Background
Methods
Study population
Selection of cases and inclusion criteria
Microbiologic data
Management of antibiotic therapy
Data collection
Statistical analysis
Results
Epidemiologic and clinical characteristics
Characteristic | De-escalation (n = 110) | No de-escalation (n = 96) | Escalation (n = 65) | No change (n = 31) |
---|---|---|---|---|
Male sex, n (%) | 61 (55) | 56 (58) | 35 (54) | 21 (68) |
Age, years, median (IQR) | 61 (47–72) | 66 (51–75)a
| 63 (47–75) | 70 (58–77) |
Comorbidities | ||||
Fatal underlying disease | 30 (27) | 32 (33) | 21 (32) | 11 (35) |
Cancer, n (%) | 37 (34) | 36 (38) | 23 (35) | 13 (42) |
Diabetes, n (%) | 17 (15) | 15 (16) | 9 (14) | 6 (19) |
Time since initial surgery, days, median (IQR) | 7 (5–12) | 7 (4–10) | 6 (3–9) | 8 (5–10) |
Antibiotic therapy before reoperation, n (%) | 73 (66) | 68 (71) | 47 (72) | 21 (68) |
Broad-spectrum interim antibiotic, n (%) | 34 (31) | 37 (39) | 27 (42) | 10 (32) |
Intraoperative diagnosis | ||||
Anastomotic leakage, n (%) | 45 (41) | 27 (28) | 19 (29) | 8 (26) |
Perforation or ischemia, n (%) | 33 (30) | 36 (38) | 20 (31) | 16 (52)a#
|
Purulent collection, n (%) | 19 (17) | 17 (18) | 13 (20) | 4 (13) |
No cause, n (%) | 19 (17) | 20 (21) | 15 (23) | 5 (16) |
Contamination below transverse mesocolon, n (%) | 82 (75) | 74 (77) | 50 (77) | 24 (77) |
Characteristics at the time of ICU admission | ||||
Bacteremia, n (%) | 26 (24) | 17 (18) | 14 (22) | 3 (10) |
SAPS II score, median (IQR) | 45 (34–54) | 47 (35–57) | 44 (34–56) | 51 (42–61)a#
|
SOFA score, median (IQR) | 7 (4–9) | 8 (4–10) | 7 (4–9) | 9 (6–10) |
Hemodynamic failureb, n (%) | 65 (59) | 65 (68) | 41 (63) | 24 (77) |
Respiratory failureb, n (%) | 54 (49) | 40 (42) | 26 (40) | 14 (45) |
Renal failureb, n (%) | 21 (19) | 18 (19) | 13 (20) | 5 (16) |
Microbiologic analysis
Microorganisms | De-escalation | No de-escalation | Escalation | No change |
---|---|---|---|---|
Total number of multidrug-resistant bacteria, n (%) | 29 (9) | 74 (24)a
| 58 (28) | 16 (16)b
|
Gram-positive bacteria, n (%) | 15 (5) | 39 (13)a
| 33 (16) | 6 (6)b
|
Enterococci, n (%) | 3 (1) | 9 (3) | 7 (3) | 2 (2) |
Staphylococci, n (%) | 12 (4) | 29 (9) | 25 (12) | 4 (4) |
Staphylococcus aureus, n (%) | – | 8 (3) | 6 (3) | 2 (2) |
Gram-negative bacteria, n (%) | 14 (5) | 35 (11)a
| 25 (12) | 10 (10) |
Enterobacteriaceae, n (%) | 10 (3) | 24 (8) | 18 (9) | 6 (6) |
Escherichia coli, n (%) | 1 (0) | 11 (4) | 8 (4) | 3 (3) |
Enterobacter spp., n (%) | 5 (2) | 9 (3) | 8 (4) | 1 (1) |
Nonfermenting Gram-negative bacilli, n (%) | 3 (1) | 11 (4) | 7 (3) | 4 (4) |
Pseudomonas spp., n (%) | 2 (1) | 6 (2) | 4 (2) | 2 (2) |
Total number of cultured bacteria, n
| 307 | 311 | 210 | 101 |
Anti-infective therapy
De-escalation (n = 110) | No de-escalation (n = 96) | Escalation (n = 65) | No change (n = 31) | |
---|---|---|---|---|
Empiric antibiotic therapy | ||||
Monotherapy, n (%) | 13 (12) | 32 (33)a
| 20 (31) | 12 (29) |
Combination of two drugs, n (%) | 40 (36) | 34 (35) | 26 (40) | 8 (26) |
Combination of three drugs or more, n (%) | 57 (49) | 30 (33)b
| 19 (29) | 11 (35) |
Carbapenem, n (%) | 35 (32) | 15 (16)a
| 10 (15) | 5 (16) |
Piperacillin-tazobactam, n (%) | 67 (61) | 60 (63) | 40 (62) | 20 (65) |
Vancomycin, n (%) | 57 (52) | 23 (24)a
| 15 (23) | 8 (26) |
Aminoglycosides, n (%) | 59 (54) | 33 (34)a
| 27 (42) | 6 (19)b
|
Fluoroquinolones, n (%) | 6 (5) | 11 (11) | 10 (15) | 1 (3) |
Antifungal therapy, n (%) | 47 (43) | 23 (24)a
| 11 (17) | 12 (39)b
|
Azoles, n (%) | 41 (37) | 20 (21)a
| 9 (14) | 11 (35)b
|
Adequate empiric therapy, n (%) | 100 (91) | 37 (39)a
| 9 (14) | 28 (90)c
|
Reevaluation of antibiotic therapy | ||||
Discontinuation of carbapenemsd, n (%) | 27/35 (77) | 4/15 (27) | 4/10 (40) | – |
Discontinuation of piperacillin-tazobactamd, n (%) | 50/67 (75) | 25/60 (42) | 25/40 (63) | – |
Discontinuation of vancomycind, n (%) | 46/57 (81) | 6/23 (26) | 6/15 (40) | – |
Discontinuation of aminoglycosidesd, n (%) | 54/59 (92) | 21/33 (64) | 21/27 (78) | – |
Discontinuation of fluoroquinolonesd, n (%) | 2/6 (33) | 6/11 (55) | 6/10 (60) | – |
Discontinuation of antifungal agent d, n (%) | 23/47 (49) | 4/23 (17) | 4/11 (36) | – |
Withdrawal of at least one agent, n (%) | 110 (100) | 42 (47)a
| 42 (65) | – |
Narrowing spectrum, n (%) | 74 (67) | 18 (19)a
| 18 (28) | – |
Switch to monotherapy, n (%) | 54 (49) | 7 (7)a
| 7 (11) | – |
Interruption of unnecessary agent, n (%) | 78 (71) | 20 (21)a
| 20 (31) | – |
Clinical changes between days 0 and 3 | ||||
Changes in SOFA score, median (IQR) | −2 (−4 to -1) | −2 (−4 to 0) | −2 (−4 to 0) | −2 (−3 to 0) |
Decreased SOFA score, n (%) | 69 (63) | 57 (59) | 38 (58) | 19 (61) |
Decreased temperature, n (%) | 69 (63) | 64 (67) | 41 (63) | 23 (74) |
Decreased WBC, n (%) | 38 (35) | 32 (33) | 23 (35) | 9 (29) |
Clinical improvement at day 3, n (%) | 17 (15) | 18 (19) | 14 (22) | 4 (13) |
Univariate analysis | Multivariate analysis | |||||
---|---|---|---|---|---|---|
De-escalation (n = 110) | No de-escalation (n = 96) | Unadjusted odds ratio (95 % CI) |
p valuea
| Adjusted odds ratio (95 % CI) |
p value | |
At time of admission | ||||||
Age, years | 61 (47–72) | 66 (51–75) | 0.98 (0.97–0.99) | 0.049 | – | – |
Emergency surgery | 37 (34) | 44 (46) | 0.60 (0.34–1.05) | 0.087 | ||
Anastomotic leakage | 45 (41) | 27 (28) | 1.76 (0.98–3.17) | 0.058 | – | – |
Empiric antibiotic monotherapy | 13 (12) | 32 (33) | 0.26 (0.13–0.54) | <0.01 | – | – |
Initial use of carbapenems | 35 (32) | 15 (16) | 2.52 (1.27–4.98) | 0.009 | 2.64 (1.01–6.91) | 0.047 |
Initial use of vancomycin | 57 (52) | 23 (24) | 3.41 (1.87–6.21) | <0.0001 | 3.39 (1.46–7.87) | 0.004 |
Initial use of aminoglycosides | 59 (54) | 33 (34) | 2.20 (1.25–3.88) | 0.007 | 2.31 (1.08–4.94) | 0.031 |
Initial use of fluoroquinolones | 6 (5) | 11 (11) | 0.44 (0.15–1.22) | 0.134 | ||
Initial use of antifungal agents | 47 (43) | 23 (24) | 2.36 (1.29–4.32) | 0.005 | – | – |
At day 3 | ||||||
Presence of Enterococcus faecium
| 8 (7) | 18 (19) | 0.33 (0.14–0.82) | 0.019 | – | – |
Presence of streptococci | 31 (28) | 17 (18) | 1.82 (0.93–3.55) | 0.098 | – | – |
Presence of staphylococci | 25 (23) | 35 (36) | 0.51 (0.27–0.94) | 0.032 | – | – |
Presence of NFGNB | 8 (7) | 22 (23) | 0.26 (0.11–0.62) | 0.002 | 0.28 (0.09–0.89) | 0.031 |
Presence of MDR strains | 23 (21) | 54 (56) | 0.20 (0.11– 0.37) | <0.0001 | 0.21 (0.09–0.52) | 0.0007 |
Presence of fungi | 30 (27) | 41 (43) | 0.50 (0.28–0.90) | 0.027 | – | – |
Adequate empiric therapy | 100 (91) | 37 (39) | 15.95 (7.38–34.40) | <0.0001 | 9.60 (4.02–22.97) | <0.0001 |
Clinical evaluation following de-escalation
De-escalation (n = 110) | No de-escalation (n = 96) | Escalation (n = 65) | No change (n = 31) | |
---|---|---|---|---|
Definitive anti-infective therapy | ||||
Monotherapy, n (%) | 64 (58) | 21 (22) | 9 (14) | 12(29)a
|
Combination of two drugs, n (%) | 33 (30) | 39 (41) | 31 (48) | 8 (26)b
|
Combination of three drugs or more, n (%) | 13 (12) | 36 (38) | 25 (38) | 11 (35) |
Use of carbapenems, n (%) | 5 (5) | 23 (24)a
| 18 (28) | 5 (16) |
Use of piperacillin-tazobactam, n (%) | 28 (25) | 39 (41)b
| 19 (29) | 20 (65)a
|
Use of vancomycin, n (%) | 11 (10) | 41 (43)a
| 33 (51) | 8 (26)b
|
Use of antifungals, n (%) | 29 (26) | 43 (45)a
| 31 (48) | 12 (39) |
Use of azoles, n (%) | 28 (25) | 39 (41)b
| 28 (43) | 11 (35) |
Use of echinocandins, n (%) | 1 (1) | 3 (3) | 3 (5) | – |
Duration of anti-infective therapy, days, median (IQR) | 10 (10–14) | 10 (10–14) | 10 (10–14) | 10 (10–14) |
Clinical changes between days 3 and 7 | ||||
Number of cases at day 7 | 91 | 75 | 51 | 24 |
Changes in SOFA scorec, median (IQR) | −1 (−3 to 0) | −2 (−4 to 0) | −2 (−4 to 0) | −2 (−4 to 0) |
Decreased SOFA scorec, n (%) | 57 (63) | 48 (64) | 35 (54) | 13 (42) |
Decreased temperaturec, n (%) | 38 (42) | 36 (48) | 27 (42) | 9 (29) |
Decreased WBCc, n (%) | 31 (34) | 17 (23) | 12 (18) | 5 (16) |
Clinical improvement at day 7c, n (%) | 16 (18) | 9 (12) | 9 (14) | – |
Discharge between days 3 and 7, n (%) | 20 (22) | 15 (20) | 10 (20) | 5 (21) |
Death between days 3 and 7, n (%) | 1 (1) | 3 (4) | 2 (4) | 1 (4) |
Medical complications | 12 (12) | 14 (16) | 6 (10) | 8 (29) |
Surgical complications | 26 (25) | 21 (24) | 10 (17) | 11 (39)b
|
Reoperation, n (%) | 38 (35) | 35 (36) | 19 (29) | 16 (51)b
|
Time to reoperation, days, median (IQR) | 6 (5–9) | 6 (5–8) | 5 (4–8) | 7 (6–10) |
Superinfection on subsequent reoperationd, n (%) | 23 (61) | 23 (66) | 13 (68) | 10 (63) |
Emergence of MDR strainsd, n (%) | 21 (55) | 20 (57) | 11 (58) | 9 (56) |
Emergence of ESBL Enterobacteriaceae
d, n (%) | 5 (13) | 5 (14) | 3 (16) | 2 (13) |
Emergence of MDR NFGNBd, n (%) | 9 (24) | 5 (14) | 2 (11) | 3 (19) |
Emergence of MRSAd, n (%) | 9 (24) | 8 (23) | 6 (32) | 2 (13) |
Duration of mechanical ventilatione, days, median (IQR) | 7 (3–13) | 7 (3–11) | 7 (2–10) | 7 (3–15) |
ICU length of staye, days, median (IQR) | 12 (8–20) | 12 (8–21) | 12 (8–21) | 14 (5–23) |
Survival at day 28, n (%) | 91 (83) | 72 (75) | 51 (78) | 21 (68) |
ICU mortality rate, n (%) | 23 (21) | 32 (33)b
| 18 (28) | 14 (45) |
Hospital mortality rate, n (%) | 25 (23) | 33 (34) | 19 (29) | 14 (45) |
Univariate analysis | Multivariate analysis | |||||
---|---|---|---|---|---|---|
Death at day 28 (n = 43) | Survival at day 28 (n = 163) | Unadjusted hazard ratio (95 % CI) |
p valuea
| Adjusted hazard ratio (95 % CI) |
p value | |
Age, years | 69 (56–78) | 62 (46–72) | 1.022 (1.003–1.042) | 0.023 | 1.034 (1.011–1.059) | 0.004 |
Emergency surgery | 24 (56) | 57 (35) | 2.072 (1.135–3.783) | 0.015 | – | – |
Surgery below the mesocolon | 28 (65) | 128 (79) | 0.561 (0.299–1.050) | 0.075 | 0.427 (0.215–0.848) | 0.015 |
SOFA score | 10 (7–11) | 7 (4–9) | 1.261 (1.153–1.380) | <0.0001 | 1.291 (1.168–1.427) | <0.0001 |
SAPS II score | 52 (45–61) | 44 (32–53) | 1.039 (1.019–1.060) | <0.0001 | – | – |
Initial use of piperacillin-tazobactam | 22 (51) | 105 (64) | 0.636 (0.350–1.157) | 0.117 | – | – |
Empiric antifungal therapy | 21 (49) | 49 (30) | 1.980 (1.089–3.601) | 0.029 | – | – |
Other Enterobacteriaceae
b
| 5 (12) | 34 (21) | 0.538 (0.212–1.366) | 0.195 | 0.342 (0.1219–0.961) | 0.0419 |
Presence of Candida spp. | 19 (44) | 52 (32) | 1.569 (0.859–2.865) | 0.150 | 2.641 (1.3471–5.179) | 0.0047 |
Decreased SOFA score at day 3 | 17 (40) | 109 (67) | 0.372 (0.202–0.686) | 0.0015 | 0.311 (0.1632–0.593) | 0.0004 |
Antibiotic strategy | 0.189 | |||||
De-escalation | 19 (44) | 91 (56) | 0.488 (0.227–1.051) | 0.566 (0.2503–1.278) | 0.171 | |
No change | 10 (23) | 21 (13) | Reference | Reference | Reference | |
Escalation | 14 (33) | 51 (31) | 0.627 (0.278–1.411) | 0.508 (0.2154–1.198) | 0.122 |
Discussion
Conclusions
Key messages
-
De-escalation is a reasonable option, even in polymicrobial infections such as health care-associated intra-abdominal infections.
-
Multidrug-resistant bacteria and nonfermenting Gram-negative bacilli remain a major obstacle in de-escalation.
-
The key determinants for de-escalation are susceptible microorganisms and adequate empiric therapy.