What diagnostic tests confirm loosening or infection of a knee prosthesis?
Executive summary
For a painful or failing knee prosthesis, no single "perfect" test confirms both loosening and infection; instead clinicians rely on a combination of history/exam, blood tests, joint aspiration with microbiology (the reference standard for infection), conventional and advanced imaging, and targeted nuclear medicine studies to separate aseptic loosening from periprosthetic joint infection (PJI) [1] [2] [3]. Aspiration with culture is considered the most accurate preoperative test for infection despite imperfect sensitivity, while radiographs, CT/DECT, MRI, SPECT/CT and labeled leukocyte scans each contribute complementary information about loosening versus infection and have variable accuracy by modality and by knee versus hip [1] [4] [5] [6].
1. Clinical and laboratory gateway: why serology and aspiration come first
Initial steps focus on ruling infection in or out because treatment pathways diverge dramatically; serologic markers—C‑reactive protein (CRP) and erythrocyte sedimentation rate (ESR)—are routine screening tests, and when abnormalities or clinical suspicion exist the next essential test is aspiration of joint fluid for cell count and culture, which remains the preoperative reference standard despite variable sensitivity and occasional false negatives or positives [2] [1] [4]. Consensus guidance and multiple reviews stress serology followed by aspiration as the practical first-line algorithm: if infection cannot be confidently excluded, plans for revision must account for PJI [7] [3].
2. Plain radiographs and CT: the structural story of loosening
Conventional radiographs are the basic, first-line imaging for suspected aseptic loosening because they reveal radiolucent lines, component migration, osteolysis or mechanical failure; when radiographs are equivocal, cross‑sectional imaging—CT or dual‑energy CT (DECT)—provides clearer visualization of bone–implant interfaces and DECT has shown improved performance and reproducibility over single‑energy CT and radiography in recent comparative studies [3] [5]. Radiography alone cannot reliably exclude infection and is often insensitive to early loosening, so CT/DECT is used to better define component position, bone loss and subtle lucencies that suggest loosening [6] [5].
3. MRI and advanced MRI signs: soft tissue and synovitis clues
MRI—now more feasible with metal‑artifact reduction—offers higher sensitivity than radiographs for component loosening and can detect synovitis patterns (for example lamellated hyperintense synovitis) that correlate with infection, though specificity varies and image quality can be limited by implants [6]. Multiple studies included in comparative meta‑analyses find MRI and SPECT/CT among the modalities with higher diagnostic accuracy for aseptic loosening, but evidence certainty is often low and results differ by study and prosthesis type [5].
4. Nuclear medicine: bone scans, SPECT/CT, labeled WBC and PET—sensitivity versus specificity tradeoffs
Technetium bone scintigraphy and three‑phase bone scans are highly sensitive for a painful prosthesis and useful as screening tests, but they have poor specificity—especially in knees where postoperative remodeling and variable uptake persist—so a negative scan is more helpful to exclude problems than a positive scan is to prove infection [1] [8] [9]. Hybrid SPECT/CT improves localization and has shown encouraging accuracy for differentiating loosening from infection (better in hips than knees), while labeled leukocyte (In‑111 or Tc‑99m WBC) scans combined with marrow (sulfur colloid) imaging increase specificity for infection by distinguishing marrow activity from infection‑related WBC uptake [4] [10] [11]. FDG‑PET and pattern analysis can help but have variable positive predictive values in knees and are not universally definitive [12] [13].
5. No single gold standard—how clinicians synthesize tests into a diagnosis
Because every modality has limits—radiographs miss early loosening, serology and aspiration can be false‑negative, and nuclear/imaging patterns overlap between inflammation and infection—diagnosis depends on an algorithmic synthesis: rule out infection with serology and aspiration (culture is the preoperative reference), use radiographs/CT/DECT or MRI to define loosening and bone loss, and deploy SPECT/CT or labeled WBC imaging when noninvasive differentiation remains uncertain, recognizing that diagnostic accuracy may be lower in knees than hips and that study heterogeneity and commercial interests can influence reported performance [7] [4] [5]. When sources diverge, the common thread is a multimodal approach led by aspiration/culture for infection confirmation and structural imaging for loosening [1] [3].