Infective Endocarditis & Cardiac Device Infection
FDG PET/CT is now part of the diagnostic pathway for infective endocarditis: the 2023 ESC (Duke-ESC) criteria count valvular, perivalvular/periprosthetic or foreign-material (including cardiac implantable electronic device, CIED) lesions on FDG PET/CT or WBC SPECT/CT as a major imaging criterion (for prosthetic valves the 2023 guideline counts intense focal or heterogeneous uptake irrespective of the interval from surgery, dropping the earlier 3-month rule; mild homogeneous post-operative uptake is expected and should not be read as infection). FDG performs best for prosthetic-valve and device (pocket/lead) infection and is far less sensitive in native-valve disease, where vegetations are small. Dietary preparation to suppress physiological myocardial uptake is essential.
Physiological myocardial FDG uptake is the main obstacle: a high-fat, very-low-carbohydrate diet with prolonged fasting (± heparin) suppresses it, so pathological uptake at a valve, prosthetic ring, device pocket or lead becomes visible. WBC SPECT/CT is a high-specificity alternative.
Clinical context & when to image
- Suspected prosthetic-valve endocarditis with a possible or non-diagnostic echocardiogram.
- Suspected CIED (pacemaker/ICD) pocket or lead infection, and assessment of systemic/embolic spread.
- Reclassifying “possible” endocarditis and finding extracardiac emboli and the portal of entry.
- Not first-line for native-valve endocarditis (low sensitivity for small vegetations).
Imaging strategy
- Echocardiography (TTE/TOE) remains first-line.
- FDG PET/CT is recommended in possible prosthetic-valve endocarditis and may be considered in possible CIED-related endocarditis (2023 ESC).
- WBC SPECT/CT should be considered for suspected PVE when echocardiography is inconclusive and FDG PET/CT is unavailable; a positive study can confirm PVE when FDG is indeterminate.
- Whole-body coverage detects septic emboli, spondylodiscitis and the source.
Protocol
- Strict preparation: high-fat, low/no-carbohydrate diet for 24–48 h, then a fast of at least 12 h (± IV heparin 50 IU/kg about 15 min before injection) to suppress myocardial glucose uptake; suppression still fails in 5–20%.
- Standard cardiac/whole-body FDG PET/CT; ECG-gating and CT angiography can add detail.
- Read recently implanted valves by pattern, not by date: mild homogeneous post-operative uptake is expected, while intense focal or heterogeneous uptake counts as a major criterion under the 2023 ESC criteria whatever the interval from surgery.
- Confirm adequacy of myocardial suppression before interpreting.
How to read it
- Infection: focal or heterogeneous FDG uptake at the prosthetic valve/ring, device pocket or along the lead.
- Homogeneous diffuse valve uptake is typical of non-infected prostheses and can persist indefinitely after surgery.
- Search the whole-body images for extracardiac foci and septic emboli.
- Correlate with echo and blood cultures within the Duke-ESC framework.
Diagnostic performance
FDG PET/CT is most useful for prosthetic-valve and device infection. Meta-analyses tabulated in the EANM/SNMMI v2.0 guideline report pooled sensitivity/specificity of ≈86%/≈84% for prosthetic-valve endocarditis (15 studies) and ≈85%/≈90% for CIED infection overall (14 studies), but only ≈36%/≈99% for native-valve endocarditis, where vegetations are small.
| Setting | Sensitivity | Specificity | Comment |
|---|---|---|---|
| Prosthetic-valve endocarditis | ≈86% | ≈84% | Recommended in possible PVE (ESC 2023) |
| Native-valve endocarditis | ≈36% | ≈99% | Small vegetations often missed |
| CIED pocket/generator | High | High | FDG very useful |
| CIED lead / systemic | Moderate–high | High | Assess spread |
Pitfalls
- Inadequate dietary preparation → physiological myocardial uptake obscures or mimics disease.
- Recent cardiac surgery causes post-operative uptake for months.
- Small native-valve vegetations fall below PET resolution.
- Interpret strictly within the clinical/echocardiographic context.
Evidence & guidelines
- The 2023 ESC guidelines for infective endocarditis include ¹⁸F-FDG PET/CT and WBC SPECT/CT among the modalities for the major imaging criterion (valvular, perivalvular/periprosthetic and foreign-material lesions); FDG PET/CT is recommended in possible PVE and may be considered in possible CIED-related IE.
- Meta-analyses support good performance in prosthetic-valve/device infection and low sensitivity in native-valve disease.
- EANM/SNMMI procedural guidance addresses preparation and interpretation.
In depth
- A 2019 meta-analysis (14 studies) reported FDG-PET/CT sensitivity 83% / specificity 89% for CIED infection — far better for pocket infection (96%/97%) than lead infection (76%/83%).
- Accuracy is lower for lead infection because lead vegetations are small, physiological myocardial uptake can obscure intracardiac leads when dietary preparation is inadequate, and prior antibiotic therapy dampens the inflammatory signal.
- Adding FDG-PET/CT to the modified Duke criteria reclassifies most 'possible' cases of prosthetic-valve or device endocarditis as definite or rejected (90% of possible cases in one prospective series of 92 patients).
- For prosthetic-valve endocarditis, Saby et al. (n=72) reported FDG-PET/CT sensitivity of 73% and specificity of 80%, and adding abnormal perivalvular uptake as a major criterion raised the sensitivity of the modified Duke criteria at admission from 70% to 97%.
- After excluding patients with low inflammatory activity (CRP <40 mg/L) or surgical adhesives used at valve implantation, visual PVE sensitivity and specificity rose from 74%/91% to 91%/95%; a separate cohort of 188 patients with prosthetic-valve or ascending-aortic-graft infection reported 93%/90%.
- Native-valve endocarditis has far lower sensitivity (as low as 22%) because leaflet vegetations are small and fibrotic with little leukocyte response.
- Preparation aims to suppress physiological myocardial glucose uptake with a high-fat, low-carbohydrate diet followed by prolonged fasting; an unfractionated heparin bolus shortly before injection is an optional adjunct that raises plasma free fatty acids (Randle cycle).
- The strongest sign of infection is focal or heterogeneous uptake at the valve, annulus or prosthetic material; mild homogeneous uptake around a prosthetic valve (mechanical or biological) can be a non-infective post-surgical finding.
- Pitfalls: attenuation-correction artefacts (compare NAC and AC images), crista terminalis/atrial-fibrillation uptake, surgical-adhesive false positives, and false negatives after adequate antibiotics.
- The 2015 ESC guideline accepted perivalvular FDG uptake as a major criterion only when the prosthesis had been implanted more than 3 months earlier; later data showed recent implantation did not predict false positives, and the 2023 ESC guideline counts intense focal or heterogeneous periprosthetic uptake as a major criterion irrespective of the interval from surgery.
Sources: Mahmood et al. J Nucl Cardiol 2019 · 26:958-70 (PMID 28913626) · Mahmood et al. J Nucl Cardiol 2019 (PMID 28913626) · Swart et al. Circulation 2018 (PMID 30018167) · Pizzi et al. Circulation 2015 · 132:1113-26 (PMID 26276890) · Saby et al. J Am Coll Cardiol 2013 · 61:2374-82 (PMID 23583251)
- Fever seven weeks after valve surgery · Advanced