Rottier et al reported 2 models for predicting bacteremia caused by a third-generation cephalosporin-resistant Enterobacteria that should be treated with a carbapenem. One model is for community-acquired infections and one for hospital-acquired. The authors are from multiple institutions in The Netherlands.
Patient selection: age >= 18 years of age with sepsis due to Enterobacteriaceae
Parameters:
(1) renal disease
(2) prior identification of a third generation resistant Enterobacteria during the past year
(3) any solid malignancy
(4) signs of hypoperfusion at infection onset
(5) surgical procedure in the past 30 days
(6) central venous catheter at infection onset
(7) use of cephalosporins in the past 2 months
(8) length of hospital stay in days prior to infection
(9) suspected source of infection
Parameter |
Finding |
Beta Coefficient |
Points |
renal disease |
no |
0 |
0 |
|
yes |
1.372 |
120 |
prior ID |
no |
0 |
0 |
|
yes |
1.353 |
120 |
any solid malignancy |
no |
0 |
0 |
|
yes |
0.722 |
80 |
signs of hypoperfusion |
no |
0 |
0 |
|
yes |
0.509 |
40 |
surgical procedure |
no |
0 |
0 |
|
yes |
0.444 |
40 |
central venous catheter |
no |
0 |
0 |
|
yes |
0.420 |
40 |
cephalosporin use |
no |
0 |
0 |
|
yes |
0.415 |
40 |
length of hospital stay |
|
0.011 * (days) |
(days) |
suspected source |
lower respiratory tract |
-1.729 |
-160 |
|
other |
0 |
0 |
total score =
= SUM(points for all 9 parameters)
X =
= SUM(beta-coefficients) - 5.807
probability of resistant bacteria causing bacteremia =
= 1/(1 + EXP((-1) * X))
Interpretation:
• minimum score: -160
• maximum score: around 500 (depends on length of hospital stay)
• A score >= 110 is associated with a 3.1% probability of resistant bacteremia. This cutoff would miss about 19% of patients with a resistant infection.
Specialty: Infectious Diseases