Diagnosing infective endocarditis — fellow's quick read

Diagnostic companion to the IE treatment digest (AHA 2026, DeSimone et al). This digest covers the workup from suspicion through classification; the treatment digest covers antimicrobial therapy, surgery indications, and POET/lipoglycopeptide continuation. The 2023 Duke-ISCVID criteria are the endorsed diagnostic framework; the 2026 AHA statement adds practical imaging, molecular, and clinical guidance. Neither carries formal Class/LOE grades (scientific statements by consensus).

In one line

Suspected IE? Get ≥2 blood-culture sets (no timing rules), TTE → TEE, and if prosthetic material is involved consider PET/CT and cardiac CT. Classify with the 2023 Duke-ISCVID framework: tally your major and minor criteria → definite, possible, or rejected. Know your typical organisms (the list expanded) and what counts as a firm alternate diagnosis to reject IE.

◆When to suspect IE

Unexplained bacteremia with a typical IE organism (see organism table below), especially without a clear removable source.
New regurgitant murmur (worsening of a preexisting murmur is not enough).
Embolic phenomena: stroke, septic pulmonary emboli, splenic/renal infarcts, Janeway lesions, conjunctival hemorrhages.
Immunologic phenomena: Osler nodes, Roth spots, unexplained AKI with hematuria (immune complex GN), positive rheumatoid factor.
Prosthetic material + fever: prosthetic valve, TAVR, CIED, recent valve repair.
Injection drug use + fever (right-sided IE: tricuspid vegetations, septic pulmonary emboli).

Special populations

TAVR: symptoms may be blunted by advanced age and comorbidities. Enterococci are the leading cause (unlike surgical PVE). Both TTE and TEE sensitivity are reduced by stent artifacts. Maintain a low threshold for multimodality imaging (PET/CT, cardiac CT).
CIED: suspect in any CIED patient with unexplained fever or bacteremia. PET/CT sensitivity is higher for pocket infection than for lead infection. Complete device extraction is preferred even without visible lead vegetations.
Hematogenous spondylodiscitis: IE prevalence up to 33%; TEE is usually warranted.

◆Initial workup: what to order

Blood cultures

≥2 blood-culture sets before antibiotics. Each set = 1 aerobic + 1 anaerobic bottle drawn simultaneously.
The 2023 Duke-ISCVID criteria eliminated timing and separate-venipuncture requirements. Separate venipuncture sites are still recommended when practical but no longer required.
If a nontypical organism is suspected, obtain ≥3 sets (3+ positive sets needed for a nontypical pathogen to qualify as a major criterion).
Prior antibiotics are one of the two main causes of culture-negative IE (the other: organisms routine culture cannot grow). If possible, store a pre-antibiotic plasma sample in the microbiology lab for future metagenomic sequencing if cultures are negative at 72 hours.

Labs

CBC, CRP/ESR (baseline for monitoring; CRP <25% of peak or <20 mg/L is a POET transition criterion).
Rheumatoid factor (minor criterion if positive).
Urinalysis with microscopy: hematuria, proteinuria, cellular casts suggest immune complex GN (a minor criterion; see expanded definition below).
Complement levels (C3, C4), cryoglobulins if GN suspected.
BMP/Cr (baseline renal function; AKI may signal GN or aminoglycoside toxicity).

Imaging algorithm

TTE first. Specificity >90%; sensitivity 40–66% depending on image quality. Sensitivity for perivalvular abscess is only 20–46%.
Proceed to TEE when: TTE negative/inadequate and suspicion persists; prosthetic valve or CIED; suspected perivalvular complications; before oral step-down (POET). Low threshold favored by the AHA 2026 writing group.
Repeat echo in 1 week if initial is negative but clinical suspicion remains high.
Beyond echo (see Advanced Imaging section below): PET/CT for prosthetic-material IE (counts as a major criterion only >3 mo post-implant); cardiac CT for perivalvular complications when TEE inconclusive; brain MRI for neurologic symptoms or pre-operatively.

◆The 2023 Duke-ISCVID criteria: expandable reference

Click any criterion to expand for details on what counts and what does not.

Major criteria

Microbiologic: positive blood cultures

Typical organism from ≥2 separate blood-culture sets. OR
Nontypical organism (occasionally/rarely causes IE) from ≥3 separate blood-culture sets.

A "blood-culture set" = 1 aerobic + 1 anaerobic bottle drawn together. A "positive" set = growth from ≥1 bottle. See the organism table below for what qualifies as "typical."

Does NOT count: a single positive blood culture for a common contaminant (e.g., a single CoNS set in a patient without prosthetic material), or sequencing-based detection of organisms that commonly contaminate cultures.

Microbiologic: positive molecular/serologic tests

PCR or nucleic-acid-based testing (16S/18S amplicon or metagenomic sequencing) for Coxiella burnetii, Bartonella spp., or Tropheryma whipplei positive from blood. OR
Coxiella burnetii anti-phase I IgG >1:800 (or equivalent titer). OR
C. burnetii isolated from a single blood culture. OR
Bartonella IFA: IgG ≥1:800 for B. henselae or B. quintana (IgM + IgG).

Does NOT count as major: serum amplicon or metagenomic sequencing positive for a typical organism other than C. burnetii, Bartonella, or T. whipplei. Those results count as a minor microbiologic criterion instead.

Imaging: echocardiography / cardiac CT findings

Any of the following on echo or cardiac CT:

Vegetation: oscillating intracardiac mass on valve, cardiac tissue, CIED, or other implant, without an alternative anatomic explanation.
Valvular/leaflet perforation: interruption of endocardial tissue continuity.
Valvular/leaflet aneurysm: elongation with saccular outpouching.
Abscess: perivalvular soft-tissue lesion (may progress to organized collection).
Pseudoaneurysm: perivalvular cavity communicating with the cardiovascular lumen.
Intracardiac fistula: communication between 2 chambers through a perforation.
Significant new valvular regurgitation vs. prior imaging.
New partial dehiscence of prosthetic valve vs. prior imaging.

Does NOT count: worsening or changing of preexisting regurgitation (must be new). Valve thickening alone is not listed as a major criterion in Duke-ISCVID 2023 (the 2023 ESC criteria do include it).

Imaging: 18F-FDG-PET/CT

Abnormal metabolic activity involving a native or prosthetic valve, ascending aortic graft (with concomitant valve involvement), intracardiac device leads, or other prosthetic material.

Prosthetic valve: intense, focal/multifocal, or heterogeneous FDG uptake.
NVE and CIED leads: any abnormal uptake pattern.

Timing caveat: Must be performed ≥3 months after prosthetic valve surgical implantation to count as a Major criterion. Within 3 months, it counts only as a Minor criterion (post-operative inflammation confounds).

Does NOT count as a Major criterion: isolated FDG-positive generator pocket without intracardiac infection. Non-diagnostic patterns: low-intensity, diffuse, homogeneous uptake.

Surgical: direct intraoperative inspection New in 2023

Evidence of IE documented by direct inspection during heart surgery. Applies only when neither major imaging criteria nor subsequent histologic or microbiologic confirmation are available.

This is not a license to skip cultures and pathology. Always send surgical specimens for histopathology and microbiologic studies.

Minor criteria

Predisposition

Previous history of IE
Prosthetic valve (surgical or transcatheter/TAVR)
Previous valve repair (surgical or transcatheter)
Congenital heart disease: cyanotic CHD, endocardial cushion defects, VSD, left-sided lesions (bicuspid aortic valve, AS, AI, MVP, MS, MI), right-sided lesions (Ebstein, pulmonary valve anomalies, congenital tricuspid disease), PDA, other congenital anomalies ± repair
More than mild regurgitation or stenosis of any etiology
Endovascular intracardiac implantable electronic device (CIED)
Hypertrophic obstructive cardiomyopathy
Injection drug use

Does NOT count: regurgitation or stenosis that is only mild. The source requires "more than mild."

Fever: >38.0 °C (100.4 °F)

Documented temperature >38.0 °C (100.4 °F). Self-reported or subjective fever does not qualify.

Vascular phenomena

Clinical or radiological evidence of:

Arterial emboli
Septic pulmonary infarcts
Cerebral abscess
Splenic abscess
Mycotic aneurysm
Intracranial hemorrhage
Conjunctival hemorrhages
Janeway lesions
Purulent purpura

Does NOT count: cerebral microbleeds on MRI (present in up to 50–60% of IE patients). AHA 2026 states explicitly that they are not a minor criterion; Duke-ISCVID 2023 does not list them.

Immunologic phenomena

Positive rheumatoid factor
Osler nodes (tender, raised lesions on finger/toe pads)
Roth spots (retinal hemorrhages with pale center)
Immune complex-mediated glomerulonephritis (expanded definition):

Either: (1) unexplained AKI (new eGFR <60 or reduction by ≥1 ordinal level) PLUS ≥2 of: hematuria, proteinuria, cellular casts on urinary sediment, serologic perturbations (hypocomplementemia, cryoglobulinemia, circulating immune complexes); OR (2) renal biopsy consistent with immune complex-mediated renal disease.

Does NOT count: AKI from nephrotoxic antibiotics or contrast alone without the supporting urinary/serologic findings.

Microbiologic evidence (minor)

Positive blood cultures for an IE-consistent organism but not meeting major criterion requirements (e.g., only 1 positive set for a typical organism, or only 2 sets for a nontypical organism). OR
Positive culture, PCR, or nucleic-acid-based test for an IE-consistent organism from a sterile body site other than cardiac tissue (e.g., joint fluid, CSF, vertebral bone). OR
PCR, amplicon, or metagenomic sequencing from blood positive for a typical IE pathogen other than C. burnetii, Bartonella, or T. whipplei (those three are major criteria). OR
A single finding of a skin bacterium by PCR on a valve or wire without additional clinical or microbiologic support (insufficient for definite IE).

Does NOT count: single positive blood culture for a common contaminant, or sequencing-based assays detecting organisms that commonly contaminate cultures or rarely cause IE.

Imaging (minor): PET/CT <3 months post-implant New in 2023

Abnormal metabolic activity on 18F-FDG-PET/CT involving prosthetic valve, ascending aortic graft, CIED leads, or other prosthetic material, when performed within 3 months of surgical implantation.

After 3 months, the same finding would be a Major criterion.

Physical examination: new murmur on auscultation New in 2023

New valvular regurgitation identified on auscultation only, applicable only when echocardiography is unavailable (expert opinion; resource-limited settings).

Does NOT count: worsening or changing of a preexisting murmur. This criterion is superseded by echocardiographic findings whenever echo is available.

◆Typical organisms: native valve vs. prosthetic material

Classification determines whether 2 or 3 positive blood-culture sets are needed for a microbiologic major criterion.

Setting	Typical organisms (2 sets needed for major criterion)
Any valve (native or prosthetic)	*S. aureus* · *S. lugdunensis* New · *E. faecalis* New · All streptococci except S. pneumoniae and S. pyogenes (includes VGS, S. gallolyticus, Groups B/C/G) · Granulicatella / Abiotrophia · Gemella spp. · HACEK (Haemophilus spp., A. actinomycetemcomitans, C. hominis, E. corrodens, K. kingae)
Prosthetic material only (prosthetic valve, TAVR, CIED)	All of the above PLUS: Coagulase-negative staphylococci · C. striatum / C. jeikeium · Serratia marcescens · P. aeruginosa · Cutibacterium acnes · Nontuberculous mycobacteria (esp. M. chimaerae) · Candida spp.

Key changes from 2000: E. faecalis is now typical regardless of primary source or healthcare setting (reclassification increased sensitivity from 70% to 96% without losing specificity). S. lugdunensis is newly typical (any bacteremia carries high IE risk). Non-faecalis enterococci (e.g., E. faecium) were deliberately excluded as typical organisms due to their rarity as a cause of IE. Any organism not listed above requires ≥3 positive blood-culture sets to meet major microbiologic criteria.

◆Classifying your case

Major

Minor

Classification

✓

any

Definite

✓

≥3

Definite

✓

Definite

Possible

Definite IE

Pathologic criteria: microorganisms identified on valve tissue, cardiac tissue, explanted prosthetic material, or arterial embolus (by culture, staining, PCR, mNGS, FISH, or in situ hybridization) in the context of active endocarditis. OR active endocarditis confirmed on histopathology.
Clinical criteria: 2 major; or 1 major + 3 minor; or 5 minor.

Caveat: molecular tests can remain positive long after successful treatment. In patients with prior IE, interpret cautiously. A single skin bacterium detected by PCR on a valve without additional supporting evidence is a minor criterion, not definite IE.

Possible IE

1 major + 1 minor; or 3 minor.
This is the category that generates the most consults. Possible IE is not a stopping point. Repeat imaging, additional cultures, and molecular diagnostics may reclassify the case. Whether to start empiric therapy in the meantime is a clinical-acuity judgment; neither source ties it to the "possible" label.

Rejected IE: when to take IE off the table

This is the decision IM residents struggle with most. "Rejected" has specific criteria; a vague "probably not IE" doesn't qualify.

Any one of these four pathways suffices to reject IE:

A. Firm alternate diagnosis

The most nuanced pathway. Two subtypes, each with specific requirements:

Microbiologic alternate — ALL THREE required:

Identifiable source for BSI with a nontypical IE pathogen (e.g., E. coli from a UTI, Klebsiella from pneumonia). Careful: CoNS is typical when prosthetic material is present, so a "contaminant" line infection does not qualify in a PVE or CIED patient.
Rapid resolution of bloodstream infection.
Absence of IE evidence on cardiac imaging.

Nonmicrobiologic alternate — BOTH required:

A non-IE cause for cardiac imaging findings (e.g., marantic/nonbacterial thrombotic endocarditis, Libman-Sacks).
Absence of microbiologic evidence for IE.

B. No recurrence after a short antibiotic course (<4 days)

Verbatim: "Lack of recurrence despite antibiotic therapy for less than 4 days." This is a retrospective criterion. It applies when antibiotics were stopped after fewer than 4 days (cultures called contaminants, an alternate source explained the picture) and the syndrome did not come back. True IE relapses after an inadequate course; non-recurrence argues against it.

Does NOT mean: defervescence on therapy. IE patients defervesce on treatment too. A S. aureus bacteremia that clears by day 2 of vancomycin has not rejected IE. The 2023 paper describes this wording as a clarification of the 2000 "resolution with antibiotic therapy" phrasing; the practical effect is to rule out the defervescence reading.

C. No pathologic evidence at surgery/autopsy

No pathologic or macroscopic evidence of IE at surgery or autopsy, with antibiotic therapy for less than 4 days. The source does not explain the <4-day qualifier; the usual clinical logic is that prolonged antibiotics can sterilize a vegetation and make pathology falsely negative.

D. Does not meet possible criteria

Fewer than 1 major + 1 minor, and fewer than 3 minor criteria met.

Remember: "Rejected" means the current evidence does not support the diagnosis. It does not mean "impossible." If new findings emerge (delayed culture positivity, new embolic event, new imaging finding), re-evaluate.

◆Advanced imaging deep dive

When to go beyond TTE/TEE, and what to expect from each modality.

18F-FDG-PET/CT

Best for: prosthetic-material IE (prosthetic valves, CIEDs, grafts). Sensitivity >80–90% with similar specificity. Utility similar for bioprosthetic and mechanical valves.
Not recommended for NVE: sensitivity only ~31%. High specificity may help in select cases, but a negative PET/CT cannot exclude NVE.
Prep: 24h ketogenic diet (high-fat, low-carb) to suppress physiologic myocardial FDG uptake. In urgent situations, imaging without dietary prep can still be interpretable.
Diagnostic uptake: high-intensity, focal/multifocal or heterogeneous. Likelihood of infection increases with higher SUV, but no standardized threshold exists.
Non-diagnostic uptake: low-intensity, diffuse, homogeneous (physiologic or non-specific).
False positives: recent surgery, sterile inflammation, AtriClip devices, BioGlue foreign body reactions, certain prosthetic valves (e.g., Medtronic Mosaic: intense but diffuse/homogeneous).
False negatives: low-grade infections, prior antibiotic therapy, inadequate myocardial suppression.
Timing: ≥3 months post-implant for a major criterion. <3 months = minor only (though recent data suggest post-surgical inflammation may be distinguishable from infection).
Bonus: identifies extracardiac foci (vertebral osteomyelitis, splenic abscess, septic emboli). Less useful for cerebral emboli (high physiologic brain FDG uptake).

Cardiac CT (CCT)

Best for: perivalvular/periprosthetic complications when TEE is inconclusive or contraindicated.
Superior to TEE for: pseudoaneurysm/abscess detection (78% vs 69%).
TEE is better for: vegetations (94% vs 64%), valvular perforation (81% vs 41%), paravalvular leaks (69% vs 44%).
CCT + echo combined has superior sensitivity for all valvular and paravalvular lesions compared with either alone.
Coronary assessment: CCT can evaluate native coronaries (alternative to invasive angiography in low/intermediate-risk patients) and CABG graft patency preoperatively.
Limitations: coronary assessment benefits from slower heart rates (harder in critically ill patients with tachycardia/AF); iodinated contrast caution in renal dysfunction/sepsis.

Brain MRI

Indications: any neurologic symptoms; reasonable pre-operatively even in asymptomatic patients.
What to look for: DWI for acute ischemic stroke (including "silent" emboli), ring-enhancing brain abscesses, MRA for mycotic aneurysms, intracranial hemorrhage.
Cerebral microbleeds: present in 50–60% of IE patients on T2*-weighted sequences. Usually asymptomatic. NOT a minor criterion.
Surgical timing: detection of silent ischemic emboli is not a reason to delay cardiac surgery (not associated with worse outcomes).

WBC SPECT/CT

Uses autologous leukocytes labeled with 99mTc-HMPAO or 111In-oxine. More specific for pyogenic infections than PET/CT (which reflects inflammation broadly).
Good for cardiovascular prosthetic-material infections. Reduced sensitivity for vegetations <1 cm, and for non-pyogenic organisms (Coxiella, T. whipplei, fungi).
Extended procedural time (up to 24 hours). Use when PET/CT is unavailable and high clinical suspicion persists despite inconclusive echo.

◆Culture-negative IE: key tests to order

Two main causes: prior antibiotics, and organisms routine culture cannot grow. The full algorithmic workup is in the 2025 AHA BCNE scientific statement.

Store a pre-antibiotic plasma sample for possible metagenomic sequencing if blood cultures are negative at 72h.
Bartonella: IFA for IgM and IgG to B. henselae and B. quintana. IgG ≥1:800 = major criterion.
Coxiella burnetii (Q fever): phase I and phase II IgG. Phase I IgG >1:800 = major criterion.
Tropheryma whipplei: PCR from blood. Positive = major criterion.
Valve tissue molecular diagnostics (if surgery performed): 16S/18S rRNA PCR, metagenomic sequencing, FISH. Can be deferred if blood and valve cultures are concordantly positive for a typical organism.
Interpretation caveat: DNA detection on resected valve tissue does not necessarily indicate viable organisms. Positive molecular diagnostics on valve tissue do not extend post-operative antibiotic duration.

The Endocarditis Team

The 2026 AHA statement emphasizes the multidisciplinary endocarditis team model: infectious diseases, cardiology, cardiothoracic surgery, neurology, radiology, microbiology, pharmacy, nursing, and where relevant addiction medicine, dentistry, and primary care. The final decision to operate rests with the cardiovascular surgeon.

For antimicrobial therapy, surgery indications, POET oral step-down, and post-operative management, see the companion IE treatment digest (AHA 2026, DeSimone et al).

Anki cards (8 cards minted 2026-09-08)

Deck: Bugs and Drugs Clinicals · Model: Basic-6610e · Tags: Guideline::AHA_IDSA::Endocarditis, Subject::Infectious_Disease

Duke-ISCVID classification thresholds — definite (2M / 1M+3m / 5m) vs possible (1M+1m / 3m), plus pathologic criteria.
Typical organism reclassification — E. faecalis and S. lugdunensis newly typical; E. faecium deliberately excluded.
Blood-culture set requirements — typical ≥2 sets, nontypical ≥3 sets; timing/venipuncture rules eliminated.
Rejected IE: firm alternate diagnosis — microbiologic (3 requirements) and nonmicrobiologic (2 requirements) pathways.
BCNE major-criterion serologies — Coxiella phase I IgG >1:800, Bartonella IFA IgG ≥1:800, T. whipplei PCR from blood.
Immune complex GN expanded definition — AKI + ≥2 of hematuria/proteinuria/casts/serologic perturbations, or renal biopsy.
Cerebral microbleeds & silent emboli — microbleeds NOT a minor criterion; silent emboli do NOT delay surgery.
TTE sensitivity limitations — 40–66% vegetations, 20–46% abscess; when to proceed directly to TEE.

See also: 11 treatment cards in the companion IE treatment digest (AHA 2026).

Sources: Fowler VG, et al. The 2023 Duke-ISCVID Criteria for Infective Endocarditis: Updating the Modified Duke Criteria. Clin Infect Dis 2023;77:518–526 (PMID 37138445; doi:10.1093/cid/ciad271). — DeSimone DC, et al. Infective Endocarditis: Diagnosis, Antibiotic Therapy, and Management. AHA Scientific Statement 2026, endorsed by ISCVID and CCS. Circulation 2026;154:e•••–e••• (doi:10.1161/CIR.0000000000001466). — Delgado V, et al. 2023 ESC Guidelines for the Management of Endocarditis. Eur Heart J 2023;44:3948–4042 (doi:10.1093/eurheartj/ehad193). — AHA 2025 Blood Culture-Negative Endocarditis Scientific Statement (referenced as Supplemental Figure 1 in DeSimone 2026).