Current issues of ACP Journal Club are published in Annals of Internal Medicine


Clinical risk factors identified patients at risk for bacteremia caused by staphylococci

ACP J Club. 1993 Nov-Dec;119:79. doi:10.7326/ACPJC-1993-119-3-079

Source Citation

Leibovici L, Gransden WR, Eykyn SJ, et al. Clinical index to predict bacteraemia caused by staphylococci. J Intern Med. 1993 Jul;234:83-9.



To identify risk factors for bacteremia caused by Staphylococcusaureus or coagulase-negative staphylococci in patients with bacteremia.


A cohort study to construct and validate clinical indices for predicting bacteremia caused by staphylococci.


A teaching hospital in Israel (index construction) and in the United Kingdom (validation).


Hospitalized patients with bacteremia were identified from daily laboratory record review. 1410 episodes of bacteremia were used to construct the index and 1040 episodes to validate it. Coagulase-negative staphylococci were considered to be significant if the patient was septic and if ≥ 2 blood culture sets were positive or a blood culture and a culture from another source showed the same microorganism. S. aureus bacteremia was defined as ≥ 1 positive blood cultures in a patient with a positive culture from an infected site or in a patient with sepsis.

Description of tests and diagnostic standard

Chart audit was used to collect patient age, sex, functional status, underlying disorders, presumptive focus of infection, temperature, leukocyte count, and hospital or community acquisition. Blood culture results were used as the diagnostic standard.

Main outcome measures

Risk factors associated with bacteremia caused by staphylococci.

Main results

Bacteremia was caused by S. aureus in 14% of derivation-set cases and 16% of the validation-set cases and by coagulase-negative staphylococci in 5% of derivation-set cases and 9% of the validation-set cases. Stepwise regression analysis showed that a high-risk site of infection (bone, intravenous catheter, skin, joint, burn, or decubitus ulcer), a focus other than the urinary tract or abdomen, hemodialysis, infection acquired in the orthopedic department, and intravenous drug abuse ( P for all ≥ 0.04) were associated with bacteremia caused by S. aureus. The risk factors for bacteremia caused by coagulase-negative staphylococci were site of infection (peripheral or central intravenous catheter), preterm neonate, presence of a central catheter, hypothermia, and leukopenia ( P for all ≥ 0.02). The prediction rules identified low- (< 3% incidence), intermediate- (10% to 25%), and high-risk (> 30%) groups for S. aureus and coagulase-negative staphylococci in both sets.


Clinical risk factors for patients at risk for bacteremia caused by Staphylococcus aureus were a high-risk site of infection, a focus other than the urinary tract or abdomen, hemodialysis, infection acquired in the orthopedic department, and intravenous drug use. Clinical risk factors for coagulase-negative staphylococci were an infected intravenous catheter, preterm neonate, presence of a central catheter, hypothermia, and leukopenia.

Source of funding: Not stated

For article reprint: Dr. L. Leibovici, Department of Medicine E, Beilinson Medical Centre, Petah Tiqva 49100, Israel. FAX 972-3-937-6505.


Clinicians need to know which settings facilitate acquisition of and which patients are at risk for staphylococcal bacteremia. S.aureus is a particularly virulent pathogen and the early empiric use of an appropriate penicillinase-resistant penicillin or a first-generation cephalosporin (or, if methicillin-resistant S. aureus is likely, vancomycin) is desirable. Coagulase-negative staphylococci are less virulent but are more likely to cause bacteremia in persons with poor immune defenses, and early empiric therapy with vancomycin is recommended.

The study by Leibovici and colleagues quantifies risk factors for patients at risk for staphylococcal bacteremia. Although the factors that they documented for these 2 conditions are not new or surprising, this study is important. The results should make clinicians more alert to the patients at risk, and they will provide clinical educators with a powerful instructional tool.

This study shows the value of including a "validation group" in risk-factor analysis. The "construction group" in Israel included no intravenous drug abusers. This risk factor was, however, strongly displayed in the United Kingdom, as it doubtlessly would have been in the United States. It is possible that additional risk factors might be detected in other site-specific or population-specific samples.

The authors self-critically speculate if the "denominator group" should have been persons suspected, rather than those proven, to have bacteremia. I doubt, however, that this would have changed the risk factors they identified. The invasive nature of S. aureus and the intravascular focus of infections caused by coagulase-negative staphylococci make them disproportionately likely to be associated with bacteremia.

Stephen R. Jones, MD
Emanuel and Good Samaritan HospitalPortland, Oregon, USA