Hostname: page-component-6766d58669-nf276 Total loading time: 0 Render date: 2026-05-20T19:40:59.870Z Has data issue: false hasContentIssue false
  • English
  • Français

Real-world clinical impact of plasma cell-free DNA metagenomic next-generation sequencing assay

Published online by Cambridge University Press:  30 January 2025

Ishminder Kaur Open the ORCID record for Ishminder Kaur [Opens in a new window] ,
Bennett Shaw Open the ORCID record for Bennett Shaw [Opens in a new window] ,
Ashrit Multani Open the ORCID record for Ashrit Multani [Opens in a new window] ,
Christine Pham Open the ORCID record for Christine Pham [Opens in a new window] ,
Sanchi Malhotra Open the ORCID record for Sanchi Malhotra [Opens in a new window] ,
Ethan Smith ,
Kristina Adachi ,
Paul Allyn ,
Zackary Bango  and
Omer Eugene Beaird Open the ORCID record for Omer Eugene Beaird [Opens in a new window]
...Show all authors
Ishminder Kaur*
Affiliation:
Division of Infectious Diseases, Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Bennett Shaw
Affiliation:
David Geffen School of Medicine at University of California, Los Angeles, CA, USA
Ashrit Multani
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Christine Pham
Affiliation:
Department of Pharmaceutical Services, University of California, Los Angeles, CA, USA
Sanchi Malhotra
Affiliation:
Division of Infectious Diseases, Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Ethan Smith
Affiliation:
Department of Pharmaceutical Services, University of California, Los Angeles, CA, USA
Kristina Adachi
Affiliation:
Division of Infectious Diseases, Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Paul Allyn
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Zackary Bango
Affiliation:
David Geffen School of Medicine at University of California, Los Angeles, CA, USA
Omer Eugene Beaird
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
JR Caldera
Affiliation:
Department of Pathology and Laboratory Medicine, University of California, Los Angeles, CA, USA
Sukantha Chandrasekaran
Affiliation:
Department of Pathology and Laboratory Medicine, University of California, Los Angeles, CA, USA
Lynn Chan
Affiliation:
Department of Pharmaceutical Services, University of California, Los Angeles, CA, USA
Rabia Cheema
Affiliation:
Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Sarah Daouk
Affiliation:
Division of Infectious Diseases, Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Jaime Deville
Affiliation:
Division of Infectious Diseases, Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Huan Vinh Dong
Affiliation:
Division of Infectious Diseases, Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Austin Fan
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Omai Garner
Affiliation:
Department of Pathology and Laboratory Medicine, University of California, Los Angeles, CA, USA
Pryce Gaynor
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Hannah Gray
Affiliation:
Department of Pathology and Laboratory Medicine, University of California, Los Angeles, CA, USA
Aleksandr Gorin
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Sowmya Kalava
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Meganne Kanatani
Affiliation:
Department of Pharmaceutical Services, University of California, Los Angeles, CA, USA
Andrew Karnaze
Affiliation:
Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Tawny Saleh
Affiliation:
Division of Infectious Diseases, Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Yamini Sharma
Affiliation:
Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Stacey Stauber
Affiliation:
Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Moises Vargas
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Monette Veral
Affiliation:
Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Drew Winston
Affiliation:
David Geffen School of Medicine at University of California, Los Angeles, CA, USA
Lauren Yanagimoto-Ogawa
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Grace Aldrovandi
Affiliation:
Division of Infectious Diseases, Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Karin Nielsen-Saines
Affiliation:
Division of Infectious Diseases, Department of Pediatrics, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Trevon Fuller
Affiliation:
Institute of the Environment and Sustainability, University of California, Los Angeles, CA, USA
Nicholas Jackson
Affiliation:
Department of Medicine Statistics Core, University of California, Los Angeles, CA, USA
Daniel Uslan
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Joanna Schaenman
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Tara Vijayan
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Ashlyn Sakona
Affiliation:
Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA
Shangxin Yang
Affiliation:
Department of Pathology and Laboratory Medicine, University of California, Los Angeles, CA, USA
*
Corresponding author: Ishminder Kaur; Email: ikaur@mednet.ucla.edu
Rights & Permissions [Opens in a new window]

Abstract

Objective:

To describe the real-world clinical impact of a commercially available plasma cell-free DNA metagenomic next-generation sequencing assay, the Karius test (KT).

Methods:

We retrospectively evaluated the clinical impact of KT by clinical panel adjudication. Descriptive statistics were used to study associations of diagnostic indications, host characteristics, and KT-generated microbiologic patterns with the clinical impact of KT. Multivariable logistic regression modeling was used to further characterize predictors of higher positive clinical impact.

Results:

We evaluated 1000 unique clinical cases of KT from 941 patients between January 1, 2017–August 31, 2023. The cohort included adult (70%) and pediatric (30%) patients. The overall clinical impact of KT was positive in 16%, negative in 2%, and no clinical impact in 82% of the cases. Among adult patients, multivariable logistic regression modeling showed that culture-negative endocarditis (OR 2.3; 95% CI, 1.11–4.53; P .022) and concern for fastidious/zoonotic/vector-borne pathogens (OR 2.1; 95% CI, 1.11–3.76; P .019) were associated with positive clinical impact of KT. Host immunocompromised status was not reliably associated with a positive clinical impact of KT (OR 1.03; 95% CI, 0.83–1.29; P .7806). No significant predictors of KT clinical impact were found in pediatric patients. Microbiologic result pattern was also a significant predictor of impact.

Conclusions:

Our study highlights that despite the positive clinical impact of KT in select situations, most testing results had no clinical impact. We also confirm diagnostic indications where KT may have the highest yield, thereby generating tools for diagnostic stewardship.

Information

Type
Original Article
Information
Infection Control & Hospital Epidemiology , Volume 46 , Issue 5 , May 2025 , pp. 504 - 511
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2025. Published by Cambridge University Press on behalf of The Society for Healthcare Epidemiology of America

Background

Prompt and accurate diagnosis of infectious diseases (ID) remains paramount to optimal management of hospitalized patients. Factors that limit the yield of standard microbiological tests include inherent delays of culture-based-diagnostic methods, pretreatment with antimicrobials, the fastidious nature of certain pathogens, challenges with invasive sampling, and insufficient clinical suspicion resulting in suboptimal selection of tests.Reference Dulanto Chiang and Dekker1 Despite widespread commercial and hospital-level availability of polymerase chain reaction (PCR)-based tests, the need for advanced ID diagnostics remains. Plasma cell-free DNA (cfDNA) metagenomic next-generation sequencing assays (cf-mNGS) are gaining attention to bridge this gap.Reference Chiu and Miller2

Currently, Karius Inc. (Redwood City, California) is the only reference laboratory that offers plasma cf-mNGS assay commercially under a CLIA license.Reference Park, Chang and Ledeboer3 Several reports evaluating the clinical utility of the Karius test (KT) have been published and certain variables are emerging as having potentially higher clinical impact but with differences across reports.Reference Edward and Handel4 Among diagnostic indications, data have shown that patients with culture-negative endocarditis, neutropenic fever, pneumonia in immunocompromised patients, and fever of unknown origin may benefit from plasma cf-mNGS testing.Reference Haddad, DeSimone and Chesdachai5–Reference Francisco, Woc-Colburn and Carlson8 Among host characteristics, immunocompromised hosts seem to benefit from the use of plasma cf-mNGS testing.Reference Vissichelli, Morales, Kolipakkam, Bryson, Sabo and Toor9 The clinical relevance of individual organisms and microbiological result patterns generated by KT have received limited investigation, even as an interpretation of polymicrobial results remains a diagnostic dilemma.Reference Vinh Dong, Saleh, Kaur and Yang10–Reference Lee, Dhaheri and Pollock12 Despite the cumulative published experience, the real-world clinical impact of KT in large patient populations is missing.

We examined the real-world clinical impact of KT in a large single-center retrospective cohort to define its clinical utility further stratified on diagnostic indications, host characteristics, and microbiologic result patterns.

Methods

Study design

This study was a retrospective review of 1000 unique cases of plasma cf-mNGS testing across adult and pediatric care settings at a large academic medical center in Los Angeles, California. Multiple KT performed on the same patient were included if they were separated by at least 6 months to ensure unique cases of plasma cf-mNGS testing were being reviewed.

These 1000 KT cases were assigned for clinical adjudication to 12 panels of investigators, with approximately 84 cases per panel. Each panel was comprised of 4 members: 2 ID specialists (at least 1 of them an attending physician), 1 clinical microbiologist, and an additional member (either an ID pharmacist, hospitalist attending physician, or medical student/resident). The clinical microbiologists were responsible for designation of the microbiological patterns of pathogens detected by KT into (A) 1 of 4 bacterial categories ((i) zoonotic/vector-borne, (ii) fastidious/slow-growing, (iii) 3+ bacterial genera detected, or (iv) other bacteria that does not fall in the previous 3 categories)), and/or (B) 1 of 3 fungal categories (mold, yeast, and dimorphic), and/or (C) viral, and/or (D) parasitic categories.Reference Vinh Dong, Saleh, Kaur and Yang10 Polymicrobial results could be assigned up to 3 microbiological patterns. Clinical adjudication was performed by the remaining 3 team members in each panel, whereby each member abstracted data for one-third of the assigned cases and reviewed the remaining cases abstracted by the other team members to ensure the validity of the abstracted data and allow for a minimum 3 assessments of each testing case.

Patient charts were reviewed to collect demographics, past medical history, clinical diagnosis, other microbiologic test results at the time of KT collection, and provider documentation pertaining to KT results. Patient comorbidities were assigned to discrete categories, including immunocompromised status, with further stratifications, or other comorbidities (Supplementary Table 1). Indications for KT were captured using 18 predefined categories, which included major clinical syndromes or concerns for fastidious organisms (Supplementary Table 2). Each patient could be assigned up to 3 comorbidity and 2 diagnostic indication categories. A positive clinical impact was defined as a KT result that led to (a) a new diagnosis not confirmed by standard microbiologic tests, (b) a new diagnosis made earlier than standard microbiologic tests, (c) additional diagnostic evaluation leading to a subsequent positive change in management, (d) avoidance of surgical/tissue biopsy, or (e) de-escalation/discontinuation of therapy. A negative clinical impact was predefined as a KT result that led to (a) unnecessary treatment, (b) unnecessary diagnostic evaluation, or (c) additional time spent in the hospital. No clinical impact was used to define cases where (a) KT results were not acted upon by clinical teams, (b) KT results confirmed a known diagnosis leading to no change in management, or (c) KT missed causative pathogen detected by standard microbiological tests. The indeterminate clinical impact was used where the clinical impact could not be determined from chart review.

Given the large number of investigators involved in clinical adjudication, we performed quality control assessments by assessing differences in the percentage of clinical impact designations by the panels. All but 2 panels were highly consistent in their clinical impact assessments. Our data validation reviews (performed by I.K., S.Y.) revealed that the differences in the 2 outlier panels stemmed from varied interpretations (result not acted upon, or indeterminate impact) of a KT test result that confirmed known diagnosis without change in management. To decrease this inconsistency without introducing further bias, we combined these 3 clinical impact categories (confirmed clinical suspicion without change in management, result not acted upon, could not determine impact) into no clinical impact.

Diagnostic stewardship of KT at our institution during the study period evolved from no oversight (before August 2018), to manual review of each order by microbiology postdoctoral staff to ensure approval by ID specialists (August 2018–February 2023), to computerized provider order entry (CPOE) restricted to ID physicians alone (after February 2023).

Statistical approach

Descriptive statistics were used to describe the clinical impact of KT. Assessment of differences in clinical impact between adult and pediatric patient cohorts and during different phases of diagnostic stewardship at an institutional level were made using χ2 tests.

Further statistical analyses were performed to assess potential associations of clinical and microbiological variables with the positive clinical impact of KT. Pairwise univariable analysis of associations of diagnostic indications with positive clinical impact of KT were performed separately for the pediatric and adult patient cohort. Odds ratios (OR) were determined for each diagnostic indication, and statistical significance was determined by χ2 test or Fisher’s exact test, as appropriate. A similar analysis was done to determine the association of host comorbidities (separately for adult and pediatric cohorts) and KT microbiologic result pattern (combined for adult and pediatric cohorts) with the positive clinical impact of KT. A multivariable logistic regression model was constructed to further investigate the combined effects of the clinical variable(s) (diagnostic indications and host comorbidities) found to have statistically significant (P < .05) associations in the pairwise univariable analyses. All analyses were evaluated using a two-sided alpha level of 0.05 and conducted in R version 4.3.2.

This study was reviewed by the UCLA Human Research Protection Program and received an Institutional Review Board exemption.

Results

Demographics and KT result summary

The final cohort included 1000 cases from 941 individual patients tested using KT between January 1, 2017, through August 31, 2023. Of these patients, 57.9% were male. Of the 1000 unique cases, 299 tests were ordered for patients <18 years of age and 701 for patients 18 years and older (Figure 1a). Most adult patient cohorts had underlying comorbidities at the time of KT (627 of 701, 89.4%), compared with the pediatric patient cohort (194 of 299, 64.9%) (P < .01) (Figure 1b). The adult patient cohort had a higher percentage of immunocompromised patients (460 of 701, 65.6%), compared with the pediatric patient cohort (112 of 299, 37.4%) (P < .01) (Figure 1b). The most common diagnostic indications among pediatric patients were unexplained fevers, culture-negative sepsis/multiorgan failure, neutropenic fever, musculoskeletal infections, and meningoencephalitis. This contrasted with the adult patient population for whom pneumonia in immunocompromised patients, unexplained fevers, concern for fastidious/zoonotic/vector-borne organisms, concern for invasive fungal infection, followed by culture-negative endocarditis were the most common diagnostic indications (Figure 1c).

Figure 1.

Demographics of the study population. Figure 1a, age distribution of the study cohort at the time of Karius testing; Figure 1b, distribution of comorbidities of the pediatric and adult cohort at the time of Karius testing; Figure 1c, distribution of diagnostic indications for Karius test in the pediatric and adult patient cohort.

At least one microorganism was detected by KT in 58.3% of cases, with the top 10 most frequently detected bacteria, fungi, and viruses shown in Supplementary Figure 1. The most common microbiologic result patterns of KT when organisms were identified were “other bacteria,” followed by viruses (Figure 2a).

Figure 2.

Microbiologic results of Karius testing. Figure 2a, Microbiologic result patterns of Karius test. Figure 2b, causative pathogens detected exclusively/earlier by Karius test in comparison with standard microbiological tests in more than one case. Figure 2c, causative pathogens missed by the Karius test in comparison with standard microbiological tests in more than one case.

Clinical impact assessment

KT results were found to have no clinical impact in 82.2% (822 of 1000) of cases, predominantly due to the result not being acted upon by the clinical team (n = 565) (Table 1). KT results missed causative pathogens detected by standard diagnostic methods in 63 cases, with most of the missed organisms (36 of 63, 57.1%) being fungi or mycobacteria. A fungal etiology was missed by KT in 28 cases (12 of which were Aspergillus spp.), and mycobacteria were missed in 8 cases (including 3 cases of Mycobacterium tuberculosis, all diagnosed by tissue biopsy). Frequently missed causative pathogens are shown in Figure 2c. The standard microbiologic tests and specimen sources that identified the frequently missed pathogens by KT are shown in Supplementary Table 3, with sites of infection summarized in a footnote. In cases where KT missed identifying the causative pathogen, it either generated results with other nonrelevant microorganisms (n = 26) or no microorganisms (n = 37).

Table 1.

Summary of clinical impact of Karius results (total N = 1000)

Positive impact was found in 162 cases (16.2%), largely due to a new diagnosis not confirmed by (n = 69) or made earlier (n = 34) than standard microbiologic tests (Table 1). The most frequent causative pathogens detected exclusively or earlier by KT are shown in Figure 2b and include Streptococcus spp. (including S. pneumoniae and S. pyogenes), Aspergillus spp., Mucorales, anaerobes, Pneumocystis jirovecii, and Legionella.

Negative impact was found in only 16 cases (1.6%). This was largely (n = 11) due to KT results leading to additional unnecessary diagnostic investigations, including quantitative PCRs for viruses found on KT or imaging to evaluate for endocarditis or occult intra-abdominal infection following detection of oral flora or enteric anaerobes by KT.

Association of clinical impact with clinical variables

KT had a positive clinical impact in 43 of 299 pediatric patients (14.4%) and 120 of 701 adult patients (17%) (Supplementary Figure 2a). These differences were not statistically significant (χ2 test, P = .17).

We observed a trend toward increasing positive impact rate with the implementation of each phase of diagnostic stewardship from 5.6% (1 of 18) at baseline to 15.9% (116 of 729) with a manual review of ID approval by microbiology postdoctoral fellows and to 17.8% (45 of 253) with CPOE restricted to ID providers alone (Supplementary Figure 2b). These differences were not statistically significant (Fisher’s exact test P = .50).

Pairwise univariable analysis of diagnostic indications with positive clinical impact of KT showed no significant association in the pediatric cohort, while culture-negative endocarditis and concern for fastidious/zoonotic/vector-borne pathogens showed significant positive association in the adult cohort (P values of .033 and .032, respectively) (Table 2). Meningoencephalitis was negatively associated with the clinical impact of KT in adult patients (P-value of .048) (Table 2). Exclusive or earlier pathogen detection by KT done for diagnostic indications that either trended toward or were associated with higher positive clinical impact are shown in Supplementary Table 4.

Table 2.

Pairwise univariable analysis of diagnostic indication with the positive clinical impact of Karius testing among pediatric and adult patients

Note. CNS, central nervous system.

a P < .05.

Pairwise univariable analysis of host comorbidities showed no association with positive clinical impact in the pediatric or adult patient cohorts (Supplementary Table 5).

Multivariable regression model was constructed only for the adult patient cohort with the input of culture-negative endocarditis, concern for fastidious/zoonotic/vector-borne pathogens, central nervous system infections (lesions or encephalitis combined based on the similar negative signal in pairwise univariable analysis), deep-seated infections, invasive fungal infection, pneumonia regardless of immunocompromised status, and host immunocompromised status. The host immunocompromised status was selected based on prior published data of association with higher KT clinical impact. Pneumonia cases were combined regardless of host immune status due to concerns of collinearity with the incorporation of immunocompromise as a variable in the model construction. Multivariable analysis showed that among the adult patient cohort, culture-negative endocarditis (OR 2.27; 95% CI, 1.11–4.53; P .022) and concern for fastidious/zoonotic/vector-borne pathogens (OR 2.07; 95% CI, 1.11–3.76; P .019) were associated with higher positive clinical impact of KT (Table 3). Host immunocompromised status was not significantly associated with the positive clinical impact of KT (OR 1.03; 95% CI, 0.83–1.29; P .7806) (Table 3).

Table 3.

Results of multivariable analysis of specific diagnostic indications and comorbidities (selected based on results of pairwise univariable analysis or known significance in prior studies) to determine association with positive clinical impact of an adult patient cohort

Note. CNS, central nervous system.

a P-value < .05.

Association of clinical impact with microbiologic patterns of KT result

Pairwise univariable analysis revealed that several microbiologic result patterns of KT were associated with a higher positive clinical impact, namely zoonotic/vector-borne bacteria, followed by fastidious/slow-growing bacteria, molds, and yeasts (Table 4). KT results designated as other bacteria, viruses, and 3+ bacterial genera were associated with lower positive clinical impact.

Table 4.

Pairwise univariable analysis of microbiologic result patterns of Karius testing with its positive clinical impact (results represent the combined pediatric and adult patient cohort)

a P < .05.

Discussion

Our study shows that in a large academic center with a high volume of KT testing, most ordered KT had no clinical impact (82.2%), while smaller percentages had a positive (16.2%) or a negative (1.6%) clinical impact. We also found a trend toward improving clinical impact over time in parallel to more ID provider oversight. Multivariable logistic regression modeling found that the highest positive clinical impact of KT in adult patients is seen when performed for select diagnostic indications, namely, culture-negative endocarditis, and concern for fastidious/zoonotic/vector-borne pathogens.

Our findings are supported by the role of KT in culture-negative endocarditis and Q fever in recent studies.Reference Eichenberger, Degner and Scott13–Reference Ranganath, Khodadadi and Abu Saleh15 We did not find an association of positive clinical impact with unexplained fevers, but this cohort was not strictly defined to include patients with fever of unknown etiology and more often included patients with unexplained hospital-onset fevers. Host immunocompromised status is often described as an important consideration in predicting the clinical impact of plasma cf-mNGS testing.Reference Vissichelli, Morales, Kolipakkam, Bryson, Sabo and Toor9 A recent pediatric study showed a limited impact of plasma cf-mNGS testing among immunocompromised children.Reference Castejon-Ramirez, Glasgow, Ferrolino, Hayden, Maron and Hijano16 Our study suggests that while KT may increase the detection of clinically relevant pathogens, the positive clinical impact of KT is modulated more by the type of infection and the nature of suspected pathogens as opposed to the host immune status. We also did not find that pneumonia in immunocompromised patients was a significant predictor of a higher positive clinical impact of KT. A recent prospective study of immunocompromised patients with pneumonia showed that KT provided an additive diagnostic value of 17.4% in 121 patients where standard methods failed to identify an etiology.Reference Bergin, Chemaly and Dadwal7 This could be in part that, fungal and mycobacterial pathogens, even though highly impactful when KT detects them, also represented most of the “missed” causative pathogens by KT. Several of these diagnoses were made in our patient cohort by pursuing invasive tissue specimens, a finding similar to other investigations.Reference Huygens, Schauwvlieghe and Wlazlo17

We found meningoencephalitis as an indication was inversely associated with positive clinical impact (0%, 0 of 24) in adult patients, but not in pediatric patients (21%, 5 of 24). No other literature has demonstrated the diagnostic utility of using plasma cf-mNGS for meningoencephalitis in adults. Instead, direct cerebrospinal fluid mNGS has shown promise in improving the diagnostic yields for meningoencephalitis.Reference Wilson, Sample and Zorn18 In pediatric patients, however, one small study showed plasma cf-mNGS may provide some added value as an adjunct to standard testing for meningitis.Reference Duster, Ramchandar and Foley19

Our study also found that false-negative KT results were seen in 63 clinical cases. This confirms the known problematic sensitivity and specificity of KT.Reference Edward and Handel4,Reference Vissichelli, Morales, Kolipakkam, Bryson, Sabo and Toor9 KT implementation at an institutional level should be accompanied by provider education highlighting that KT does not have the sensitivity of dedicated single microbiologic tests to detect specific targets and should not be relied upon as the sole modality to rule out an infectious etiology in the context of high clinical suspicion. Increased uptake of dedicated testing, such as targeted 16S rRNA for infective endocarditis or a targeted PCR for zoonotic pathogens has the potential to decrease the positive impact of KT.Reference Flurin, Fisher and Wolf20

Our study also lends itself to the comparison of the overall clinical utility of KT testing between adult and pediatric patient populations. Complicated pneumonia in immunocompetent patients and concern for fastidious/zoonotic/vector-borne pathogens trended toward association with positive clinical impact among pediatric patients, but results did not reach significance, likely in part due to smaller sample size. Other investigators have shown positive impact in pediatric patients with complicated pneumonia.Reference Dworsky, Lee, Ramchandar, Rungvivatjarus, Coufal and Bradley21 We also noted that a substantial number of KT were sent among infants. This warrants further study as the role of plasma cf-mNGS has received limited investigation in this cohort.Reference Zhu, Gan and Ge22

Our study has several limitations, including the retrospective and heterogeneous nature of the study cohort. The timing of the KT in the clinical course was up to the discretion of the ordering providers with variability between cases and could have affected its clinical impact. The large number of investigators could be seen as a limitation, but the clinical adjudication was highly aligned and data validation revealed discrepancies that were resolved prior to analysis. The statistical analysis of predictive factors for the positive impact of KT was performed by comparing it to clinically adjudicated diagnostic results in a real-world clinical setting. The standard microbiological testing available to the providers includes a comprehensive repertoire of tests as would be expected for a large academic health system. The results of the multivariate regression modeling should be interpreted in the light of the limitations of the study design and without a priori definition of power needed for such analysis. Further, the real-world clinical impact of the KT will likely be different for centers that have had more intensive diagnostic stewardship oversight of the test upstream—before broad application by front line providers. Lastly, this study was not designed to be a comparison of the clinical performance of KT compared to the standard microbiologic tests because such work has been previously published.Reference Edward and Handel4,Reference Huygens, Schauwvlieghe and Wlazlo17,Reference Blauwkamp, Thair and Rosen23–Reference Rossoff, Chaudhury and Soneji25

Importantly, we showed the lack of clinical impact of plasma cf-mNGS testing with the trend toward improved impact after implementing increased oversight. This highlights the complex role for plasma cf-mNGS testing and the importance of optimizing diagnostic stewardship upfront. We have hopefully generated useful information for diagnostic stewardship of plasma cf-mNGS testing. These data are important as the cost-effectiveness of KT is being explored and given the significant cost of KT.Reference Wright, Simner and Carroll26 Our findings suggest that successful implementation of KT in any institution requires a nuanced, multidisciplinary approach involving ID providers and clinical microbiologists to ensure optimal patient selection and ongoing provider education to aid in test result interpretation. Future prospective studies are needed to better define the role of KT.

Supplementary material

The supplementary material for this article can be found at https://doi.org/10.1017/ice.2024.242.

Author contribution

I.K.: Conceptualization, data collection, clinical adjudication, methodology, data interpretation, statistical analysis, and writing. S.Y.: Conceptualization, data collection, methodology, statistical analysis, data interpretation, and writing. B.S.: Data collection, clinical adjudication, statistical analysis, data interpretation, and writing. A.M. and A.S.: Data collection, clinical adjudication, methodology, data interpretation, and writing. N.J., T. F., J.S., and K.N.S.: Statistical analysis and manuscript revision. C.P., T.V., D.U., S.M., E.S., K.A., and H.V.D.: Data collection, clinical adjudication, and significant contributions to manuscript revision. D.W.: Data interpretation and manuscript revision. All other co-authors participated in chart abstraction and clinical adjudication, revised the manuscript, and approved the manuscript. No form of payment was given to anyone to produce this work or creation of the manuscript. The authors report no conflicts of interest related to this work.

References

Dulanto Chiang, A, Dekker, JP. From the pipeline to the bedside: advances and challenges in clinical metagenomics. J Infect Dis 2020;221:S331–S340.CrossRefGoogle Scholar
Chiu, CY, Miller, SA. Clinical metagenomics. Nat Rev Genet 2019;20:341–355.CrossRefGoogle ScholarPubMed
Park, SY, Chang, EJ, Ledeboer, N, et al. Plasma microbial cell-free DNA sequencing from over 15,000 patients identified a broad spectrum of pathogens. J Clin Microbiol. 2023;1:e01855–22.Google Scholar
Edward, P, Handel, AS. Metagenomic next-generation sequencing for infectious disease diagnosis: a review of the literature with a focus on pediatrics. J Pediatric Infect Dis Soc 2021;10:S71–S77.CrossRefGoogle ScholarPubMed
Haddad, SF, DeSimone, DC, Chesdachai, S, et al. Utility of metagenomic next-generation sequencing in infective endocarditis: a systematic review. Antibiotics (Basel) 2022;11:1798.CrossRefGoogle ScholarPubMed
Benamu, E, Gajurel, K, Anderson, JN, et al. Plasma microbial cell-free DNA next-generation sequencing in the diagnosis and management of febrile neutropenia. Clin Infect Dis 2022;74:1659–1668.CrossRefGoogle ScholarPubMed
Bergin, SP, Chemaly, RF, Dadwal, SS, et al. Plasma microbial cell-free DNA sequencing in immunocompromised patients with pneumonia: a prospective observational study. Clin Infect Dis 2024;78:775–784.CrossRefGoogle ScholarPubMed
Francisco, DMA, Woc-Colburn, L, Carlson, TJ, et al. The effect of a plasma next-generation sequencing test on antimicrobial management in immunocompetent and immunocompromised patients-A single-center retrospective study. Antimicrob Steward Healthc Epidemiol 2023;3:e31.CrossRefGoogle ScholarPubMed
Vissichelli, NC, Morales, MK, Kolipakkam, B, Bryson, A, Sabo, RT, Toor, AA. Cell-free next-generation sequencing impacts diagnosis and antimicrobial therapy in immunocompromised hosts: a retrospective study. Transpl Infect Dis 2023;25:e13954.CrossRefGoogle ScholarPubMed
Vinh Dong, H, Saleh, T, Kaur, I, Yang, S. Elucidating the clinical interpretation and impact of a positive plasma cell-free DNA metagenomics test result-a single center retrospective study. J Appl Lab Med 2024;9:14–27.CrossRefGoogle ScholarPubMed
Hogan, CA, Yang, S, Garner, OB, et al. Clinical impact of metagenomic next-generation sequencing of plasma cell-free DNA for the diagnosis of infectious diseases: a multicenter retrospective cohort study. Clin Infect Dis 2021;72:239–245.CrossRefGoogle ScholarPubMed
Lee, RA, Dhaheri, FA, Pollock, NR, et al. Assessment of the clinical utility of plasma metagenomic next-generation sequencing in a pediatric hospital population. J Clin Microbiol 2020; 58:e00419–20.CrossRefGoogle Scholar
Eichenberger, EM, Degner, N, Scott, ER, et al. Microbial cell-free DNA identifies the causative pathogen in infective endocarditis and remains detectable longer than conventional blood culture in patients with prior antibiotic therapy. Clin Infect Dis 2023;76:e1492–e1500.CrossRefGoogle ScholarPubMed
Eichenberger, EM, de Vries, CR, Ruffin, F, et al. Microbial cell-free DNA identifies etiology of bloodstream infections, persists longer than conventional blood cultures, and its duration of detection is associated with metastatic infection in patients with staphylococcus aureus and gram-negative bacteremia. Clin Infect Dis 2022;74:2020–2027.CrossRefGoogle ScholarPubMed
Ranganath, N, Khodadadi, RB, Abu Saleh, OM. Karius with a Q: role for microbial cell-free DNA next-generation sequencing in diagnosis of acute Q fever. Open Forum Infect Dis 2022;10:ofac666.CrossRefGoogle Scholar
Castejon-Ramirez, S, Glasgow, HL, Ferrolino, JA, Hayden, RT, Maron, G, Hijano, DR. Plasma metagenomic sequencing in immunocompromised children. J Pediatric Infect Dis Soc 2024;13:28–29.CrossRefGoogle ScholarPubMed
Huygens, S, Schauwvlieghe, A, Wlazlo, N, et al. Diagnostic value of microbial cell-free DNA sequencing for suspected invasive fungal infections: a retrospective multicenter cohort study. Open Forum Infect Dis 2024;11:ofae252.CrossRefGoogle ScholarPubMed
Wilson, MR, Sample, HA, Zorn, KC, et al. Clinical metagenomic sequencing for diagnosis of meningitis and encephalitis. N Engl J Med 2019;380:2327–2340.CrossRefGoogle Scholar
Duster, N, Ramchandar, N, Foley, J, et al. Application of cell-free plasma next-generation sequencing technology in the diagnosis and management of pediatric meningitis. J Microbiol Methods 2024;217–218:106887.CrossRefGoogle ScholarPubMed
Flurin, L, Fisher, CR, Wolf, MJ, et al. Comparison of blood-based shotgun and targeted metagenomic sequencing for microbiological diagnosis of infective endocarditis. Open Forum Infect Dis 2023;10:ofad546.CrossRefGoogle ScholarPubMed
Dworsky, ZD, Lee, B, Ramchandar, N, Rungvivatjarus, T, Coufal, NG, Bradley, JS. Impact of cell-free next-generation sequencing on management of pediatric complicated pneumonia. Hosp Pediatr 2022;12:377–384.CrossRefGoogle ScholarPubMed
Zhu, Y, Gan, M, Ge, M, et al. Diagnostic performance and clinical impact of metagenomic next-generation sequencing for pediatric infectious diseases. J Clin Microbiol 2023;61:e0011523.CrossRefGoogle ScholarPubMed
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;4:663–674.CrossRefGoogle ScholarPubMed
Langelier, C, Fung, M, Caldera, S, et al. Detection of pneumonia pathogens from plasma cell-free DNA. Am J Respir Crit Care Med 2020;201:491–495.CrossRefGoogle ScholarPubMed
Rossoff, J, Chaudhury, S, Soneji, M, et al. Noninvasive diagnosis of infection using plasma next-generation sequencing: a single-center experience. Open Forum Infect Dis 2019;6:ofz327.CrossRefGoogle ScholarPubMed
Wright, WF, Simner, PJ, Carroll, KC, et al. Progress report: next-generation sequencing, multiplex polymerase chain reaction, and broad-range molecular assays as diagnostic tools for fever of unknown origin investigations in adults. Clin Infect Dis 2022;74:924–932.CrossRefGoogle ScholarPubMed
Figure 0

Figure 1. Demographics of the study population. Figure 1a, age distribution of the study cohort at the time of Karius testing; Figure 1b, distribution of comorbidities of the pediatric and adult cohort at the time of Karius testing; Figure 1c, distribution of diagnostic indications for Karius test in the pediatric and adult patient cohort.

Figure 1

Figure 2. Microbiologic results of Karius testing. Figure 2a, Microbiologic result patterns of Karius test. Figure 2b, causative pathogens detected exclusively/earlier by Karius test in comparison with standard microbiological tests in more than one case. Figure 2c, causative pathogens missed by the Karius test in comparison with standard microbiological tests in more than one case.

Figure 2

Table 1. Summary of clinical impact of Karius results (total N = 1000)

Figure 3

Table 2. Pairwise univariable analysis of diagnostic indication with the positive clinical impact of Karius testing among pediatric and adult patients

Figure 4

Table 3. Results of multivariable analysis of specific diagnostic indications and comorbidities (selected based on results of pairwise univariable analysis or known significance in prior studies) to determine association with positive clinical impact of an adult patient cohort

Figure 5

Table 4. Pairwise univariable analysis of microbiologic result patterns of Karius testing with its positive clinical impact (results represent the combined pediatric and adult patient cohort)

Supplementary material: File

Kaur et al. supplementary material

Kaur et al. supplementary material
Download Kaur et al. supplementary material(File)
File 66.7 KB