Table 1: Characteristics of patients according to empiric therapy: Patients who received Carbapenem (CRB) vs. patients treated with Non-CRB.

Characteristic CRBa
N = 145
Non-CRBb
N = 402
P value
Age 80 (75-85) 83 (76.2-87) 0.006
Male sex 68 (46.9%) 175 (43.5%) 0.50
ESBL-E colonization recognized at admission 119 (82.1%) 100 (24.9%) < 0.001
Charlson index score 6 (5-7) 6 (5-7) 0.45
Pitt severity score 1 (0-2) 1 (0-2) 0.28
Site of sepsis origin < 0.001
UTI 82 (56.5%) 113 (28.1%)
RTI 34 (23.4%) 219 (54.5%)
Other 18 (12.4%) 32 (8.0%)
Unknown 11 (7.6%) 38 (9.4%)
Aetiology of sepsis < 0.001
Identified, ESBL-E 58 (40%)c 97 (24.1%)c
Escherichia coli 42 67
Klebsiella pneumoniae 16 29
Identified, not ESBL-E 51 (35.2%) 134 (33.3%)
Escherichia coli 23 41
Unidentified 36 (24.8%) 171 (42.5%)
Healthcare-associated infection 118 (81.4%) 344 (85.6%) 0.23
ESBL-E bacteraemia 6 (4.1%) 18 (4.5%) 1.00

Data are median (interquartile range), or number (%).

CRB: carbapenem; ESBL-E: extended-spectrum β-lactamase-producing Enterobacteriaceae; RTI: respiratory tract infection; UTI: urinary tract infection.

aErtapenem 130, imipenem 13, meropenem 2 patients; in addition to CRB 1 patient received azithromycin, 1 flucloxacillin, 1 vancomycin; bβ-lactam-β-lactamase inhibitor combinations (BLI) were prescribed empirically in 266 patients: amoxicillin-clavulanate in 198, piperacillin-tazobactam in 68. In addition to BLI, 5 patients received azithromycin, 2 vancomycin, 1, linezolid, 1 fluoroquinolone, 1 gentamicin. Other antibiotics with Gram-negative activity were prescribed empirically in 136 patients: fluoroquinolone alone in 84, fluoroquinolone in combination with other antibiotics in 26 (anti-Gram-positive in 20, metronidazole in 6), gentamicin in 11, trimethoprim-sulfamethoxazole in 2, other antibiotic combinations in 13.

cOne patient had sepsis caused by both ESBL-E. coli and ESBL-K. pneumoniae.