Efficacy of Tigecycline for Secondary Acinetobacter Bacteremia and Factors Associated with Treatment Failure

Abstract

We describe the clinical outcome of 17 patients with secondary Acinetobacter bacteremia whose isolates had a tigecycline MIC of ≤2 mg/liter and who received tigecycline within 2 days of bacteremia onset. The 14-day mortality rate of the tigecycline cohort was 41.2% (7/17), which was significantly higher than that of those receiving other appropriate antimicrobial agents (13.8%, 9/65; P = 0.018). However, the percentages of end-stage renal disease and congestive heart failure were higher in the tigecycline cohort. The efficacy of tigecycline was contingent upon the illness severity and bacterial species. Tigecycline should be applied cautiously for treatment of Acinetobacter bacteremia.

TEXT

Several recent studies have shown tigecycline to be effective in subjects with secondary bacteremia in complicated skin infections, intra-abdominal infections, and community-acquired pneumonia (1–5). However, the efficacy of tigecycline specifically in patients with secondary Acinetobacter bacteremia is not known. Here, we report the overall 14-day mortality of patients with secondary Acinetobacter bacteremia treated with tigecycline, initiating within 2 days of the onset of bacteremia, and compared it with mortality for those treated appropriately with other antimicrobial agents. We also investigated risk factors for a poor outcome after tigecycline therapy.

This retrospective study was conducted at the Taipei Veterans General Hospital in Taiwan from September 2007 to August 2013. We reviewed the charts of patients with sepsis and a positive blood culture for Acinetobacter baumannii, Acinetobacter nosocomialis, or Acinetobacter pittii. Those patients who received tigecycline as appropriate therapy within 2 days of the onset of bacteremia were included in the study. The clinical data collection, identification of bacterial isolates, and antimicrobial susceptibilities were obtained as described previously (6). The onset of bacteremia was defined as the day when the blood culture that eventually yielded Acinetobacter was drawn. Appropriate antimicrobial therapy was defined as intravenous administration of at least one antimicrobial, to which the bacterium was susceptible, at an appropriate dose within 2 days of the onset of bacteremia. The MIC for tigecycline was determined by the Etest method (bioMérieux). The U.S. Food and Drug Administration (FDA) breakpoint for tigecycline was applied (susceptibility, ≤2 mg/liter).

Continuous variables were reported with the median and interquartile range (IQR) and compared using the Mann-Whitney U test. Categorical variables were expressed as percentages and calculated by Fisher's exact test. A P value of <0.05 was considered significant.

In total, 17 consecutive patients whose Acinetobacter isolates had tigecycline MICs of ≤2 mg/liter were appropriately treated with tigecycline during the 6-year study period (Table 1). All of them received tigecycline within 2 days of bacteremia onset, and 12 (70.6%) had tigecycline administration at the same day of bacteremia onset. The median duration of tigecycline therapy was 12 days (IQR, 7.5 to 18). All treated episodes of bacteremia were hospital acquired, and pneumonia was the most prevalent source (n = 14, 82.4%). A. baumannii accounted for 64.7% (n = 11) of isolates, A. nosocomialis for 29.4% (n = 5), and A. pittii for 5.9% (n = 1). Tigecycline MICs for these isolates were between 0.1 and 2 mg/liter.

TABLE 1.

Description of 17 patients with monomicrobial Acinetobacter calcoaceticus-Acinetobacter baumannii complex bacteremia treated with tigecycline within 2 days of bacteremia onset

No.	Sexa	Age (yr)	APACHE II score	Origin of bacteremiab	Species	TG MIC (mg/liter)c	Duration of TG treatment (days)	Concomitant appropriate antibioticd	Outcome
									Day 14	Day 28
1	M	87	27	Pneumonia	A. nosocomialis	2	20	None	Alive	Alive
2	M	53	33	Pneumonia	A. baumannii	2	15	Ampicillin-sulbactam	Dead	Dead
3	M	76	25	Pneumonia	A. baumannii	2	7	None	Alive	Alive
4	F	76	14	Pneumonia	A. nosocomialis	1.5	12	Sulbactam	Alive	Alive
5	M	83	26	Pneumonia	A. baumannii	1.5	25	None	Alive	Alive
6	M	80	29	Pneumonia	A. baumannii	1.5	12	Levofloxacin	Dead	Dead
7	M	71	29	Pneumonia	A. baumannii	1.5	9	None	Dead	Dead
8	F	63	16	Pneumonia	A. nosocomialis	0.2	8	None	Alive	Alive
9	M	89	38	IAI	A. baumannii	1.5	6	None	Dead	Dead
10	M	62	33	Pneumonia	A. nosocomialis	0.5	31	None	Alive	Alive
11	M	80	33	Pneumonia	A. baumannii	2	4	None	Dead	Dead
12	M	83	32	UTI	A. baumannii	2	14	None	Alive	Dead
13	M	80	22	Pneumonia	A. baumannii	1.5	9	None	Alive	Alive
14	F	45	19	CR-BSI	A. pittii	0.1	8	Ceftazidime	Alive	Alive
15	F	84	23	Pneumonia	A. baumannii	2	19	None	Dead	Dead
16	M	83	26	Pneumonia	A. nosocomialis	0.1	17	None	Alive	Alive
17	F	61	31	Pneumonia	A. baumannii	1.5	4	None	Dead	Dead

^aM, male; F, female.

^bIAI, intra-abdominal infection; UTI, urinary tract infection; CR-BSI, catheter-related bloodstream infection.

^cTigecycline (TG) MIC was determined by Etest.

^dConcomitant appropriate antibiotic was defined as intravenous antibiotic, to which the bacterium was susceptible, administered at an appropriate dose within 2 days of the bacteremia onset.

To evaluate the efficacy of tigecycline, we compared the 14-day mortality rates of patients treated with tigecycline to those appropriately treated with other antimicrobial agents, whose data were extracted from a previous study (6). Of those patients treated with other antimicrobial agents, three or four cases were matched to a tigecycline case according to sex, age (±15 years), and acute physiology and chronic health evaluation II (APACHE II) score (±10). The comparison group consisted of 65 patients, while the tigecycline group consisted of 17 cases. Antipseudomonal carbapenem (27/65, 41.5%) and sulbactam (18/65, 27.7%) were the most commonly used antibacterial agents in the comparison group (see Table S1 in the supplemental material). Each group displayed similar demographic characteristics, comorbid conditions, and infection sources. Notably, in the comparison group, fewer patients had hospital-acquired pneumonia (53.8% [35/65] versus 82.4% [14/17]; P = 0.05) or ventilator-associated pneumonia (27.7% [18/65] versus 58.8% [10/17]; P = 0.023), and fewer patients had comorbidities of end-stage renal disease and congestive heart failure (Table 2).

TABLE 2.

Clinical characteristics in the tigecycline and comparison groups

Characteristic	Tigecycline group (n = 17)	Comparison group (n = 65)	P value
Demographic characteristic
Age (median [IQR]) (yr)	80 (62.5–83)	76 (68.5–82)	0.693
Male (no. [%])	12 (70.6)	50 (76.9)	0.752
APACHE II score within 24 h of bacteremia onset (median [IQR])	27 (22.5–32.5)	24 (21.5–30)	0.375
APACHE II score > 25 (no. [%])	11 (64.7)	29 (44.6)	0.177
Comorbid condition (no. [%])
Type 2 diabetes mellitus	5 (29.4)	18 (27.7)	>0.999
Chronic obstructive pulmonary disease	5 (29.4)	8 (12.3)	0.130
Coronary artery disease	4 (23.5)	6 (9.2)	0.204
Congestive heart failure	5 (29.4)	4 (6.2)	0.016
Renal impairment	5 (29.4)	11 (16.9)	0.304
End-stage renal disease	4 (23.5)	1 (1.5)	0.006
Cerebrovascular accident	3 (17.6)	14 (21.5)	>0.999
Collagen vascular disease	0 (0.0)	1 (1.5)	>0.999
Malignancy	7 (41.2)	25 (38.5)	>0.999
Median hospital duration prior to bacteremia, days (IQR)	22 (15–59.5)	14 (6–36)	0.081
Mechanical ventilator use at bacteremia onset (no. [%])	13 (76.5)	32 (49.2)	0.057
Acquired in intensive care unit (no. [%])	13 (76.5)	53 (81.5)	0.732
Infection source (no. [%])
Respiratory tract	14 (82.4)	39 (60.0)	0.098
Urinary tract	1 (5.9)	6 (9.2)	>0.999
Intra-abdominal	1 (5.9)	4 (6.2)	>0.999
Skin and soft tissue	0 (0.0)	1 (1.5)	>0.999
Catheter related	1 (5.9)	2 (3.1)	>0.999
Central nervous system	0 (0.0)	2 (3.1)	>0.999
Primary bacteremia	0 (0.0)	11 (16.9)	0.109
Isolates with Acinetobacter baumannii (no. [%])	11 (64.7)	65 (100.0)	<0.001
Susceptible antimicrobial agent at the day of bacteremia onset (no. [%])	12 (70.6)	26 (40.0)	0.031
Dual appropriate antimicrobial agents (no. [%])a	4 (23.5)	4 (6.2)	0.054
14-day mortality rate (no. [%])	7 (41.2)	9 (13.8)	0.018

^aDual appropriate antimicrobial agents was defined as ≥2 intravenous antibiotics, to which the bacterium was susceptible, at an appropriate dose within 2 days of bacteremia onset.

The tigecycline group had a significantly lower rate of A. baumannii isolates (64.7% versus 100%; P < 0.001) and a higher percentage of patients receiving antimicrobial agents to which the causative agent was susceptible at the day of bacteremia onset (70.6% versus 40.0%; P = 0.031). Despite these favorable conditions, the tigecycline group had a 14-day mortality rate of 41.2% (7/17), which was significantly higher than that of the comparison group (13.8% [9/65]; P = 0.018). The subgroup analysis stratified by APACHE II score also showed a persistent trend toward higher 14-day mortality in the tigecycline group (Table 3). For patients with hospital-acquired pneumonia, the mortality rate was also higher in the tigecycline group (42.9% [6/14] versus 14.3% [5/35]; P = 0.054).

TABLE 3.

Fourteen-day mortality rates for tigecycline and comparison groups, stratified by APACHE II score

APACHE II score	Tigecycline group (n = 17)		Comparison group (n = 65)		P value
	No. of patients	14-day mortality (no. [%])	No. of patients	14-day mortality (no. [%])
≤25	6	1 (16.7)	36	2 (5.6)	0.378
26–30	5	2 (40.0)	16	4 (25.0)	0.598
>30	6	4 (66.7)	13	3 (23.1)	0.129

For the tigecycline group, the clinical characteristics of patients, predisposing factors, microbiologic data, and combination use of other antimicrobial agents to which the causative bacteria were susceptible are shown in Table 4, stratified by 14-day survival. The APACHE II score was significantly related to survival (median, 25.5 [IQR, 18.25 to 28.25] in survivors versus 31 [IQR, 29 to 33] in nonsurvivors; P = 0.031). In addition, 100% of nonsurvivors were infected by A. baumannii, while only 40% of survivors were infected with this strain (P = 0.035).

TABLE 4.

Clinical characteristics of patients with Acinetobacter calcoaceticus-Acinetobacter baumannii complex bacteremia, stratified by 14-day mortality rate

Variable	Tigecycline group (n = 17)			Comparison group (n = 65)
	Survivors (n = 10)	Nonsurvivors (n = 7)	P value	Survivors (n = 56)	Nonsurvivors (n = 9)	P value
Male (no. [%])	7 (70.0)	5 (71.4)	>0.999	42 (75.0)	8 (88.9)	0.446
Age (median [IQR]) (yr)	78 (62.75–83)	80 (61–84)	>0.999	77.5 (70–82)	65 (57–78.5)	0.089
APACHE II score within 24 h of the onset of bacteremia (median [IQR])	25.5 (18.25–28.25)	31 (29–33)	0.031	24 (21–29.75)	30 (25.5–35)	0.047
Duration of tigecycline therapy (median [IQR]) (days)	13 (8–21.25)	9 (4–15)	0.171
Mechanical ventilator use at bacteremia onset (no. [%])	7 (70.0)	6 (85.7)	0.603	27 (48.2)	5 (55.6)	0.733
Acquired in intensive care unit (no. [%])	7 (70.0)	6 (85.7)	0.603	47(83.9)	6 (66.7)	0.349
Isolates with A. baumannii (no. [%])	4 (40.0)	7 (100.0)	0.035	56 (100)	9 (100)
Dual appropriate antimicrobial agent (no. [%])	2 (20.0)	2 (28.6)	>0.999	4 (7.1)	0 (0)	>0.999

Tigecycline is one of the few remaining options to treat multidrug-resistant Acinetobacter. However, our data suggest that tigecycline treatment adversely affects 14-day survival in patients with Acinetobacter infection and secondary bacteremia. We also observed a trend of higher mortality in patients receiving tigecycline for hospital-acquired pneumonia. Although the discrepancy in disease severity may be a sufficient confounder to explain the difference in mortality, subgroup analysis stratified by disease severity (Table 3) still showed a persisted trend of higher mortality in the tigecycline group. The FDA-approved breakpoint of ≤2 mg/liter did not predict favorable outcomes in this setting. This result is in accordance with the warning issued by the FDA (7) about the increased risk of poor outcomes associated with tigecycline, especially in hospital-acquired pneumonia. The bacteriostatic property of tigecycline may be one of the reasons. Compared with studies that favored the use of tigecycline in patients with bacteremia, our study cohort acquired nosocomial Acinetobacter bacteremia and had a greater severity of illness (1–5). For such patients, appropriate and bactericidal antimicrobial therapy was of the utmost importance for preventing a worse outcome (6, 8).

The risk factors associated with worse outcomes in the present study are consistent with other studies (6, 9, 10). Some may argue about whether Acinetobacter bacteremia is an independent risk for mortality or merely a marker of the severity of illness. Secondary bacteremia, specifically with A. baumannii, might be an additional risk factor because of its increased virulence compared to other Acinetobacter species. Patients infected with A. baumannii have a higher risk of mortality than patients infected with other Acinetobacter species, after adjustment for appropriate therapy (9). The mortality of patients that receive appropriate antimicrobial therapy against Acinetobacter bacteremia is also independently affected by disease severity (6). However, we may not have been able to detect other factors with less impact because of the limited number of patients in the present study.

The major strength of our study was the stringent inclusion criteria. The outcome of antimicrobial therapy is related to the timing and adequacy of treatment and disease severity. Therefore, we limited this study to patients with monomicrobial Acinetobacter bacteremia treated with a standard dose of tigecycline within 2 days of a positive blood culture. To minimize the confounding factors, we selected a comparison group from the same hospital with regard to age, sex, and APACHE II score. However, the relatively small case numbers and the retrospective nature limited our study. Use of tigecycline in patients with bacteremia is often deferred, so conducting a prospective controlled clinical trial with an appropriate comparator drug is difficult.

In conclusion, tigecycline is not better than other appropriate antimicrobial therapies for the treatment of secondary Acinetobacter bacteremia. The efficacy of tigecycline for Acinetobacter bacteremia is contingent upon the illness severity and bacterial species. Attending physicians should use tigecycline with caution only after other reasonable therapeutic options have been exhausted.