mNGS offers distinct advantages: minimal invasiveness, rapid turnaround time, and the avoidance of invasive spinal biopsy. Blood mNGS retains potential utility during early-stage spinal infections and when biopsy is contraindicated or unfeasible. In fact, early antibiotic use did not significantly reduce the sensitivity of each test in our study. A total of five patients had received oral antibiotics early, and the pathogens that were eventually detected were Brucella, Staphylococcus aureus, Proteus mirabilis, Escherichia coli, Staphylococcus epidermidis (Table 5).

Table 5.

The Results of Different Techniques of the Patients Using Antibiotics Within One Month

Patients	Blood culture	Blood mNGS	Tissue culture	Tissue mNGS
PT-4	Negative	Negative	Brucella	Brucella
PT-6	Staphylococcus aureus	Torque teno virus	Staphylococcus aureus	Staphylococcus aureus
PT-7	Negative	Epstein-barr virus	Negative	Proteus mirabilis
PT-10	Negative	Negative	escherichia coli	escherichia coli
PT-15	Negative	Negative	Staphylococcus epidermidis	Staphylococcus epidermidis

The operational costs associated with mNGS remain a significant consideration. While mNGS offers higher throughput, its complex workflow typically requires 1-2 days for completion and demands skilled personnel, thus entailing significantly higher operational and economic costs.

Our study also has several limitations. First, a low fever incidence (9.53%) means fewer patients with bacteraemia, which causes selection bias concerns and further affects the sensitivity of blood mNGS. In other words, for primary infection, there is no reference test for aetiological diagnosis, and clinical criteria cannot confirm the source of the pathogen. If the results of tissue mNGS and blood mNGS are inconsistent, we recognized the results of the tissue mNGS as aetiological diagnoses which may introduce bias. Our limited cohort size requires validation through prospective STARD-compliant studies (Standards for Reporting of Diagnostic Accuracy Studies)³⁷ to establish blood mNGS's clinical utility.

Conclusion

Compared with tissue mNGS, blood mNGS has significantly inferior diagnostic performance compared to tissue mNGS for primary spinal infections. Although blood sample collection offers procedural convenience, the constrained diagnostic value of blood mNGS necessitates prioritizing spinal tissue mNGS as the primary recommendation for aetiological diagnosis when feasible. However, it presents a viable alternative during the early bacteremia phase of spinal infection or when obtaining tissue samples from the infection site is impractical.

Footnotes

ORCID iDs

Jiang Long

Xue Leng

Yue Zhou

Chencheng Feng

Bo Huang

Ethical Considerations

This study was approved by the Institutional Review Board of Xinqiao Hospital (No. 2025-093-02).

Consent to Participate

All the patients completed the signing of the informed consent form.

Author Contributions

Chang Liu: Dr Liu spearheaded study design, coordinated implementation, and led manuscript preparation. Jiang Long, Yuan Li, Xue Leng, Shipeng Chen, Jiawei Fu, Changqing Li, Yue Zhou: These authors contributed equally to data acquisition, experimental execution, and critical technical/logistical support. Jiajie Zhang: Mr. Zhang held primary responsibility for statistical methodology, data processing, and computational validation. Bo Huang and Chencheng Feng: As corresponding authors, Dr Huang and Feng oversaw all stages of the research, provided senior mentorship, secured resources, and finalized the manuscript for intellectual integrity. All authors reviewed the manuscript, approved the final version, and agreed to accountability for all aspects of the work.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study is supported by the Special Project for Discipline Talent Construction of the Second Affiliated Hospital of Army Medical University (2023XKRC013).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Arbelaez

Restrepo

Castillo

. Spinal infections: clinical and imaging features. Top Magn Reson Imag. 2014;23(5):303-314. doi:10.1097/rmr.0000000000000032

Babic

Simpfendorfer

. Infections of the spine. Infect Dis Clin. 2017;31(2):279-297. doi:10.1016/j.idc.2017.01.003

Duarte

Vaccaro

. Spinal infection: state of the art and management algorithm. Eur Spine J. 2013;22(12):2787-2799. doi:10.1007/s00586-013-2850-1

Carragee

. Pyogenic vertebral osteomyelitis. J Bone Joint Surg Am. 1997;79(6):874-880. doi:10.2106/00004623-199706000-00011

Fantoni

Trecarichi

Rossi

, et al. Epidemiological and clinical features of pyogenic spondylodiscitis. Eur Rev Med Pharmacol Sci. 2012;16(Suppl 2):2-7.

Tsantes

Papadopoulos

Vrioni

, et al. Spinal infections: an update. Microorganisms. 2020;8(4):476. doi:10.3390/microorganisms8040476

Lener

Hartmann

Barbagallo

GMV

Certo

Thomé

Tschugg

. Management of spinal infection: a review of the literature. Acta Neurochir. 2018;160(3):487-496. doi:10.1007/s00701-018-3467-2

Zhou

Wang

Song

Tan

. Improved accuracy of etiological diagnosis of spinal infection by metagenomic next-generation sequencing. Front Cell Infect Microbiol. 2022;12:929701. doi:10.3389/fcimb.2022.929701

Kwon

Hyun

Han

Kim

Jahng

. Pyogenic vertebral osteomyelitis: clinical features, diagnosis, and treatment. Korean J Spine. 2017;14(2):27-34. doi:10.14245/kjs.2017.14.2.27

10.

Eren Gök

Kaptanoğlu

Celikbaş

, et al. Vertebral osteomyelitis: clinical features and diagnosis. Clin Microbiol Infect. 2014;20(10):1055-1060. doi:10.1111/1469-0691.12653

11.

Sobottke

Seifert

Fätkenheuer

Schmidt

Gossmann

Eysel

. Current diagnosis and treatment of spondylodiscitis. Dtsch Arztebl Int. 2008;105(10):181-187. doi:10.3238/arztebl.2008.0181

12.

Nagashima

Tanishima

Tanida

. Diagnosis and management of spinal infections. J Orthop Sci. 2018;23(1):8-13. doi:10.1016/j.jos.2017.09.016

13.

Gupta

Kowalski

Osmon

, et al. Long-term outcome of pyogenic vertebral osteomyelitis: a cohort study of 260 patients. Open Forum Infect Dis. 2014;1(3):ofu107. doi:10.1093/ofid/ofu107

14.

Kim

Song

Park

, et al. Microbiologically and clinically diagnosed vertebral osteomyelitis: impact of prior antibiotic exposure. Antimicrob Agents Chemother. 2012;56(4):2122-2124. doi:10.1128/aac.05953-11

15.

Grigoryan

Germanos

Zoorob

, et al. Use of antibiotics without a prescription in the U.S. population: a scoping review. Ann Intern Med. 2019;171(4):257-263. doi:10.7326/m19-0505

16.

, et al. Responsive assembly of silver nanoclusters with a biofilm locally amplified bactericidal effect to enhance treatments against multi-drug-resistant bacterial infections. ACS Cent Sci. 2019;5(8):1366-1376. doi:10.1021/acscentsci.9b00359

17.

Goldberg

Sichtig

Geyer

Ledeboer

Weinstock

. Making the leap from research laboratory to clinic: challenges and opportunities for next-generation sequencing in infectious disease diagnostics. mBio. 2015;6(6):e01888. doi:10.1128/mBio.01888-15

18.

Deng

Lee

, et al. Rapid pathogen detection by metagenomic next-generation sequencing of infected body fluids. Nat Med. 2021;27(1):115-124. doi:10.1038/s41591-020-1105-z

19.

Wilson

Naccache

Samayoa

, et al. Actionable diagnosis of neuroleptospirosis by next-generation sequencing. N Engl J Med. 2014;370(25):2408-2417. doi:10.1056/NEJMoa1401268

20.

Hoffmann

Tappe

Höper

, et al. A variegated squirrel bornavirus associated with fatal human encephalitis. N Engl J Med. 2015;373(2):154-162. doi:10.1056/NEJMoa1415627

21.

Wilson

Suan

Duggins

, et al. A novel cause of chronic viral meningoencephalitis: cache valley virus. Ann Neurol. 2017;82(1):105-114. doi:10.1002/ana.24982

22.

Han

Tan

Zhang

. mNGS in clinical microbiology laboratories: on the road to maturity. Crit Rev Microbiol. 2019;45(5-6):668-685. doi:10.1080/1040841x.2019.1681933

23.

Simner

Miller

Carroll

. Understanding the promises and hurdles of metagenomic next-generation sequencing as a diagnostic tool for infectious diseases. Clin Infect Dis. 2018;66(5):778-788. doi:10.1093/cid/cix881

24.

Miller

Chiu

. Clinical metagenomic next-generation sequencing for pathogen detection. Annu Rev Pathol. 2019;14:319-338. doi:10.1146/annurev-pathmechdis-012418-012751

25.

Govender

Street

Sanderson

Eyre

. Metagenomic sequencing as a pathogen-agnostic clinical diagnostic tool for infectious diseases: a systematic review and meta-analysis of diagnostic test accuracy studies. J Clin Microbiol. 2021;59(9):e0291620. doi:10.1128/jcm.02916-20

26.

Chiu

Miller

. Clinical metagenomics. Nat Rev Genet. 2019;20(6):341-355. doi:10.1038/s41576-019-0113-7

27.

Wang

Long

Zhuang

, et al. Application of metagenomic next-generation sequencing in the detection of pathogens in spinal infections. Spine J. 2023;23(6):859-867. doi:10.1016/j.spinee.2023.02.001

28.

Zhang

, et al. Clinical application value of metagenomic next-generation sequencing in the diagnosis of spinal infections and its impact on clinical outcomes. Front Cell Infect Microbiol. 2023;13:1076525. doi:10.3389/fcimb.2023.1076525

29.

Wang

Chen

Yuan

. Application of next-generation metagenomic sequencing in the diagnosis and treatment of acute spinal infections. Heliyon. 2023;9(3):e13951. doi:10.1016/j.heliyon.2023.e13951

30.

Chen

Ding

Lei

, et al. Blood and bronchoalveolar lavage fluid metagenomic next-generation sequencing in pneumonia. Can J Infect Dis Med Microbiol. 2020;2020:6839103. doi:10.1155/2020/6839103

31.

Snopko

Kolarovszki

Opšenák

Hanko

Richterová

. Surgical site infections after degenerative lumbar spine surgery. Rozhl Chir. 2018;97(12):539-545.

32.

Zhang

Ding

. The diagnostic value of blood next-generation sequencing in early surgical site infection after spine surgery. Int J Gen Med. 2023;16:37-45. doi:10.2147/ijgm.S394255

33.

Podnar

Deiderick

Hunicke-Smith

. Next-generation sequencing fragment library construction. Curr Protoc Mol Biol. 2014;107:1-7. doi:10.1002/0471142727.mb0717s107

34.

Qian

Zhang

, et al. Blood metagenomics next-generation sequencing has advantages in detecting difficult-to-cultivate pathogens, and mixed infections: results from a real-world cohort. Front Cell Infect Microbiol. 2023;13:1268281. doi:10.3389/fcimb.2023.1268281

35.

Zhang

Chen

Yao

, et al. The diagnostic value of blood metagenomic next-generation sequencing in patients with acute hematogenous osteomyelitis. Front Cell Infect Microbiol. 2023;13:1106097. doi:10.3389/fcimb.2023.1106097

36.

Coburn

Morris

Tomlinson

Detsky

. Does this adult patient with suspected bacteremia require blood cultures? JAMA. 2012;308(5):502-511. doi:10.1001/jama.2012.8262

37.

Cohen

Korevaar

Altman

, et al. STARD 2015 guidelines for reporting diagnostic accuracy studies: explanation and elaboration. BMJ Open. 2016;6(11):e012799. doi:10.1136/bmjopen-2016-012799