Infective endocarditis: splinter haemorrhages were not helpful.

Clinical bottom line (level 4)

  1. Splinter haemorrhages were common (occurring in 10% of medical patients).
  2. No patients with splinter haemorrhages in this study had infective endocarditis.
Kilpatrick et al: Archives of Internal Medicine 1965; 115: 730-735
Expires July 2003

The study

Setting: acute hospital, USA

574 patients (aged range 13 to 90 years; mean 57, 55% female) admitted to the medical service

Excluded if
  • died
  • no records available



    • ?independent unblinded reference standard, applied in all patients from a consecutive appropriate spectrum.
    Reference standard:
    • clinical findings and positive blood cultures
    Diagnostic test: splinter haemorrhages- homogeneous linear subungual streak of red, brown, purple or black colour, the long axis of which lies in the long axis of the finger

    The evidence


    differential diagnosis number of patients prevalence
    (95% CI)
    splinter haemorrhage 59 10%
    (7.8% to 13%)

    • Three patients had positive blood cultures- none was found to be infective endocarditis (95% CI: 0.0% to 5.0%).
    • Peritoneal dialysis patients accounted for 8/59 (13.6%).

    Citation

    1. Kilpatrick ZM, et al: Splinter hemorrhages- their clinical significance. Archives of Internal Medicine 1965; 115: 730-735
    Search Terms: endocarditis, splinter hemorrhages
    Contributor: Carl Heneghan, Sumit Dhingra and Chris Ball, July 2000
    Reviewer:

    Clinical Question.
    Patient suspected endocarditis
    Intervention or Exposure splinter haemorrhages
    Outcome endocarditis, positive blood cultures