Infective endocarditis: inadequate antibiotic therapy reduced the chance of positive blood cultures.

Clinical bottom line (level 4)

  1. Patients suspected of having bacterial endocarditis partially treated with antibiotics were less likely to have positive blood cultures (NNH = 18 at 3 weeks) .
Werner et al: Journal of the American Medical Association 1967; 202: 199-203
Expires July 2003

The study

Case-control study with objective outcomes, not adjusted for confounding factors, not validated in an independent set of patients.
Setting: university hospital, USA, 1944 to 1960

178 patients (aged ?, ?% male) proven bacterial endocarditis, diagnosed by isolation of Streptococcus from at least one blood culture and a clinical syndrome consistent with bacterial endocarditis

Excluded if
  • relapse of endocarditis


    • Control Group: (n = 129, 129 analysed): no antibiotic therapy during the two weeks prior to blood culture
    Experimental Group: (n = 49, 49 analysed): oral or parenteral antimicrobial agent (sulphonamides, penicillin in small doses, erythromycin, streptomycin, tetracycline derivatives and chloramphenicol) within two weeks prior to blood culture (therapy was not adequate for treatment of bacterial endocarditis)

    100% followed for 3 weeks
    Outcome notes:
    • positive blood cultures : blood cultures were prepared by inoculating 5 ml into each of two bottles (anaerobic and aerobic) and were intubated for up to three weeks.

    The evidence

    Outcome Time to outcome CEREERRRR
    (95% CI)    		ARR
    (95% CI)    		NNH
    (95% CI)
    positive blood cultures 3 weeks 472
    (96.9%)    		 191
    (91.4%)    		 6%
    (1% to 10%)    		 5.53%
    (1.43% to 9.63%)    		 18
    (10 to 70)

  • It is not clear exactly how many blood samples were taken from each patients. There were 487 cultures in the control group and 209 in the antibiotic group.
  • Streptococci were isolated from the first blood culture in 96% of the cases and from one of the first two blood cultures in 98%.
  • The 33 negative blood cultures occurred in 13 cases of endocarditis (10 viridans group and 3 enterococcus). In seven of these cases, the patients had received antimicrobial agent within 14 days. The mean number of blood cultures required to obtain a positive sample in these cases was 2.47.

    • Comments

    1. Numbers were hugely increased by the many blood cultures used- may make the antibiotics appear to reduce chance of detection more than they really do.

    Citation

    1. Werner AS, et al: Studies of the bacteremia of bacterial endocarditis. Journal of the American Medical Association 1967; 202: 199-203
    Search Terms: blood cultures and endocarditis
    Contributor: Sumit Dhingra, Carl Heneghan and Chris Ball, July 2000
    Reviewer:

    Clinical Question.
    Patient suspected endocarditis
    Intervention or Exposure pre-treatment with antibiotics
    Outcome reduced chance of positive cultures