Bringing doctors together

or

A free, private network for verified physicians

Leading voices

Dr. Christopher Graber

Antibiotic Stewardship · Greater Los Angeles VA

View

Dr. Jonathan Ryder

Antibiotic Stewardship · UNMC

View

Dr. Chris Beyrer

Global Health · Duke University

View

Dr. Lance Okeke

HIV/AIDS Medicine · Duke University

View

Dr. Raphael J. Landovitz

HIV/AIDS Medicine · UCLA

View

Dr. Keith Kaye

General ID · RU

View

Dr. Helen Boucher

Transplant/Immunocompromised States · Tufts

View

Dr. Sara Dong

Transplant/Immunocompromised States · Boston Children's Hospital

View

Dr. Aniruddha Hazra

HIV/AIDS Medicine · UCMC

View

Dr. Rochelle Walensky

General ID · Harvard Medical School

View
  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    Quantitative Microbial Cell-Free DNA Sequencing From Plasma: A Potential Biomarker for the Diagnosis of Staphylococcal Infection of Cardiac Implantable Electronic Devices.

    Nov 6, 2025DOI: 10.1093/cid/ciaf113Journal ArticleCID

    Written by: Adolf W KarchmerN Jenifer KaufmanSarah Y ParkRonuk M ModiAhmed Abdul-AzimPolly van den BergJason D MatosDaniel B KramerPeter J Zimetbaum

    Find these authors on Roon

    AA
    Dr. Ahmed Abdul Azim
    Antibiotic Stewardship · RU

    MATERIALS AND METHODS

    We prospectively studied CIED patients with staphylococcal bacteremia admitted to Beth Israel Deaconess Medical Center from November 2019 through December 2023 for treatment of CIED-IE. Treating physicians determined patient management. We screened all patients with a CIED who were hospitalized with staphylococcal bacteremia. Patients aged ≥18 years with suspected CIED-IE for whom CIED removal was recommended and who provided informed written consent were enrolled if a positive blood culture was obtained within 7 calendar days, regardless of duration of antibiotic therapy. Patients were eligible for enrollment if entry portals for bacteremia and concurrent staphylococcal infections were treated. We excluded patients with deep staphylococcal infection, including left-sided endocarditis, wherein total source control was not assured.Blood specimens were collected from patients on consenting; every 24 to 48 hours on therapy prior to extraction; immediately before and following CIED extraction (the latter termed “on extraction”); and 1 hour, 2 hours, 12–24 hours, and 48–96 hours after extraction. Specimens were collected by venipuncture or, when necessary, from a central line, using a 10-mL K2 ethylenediaminetetraacetic acid tube. Specimens were promptly centrifuged and frozen at −20°C. When specimen collections for a patient were complete, specimens were shipped overnight, frozen, to the Karius, Inc laboratory.The study was approved by Beth Israel Deaconess Medical Center Committee on Clinical Investigation (#FWA00003245) and was Helsinki Declaration compliant. Each patient or their legally authorized representative provided written informed consent.Clinical data were abstracted from patients’ electronic medical records; these included medical history and physical examination findings, evidence of portal of entry for bacteremia and sites of infection, dates and results of blood and other cultures, results of diagnostic imaging including echocardiography, operative findings, pathology reports, and details of antibiotic treatment. Our hospital clinical microbiology laboratory used standard techniques to speciate recovered bacteria and determined staphylococcal antimicrobial susceptibility with Vitek (bioMérieux) or Phoenix (Becton Dickinson) gram-positive panels. Microbiology data from referral hospitals were determined in their diagnostic laboratories. All patients were hospitalized through emergency departments where generally 2 blood culture sets were collected simultaneously and promptly followed by antibiotic therapy. Accordingly, positive blood cultures on admission were considered a single set, if not indicated otherwise, and reported as a calendar day of bacteremia.
    Duration of bacteremia was defined as calendar days of positive blood cultures until cultures were consistently negative (minimum of 48 hours). Duration of antibiotic therapy to a clinical event was calculated as calendar days of pathogen-appropriate, guideline-consistent endocarditis therapy, starting the day therapy was initiated. Using the EHRA 2019 criteria [9] and medical record data, we classified suspected CIED-IE during treatment and before mcfDNA data were known. Definite CIED-IE required 2 major criteria (bacteremia on at least 2 days plus direct anatomic or microbiologic evidence of lead infection). Possible CIED-IE required 1 major plus a minor criterion or 3 minor criteria (admission day bacteremia, predisposition, or temperature >38°C).
    Blood specimens were shipped to the Karius, Inc laboratory (Redwood City, California), certified under the Clinical Laboratory Improvement Amendments of 1988 and accredited by the College of American Pathologists, for detection, identification, and quantitation of mcfDNA in plasma as previously described [21, 22]. Sequencing data were analyzed using the Karius Test bioinformatic pipeline 3.15, designed to report detected microbes from among >1000 microbes as long as the minimum sequencing coverage metrics required by quality controls are met and to provide absolute plasma concentration of mcfDNA in DNA molecules per microliter (MPM) for each microbe detected. The test detects 7 antimicrobial resistance markers, including SCCmec, mecA, and mecC, and reports S. aureus methicillin susceptibility or resistance [23]. The inability to determine antimicrobial resistance (eg, because of insufficient DNA sequences or ambiguous linkages to a qualified pathogen) is reported as indeterminant. Our cases experienced only staphylococcal bacteremia; thus, we report staphylococcal mcfDNA. Sequencing mcfDNA is pathogen agnostic; other microbes, not considered clinically relevant, were occasionally identified along with staphylococci, particularly from specimens collected from indwelling central lines. These mcfDNA data were not analyzed for this study.
    Continuous data are presented as median with interquartile range (IQR [ie, 25th–75th percentiles]), expressed as days. To compare the kinetic patterns of mcfDNA between groups (categorical data), the Fisher exact test was used with statistical significance set at P < .05.

    RESULTS

    Twenty-one patients with a CIED and staphylococcal bacteremia were enrolled. In 1 patient, extraction resulted in fracture of a lead with a fragment remaining in the coronary sinus. This patient relapsed after 6 weeks of vancomycin therapy and was studied again; thus, we enrolled 22 episodes (Figure 1). Three patients became unavailable for specimen collection and thus were not studied. Three patients with deep infections diagnosed after enrollment were excluded, per protocol, because the infections would confound assessing the relationship of mcfDNA to CIED infection. Of the remaining 16 cases, 3 were admitted directly to our hospital and 13 were transferred from other hospitals. Each patient received pathogen-specific recommended endocarditis antibiotic therapy beginning with their presentation for care.
    Consolidated Standards of Reporting Trials (CONSORT) diagram demonstrating patients enrolled in the study. *One patient was enrolled twice. Abbreviation: CIED, cardiac implantable electronic device.
    Eleven patients were classified as definite CIED-IE (Table 1). Each had multiple days of positive blood cultures for S. aureus or Staphylococcus epidermidis and direct evidence of lead infection. Infection arose from the generator pocket in 3 cases, a defined entry portal in 5 episodes, and an unknown source in 3 episodes. The median total duration of bacteremia was 7 days (IQR, 4–14 days; range, 3–17 days).
    Patients With Definite Cardiac Implantable Electronic Device Infective Endocarditis: Demographics and Clinical Features
    Abbreviations: CIED, cardiac implantable electronic device; CoNS, coagulase-negative staphylococci; CRT-D, cardiac resynchronization therapy–defibrillator; Echo, echocardiogram; F, female; ICD, implantable cardioverter-defibrillator; M, male; MRSA, methicillin-resistant Staphylococcus aureus; MSSA, methicillin-sensitive Staphylococcus aureus; PPM, pacemaker; SUD, substance use disorder; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography; TV, tricuspid valve.
    aTotal calendar days of bacteremia.
    bCase D-011 is same patient as D-008 in whom a fragment of the fractured infected coronary sinus lead was retained during 6 weeks of vancomycin therapy. Despite microbial cell-free DNA rapidly becoming undetectable during antibiotic therapy of D-008 after extraction (see Table 3), the patient represented as case D-011 with relapsed CIED infective endocarditis with bacteremia.
    cD-012 lead explanted and tricuspid valve (with vegetation) replaced at surgery.
    Five patients classified as possible CIED-IE had 1 major criterion plus 2 minor criteria or only 3 minor criteria (Table 2). In case P-006 with bacteremia on 2 days, transesophageal echocardiography disclosed a vegetation on the tip of a flail tricuspid valve leaflet but no lead vegetation. In the absence of direct evidence of lead infection, this patient was classified as possible CIED-IE. The median total duration of bacteremia was 1 day (IQR, 1–1.5 days). The end of bacteremia was confirmed by negative blood cultures on consecutive days (median, 4 days [IQR, 3–5 days]).
    Cases Classified as Possible Cardiac Implantable Electronic Device Endocarditis: Demographics and Clinical Features
    Abbreviations: AVR, aortic valve replacement; CIED, cardiac implantable electronic device; CRT-P, cardiac resynchronization therapy–pacemaker; Echo, echocardiogram; ICD, implantable cardioverter-defibrillator; IE, infective endocarditis; M, male; MRSA, methicillin-resistant Staphylococcus aureus, MSSA, methicillin-sensitive Staphylococcus aureus; PPM, pacemaker; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography; TV, tricuspid valve.
    aTotal calendar days of bacteremia.
    Table 3 shows mcfDNA concentrations in serial plasma specimens for each case. The days of endocarditis guideline–consistent antibiotic therapy on consenting and at extraction are shown. Supplementary Table 1 demonstrates the days of sterile blood cultures before extraction. In each case, mcfDNA, when detected, confirmed the species isolated from the patient's blood. In 7 of 10 cases where S. aureus mcfDNA was detected, it confirmed the methicillin susceptibility status reported for the respective blood isolate, and in 3 cases it was indeterminant.
    Microbial Cell-Free DNA in Plasma Specimens During Antibiotic Therapy on Days After Enrollment and Time After Cardiac Implantable Electronic Device Extraction
    Bolded values are quantitative estimates wherein qualitative detection was confirmed but minimum quality control sequence coverage was not met.
    Abbreviations: CIED, cardiac implantable electronic device; mcfDNA, microbial cell-free DNA; MPM, molecules per microliter; MRSA, methicillin-resistant Staphylococcus aureus; MSSA, methicillin-susceptible Staphylococcus aureus; S. epid, Staphylococcus epidermidis; Rx, antibiotic therapy.
    aExtraction via sterile generator pocket: all cases except at cardiac surgery (cases D-011, D-012), and infected pocket (cases D-008, D-019).
    bOn extraction: immediately following extraction.
    cSame specimen baseline/preextraction.
    dCase D-011 is same patient as D-008 in whom lead fractured on extraction and was retained in coronary sinus during 6 weeks of vancomycin therapy. mcfDNA rapidly became undetectable during vancomycin therapy. Case D-011 is relapsed CIED infective endocarditis with bacteremia after discontinuation of vancomycin therapy.
    eBoth specimens on day 5 of therapy.
    fSpecimen collected after transvascular aspiration debridement of vegetation before extraction.
    gSpecimen collected after extraction.
    hSpecimen collected day 3 after enrollment (day 6 of treatment).
    In 10 of the 11 definite CIED-IE cases, plasma mcfDNA persisted during antibiotic treatment and beyond the period of bacteremia (Table 3). In these 10 cases, plasma mcfDNA was detected at consent after receiving antibiotic therapy for a median of 6 days (IQR, 4–11 days) and on CIED extraction after a median of 7 days (IQR, 4–13 days) of therapy. In the 11th case (D-017), mcfDNA was not detected at consent on day 13 of antibiotic therapy (6 days after the last positive blood culture), but S. aureus mcfDNA was detected later that day in the “on extraction” specimen. Including the period after device extraction, mcfDNA was detected during antibiotic therapy for a median of 11 days (IQR, 7.5–15 days) and for a median of 4 days (IQR, 1.5–6 days) of therapy after the last positive blood culture (Supplementary Table 1). With ongoing antibiotic treatment following device removal, plasma mcfDNA concentration rapidly decreased, becoming undetectable by day 4 in both S. epidermidis infections and in 3 S. aureus infections, including D-012 with surgical lead extraction and tricuspid valve replacement. High mcfDNA plasma concentrations persisted after device extraction in cases D-005 and D-019, reflecting residual tricuspid valve vegetations and continued postextraction bacteremia for 2 and 12 days, respectively.
    In each case of definite CIED-IE, mcfDNA MPM in either the “on extraction” or 1-hour postextraction specimen increased significantly compared to the respective preextraction specimen MPM and was greater than each patient's maximum preextraction MPM (Table 3, Figure 2). These increases exceeded test variability versus the published coefficient of variation [21] (Supplementary Figure 1). In 9 of these cases, the maximum MPM occurred following lead extraction through the generator pocket. Increased MPMs occurred with documented manipulation of lead vegetations in 3 cases. In case D-019 the increase occurred with lead vegetation debridement by transcatheter aspiration and increased further with subsequent transvenous lead extraction. In cases D-011 and D-012, MPM increased with disruption of lead vegetations during CIED removal under direct vision at cardiac surgery.
    The kinetics of microbial cell-free DNA (mcfDNA) molecules per microliter (MPM) (log10) during antibiotic therapy for individual cases classified as definite cardiac implantable electronic device (CIED) infective endocarditis. The x-axis indicates the plasma specimen collection time from day of consent through 4 days postextraction (Post) of the CIED. The vertical red line indicates the time of device extraction and the immediate postextraction specimen (on extraction). Case D-017 is not shown because in all specimens except that on extraction, mcfDNA were 0 (undetectable) and these cannot be shown on log scale. Persistent high mcfDNA MPM after extraction in cases D-005 and D-019 is associated with residual tricuspid valve vegetations and bacteremia after extraction. In D-012 the device was explanted and tricuspid valve and vegetation resected at cardiac surgery. Data presented are from Table 3, which by individual case number provides days of therapy from initial bacteremia detection to consent and other time-specific relationships.
    In cases of possible S. aureus CIED-IE, mcfDNA was initially undetectable after a median of 6 days (IQR, 5.5–7.5 days) of antibiotic therapy prior to extraction (Table 3). It was not detected “on extraction” after a median of 7 days (IQR, 6.5–8.5 days) of antibiotic therapy, nor after a median of 6 days (IQR, 5.5–7.0 days) of therapy following the last positive blood culture (Supplementary Table 1).
    The kinetics of staphylococcal plasma mcfDNA differ significantly between our cases of definite and possible CIED-IE, even with relatively similar periods of appropriate antibiotic treatment after documentation of bacteremia (Fisher exact test, P = .001). During antibiotic therapy of patients with definite CIED-IE, mcfDNA detection was prolonged and persisted beyond device extraction, and the concentration increased early after device extraction (Figure 2). In contrast, in our cases of possible CIED-IE, mcfDNA was undetectable after a brief period of antibiotic therapy and remained undetectable after CIED extraction.

    DISCUSSION

    Our study prospectively evaluated mcfDNA sequencing and quantitation in serial plasma specimens to clarify the likelihood of CIED-IE in patients with a CIED and staphylococcal bacteremia and thus inform therapy. In patients classified as definite CIED-IE, we demonstrated the prolonged detectability of plasma mcfDNA during antibiotic therapy, even after treatment rendered blood cultures negative. Moreover, the concentration of plasma mcfDNA early after device extraction increased to levels exceeding those immediately before extraction. In patients classified as possible CIED-IE, we found that plasma mcfDNA became uniformly undetectable by day 6 of antibiotic therapy and remained undetectable after device extraction. Sequencing of detected mcfDNA confirmed the Staphylococcus species isolated from prior blood cultures as well as the methicillin susceptibility status of the majority of S. aureus. The kinetics of plasma staphylococcal mcfDNA can possibly serve as a biomarker for diagnostic and therapy decisions in patients with staphylococcal bacteremia and potential CIED-IE.
    Given the absence of confounding extracardiac deep S. aureus infection in our definite CIED-IE cases, the prolonged duration of S. aureus mcfDNA in spite of antibiotic therapy is consistent with the long duration of S. aureus mcfDNA in plasma or blood of patients with complicated S. aureus bacteremia, including infective endocarditis [18, 20, 24]. The observed rapid decline of mcfDNA after device removal is likewise similar to that following definitive source control in patients with complicated bacteremia [18, 20, 24]. These data provide indirect evidence of CIED-IE.
    In definite CIED-IE cases, the increase of plasma mcfDNA concentrations early after transvenous CIED extraction suggests the release of mcfDNA as a result of disruption of lead infection. This inference is supported by similar increases in plasma mcfDNA concentrations when lead infection was disrupted with transcatheter vegetation debridement or lead explantation at cardiac surgery. These observations are further evidence that the pattern of plasma staphylococcal mcfDNA in patients classified as definite CIED-IE indicates lead infection.
    The mcfDNA pattern in cases classified as possible CIED-IE differs from that in definite CIED-IE cases. In our possible CIED-IE cases, plasma mcfDNA was undetectable by day 6 of antibiotic therapy and remained undetectable after transvenous lead extraction. The significant difference in the kinetics of plasma mcfDNA in the 2 clinically classified groups suggests that they are different clinical entities. Among cases classified as definite CIED-IE, the release of mcfDNA with device extraction strongly supports CIED lead infection. In contrast, the mcfDNA pattern noted among our patients classified as possible CIED-IE implies that bacteremia was uncomplicated and had not infected the CIED leads.
    The timing of specimen collection is important if using mcfDNA as a biomarker to enhance diagnostic precision. Definite CIED-IE case D-017 demonstrates that prolonged antibiotic therapy after bacteremia is quenched can render mcfDNA undetectable. This illustrates the importance of specimens early after antibiotic treatment for distinguishing mcfDNA clearance effected by prolonged treatment versus rapid clearance associated with uncomplicated bacteremia. Patients without CIED infection can have detectable mcfDNA very early in antibiotic therapy (case P-015), underscoring the importance of specimen timing when assessing for early mcfDNA clearance associated with treatment of uncomplicated bacteremia. Our data do not suggest that all patients classified as possible S. aureus CIED-IE are free from lead infection, but rather that strategically timed mcfDNA testing can confirm definite CIED-IE and in possible CIED-IE, can potentially differentiate between patients with and without lead infection.
    If the mcfDNA kinetics we observed are confirmed, plasma mcfDNA can potentially serve as a biomarker to clarify diagnoses in CIED patients being treated for S. aureus bacteremia. Testing could be utilized selectively where clinical data do not provide diagnostic certainty. For example, Nakajima et al reported 178 patients hospitalized with S. aureus bacteremia without evident CIED infection [4]; however, patient follow-up suggested diagnostic imprecision. Empiric device removal was associated with decreased 1-year mortality and treatment without device removal was associated with bacteremia relapse, often with evidence of CIED-IE [4]. Our data suggest that as few as 1 to 3 strategically timed mcfDNA tests, in conjunction with standard clinical evaluation, might enhance diagnostic precision in such cases. Persistence versus early clearance of mcfDNA on day 6 of antibiotic therapy might help distinguish patients with or without CIED-IE and inform decisions regarding device extraction. If residual diagnostic uncertainty prompts empiric device extraction, assessing mcfDNA concentrations in early postextraction versus preextraction plasma could clarify the CIED-IE question and guide postextraction treatment. Among Nakajima and colleagues' patients, such clarifications might expand the candidates for extraction, spare unnecessary device extraction in patients with unlikely lead infection, or inform antibiotic treatment decisions after empiric extraction. In patients with evident pocket infection and staphylococcal bacteremia, absence of increased mcfDNA concentration on extraction might negate intracardiac infection, permitting shorter duration antibiotic therapy and hospitalization. Extraction and replacement of an infected device (eg, for a cardiac resynchronization device average cost is $67 586 [25]) and prolonged intravenous antibiotics are costly and potentially associated with adverse effects. Although mcfDNA testing costs exceed those of conventional diagnostic tests, appropriate care resulting from strategically timed testing in selected patients could be highly cost-effective and improve care.
    Our study is limited by the small population from a single institution; thus, confirmation and refinement are required for generalization. Additionally, concurrent S. aureus extracardiac deep infection or left-sided endocarditis confound mcfDNA kinetics as a biomarker for CIED-IE. Although our data do not address mcfDNA in non-staphylococcal CIED infection, staphylococci are the dominant cause of CIED-IE. Alternatively, the similarity of our findings to reported plasma mcfDNA in patients with complicated versus uncomplicated S. aureus bacteremia and the uniformity of the mcfDNA patterns distinguishing CIED-IE from rejected lead infection suggest that mcfDNA as a biomarker can enhance diagnostic accuracy.

    CONCLUSIONS

    Using serial plasma specimens for mcfDNA sequencing and quantitation in CIED patients with staphylococcal bacteremia, we demonstrated that mcfDNA patterns can serve as biomarkers to improve diagnostic accuracy in patients with suspected CIED-IE. If our findings are confirmed, selective use of strategically timed mcfDNA testing, in conjunction with clinical features, can improve diagnostic precision for CIED-IE and facilitate treatment trials evaluating device removal and duration of antibiotic therapy.
    AADr. Ahmed Abdul Azim

    Dr. Ahmed Abdul Azim

    Infectious Disease · Antibiotic Stewardship
    Rutgers, The State University of New Jersey
    ahmedabdulazim