Description of included patients with BSI due to S. maltophilia
Within the study period, a total of 1091 ICU patients with S. maltophilia BSI were identified and a total of 214 were included for analyses. A flowchart for the study is shown in Fig. 1. The characteristics of the ICU patients with BSI by S. maltophilia are shown in Table 1. The patient population was predominantly male (65.4%), with a median age of 74 years (IQR 57–83). The major reason for admission was infection (51.9%). Prolonged hospital stay and ICU stay before BSI were noted with a median of 24 days (IQR 15–38) and 15.5 days (IQR 9–24), respectively. Composite comorbidities at admission and severity at the BSI onset are illustrated by the CCI being 3 (IQR 2–6) and the APACHE score being 22 (IQR 18–28). The most common source of BSI was from respiratory tract (55.6%), followed by infection of primary origin (34.1%). The 14-day and in-hospital mortality of the included patients with BSI were 24.3% and 52.3%, respectively. The median hospital stay after BSI was 25.5 days (IQR 14–52).
Table 1
Characteristics of ICU patients with BSI by S. maltophilia
Variable | Total (n = 214) |
Demographics | |
Age, years, median (IQR) | 74 (57–83) |
Male, n (%) | 140 (65.4) |
Comorbidities, n (%) | |
Heart failure | 44 (20.6) |
Old cerebrovascular accident | 42 (19.6) |
Chronic obstructive pulmonary disease | 32 (15.0) |
Connective tissue disease | 8 (3.7) |
Diabetes mellitus | 78 (36.4) |
Liver cirrhosis | 17 (7.9) |
Chronic kidney disease | 39 (18.2) |
Charlson Comorbidity Index, median (IQR) | 3 (2–6) |
Setting from which patient was admitted to ICU, n (%) | |
Other hospital | 17 (7.9) |
Ward | 100 (46.7) |
Emergency department | 97 (45.3) |
Hospital stay before BSI onset, days, median (IQR) | 24 (15–38) |
ICU stay before BSI onset, days, median (IQR) | 15.5 (9–24) |
Reason for ICU admission, n (%) | |
Infection | 111 (51.9) |
Respiratory failure | 10 (4.8) |
Cardiovascular disease | 33 (15.4) |
Postoperative monitoring | 25 (11.7) |
Gastrointestinal bleeding | 9 (4.2) |
Cerebrovascular accident | 14 (6.5) |
Trauma | 7 (3.3) |
Others | 5 (2.3) |
Table 1
Characteristics of ICU patients with BSI by S. maltophilia (continued)
Variable | Total (n = 214) | |
Source of BSI, n (%) | | |
Respiratory tract | 119 (55.6) |
Urinary tract | 3 (1.4) |
Catheter related | 11 (5.1) |
Intra-abdominal | 3 (1.4) |
Primary | 73 (34.1) |
Skin and soft tissue | 1 (0.5) |
Disease severity at BSI onset, n (%) | |
APACHE score, median (IQR) | 22 (18–28) |
Acute kidney injury | 51 (23.8) |
Shock | 74 (34.6) |
Mechanical ventilation | 142 (66.4) |
Clinical outcome, n (%) | |
14-day mortality | 52 (24.3) |
In-hospital mortality | 112 (52.3) |
Hospital stay after BSI among survivals, days, median (IQR) | 25.5 (14–52) |
Other causes of ICU admission include 1 case of carbon monoxide poisoning, 1 case of hyperosmolar hyperglycemic state, 2 cases of acute renal failure, and 1 case of spinal shock. The assessment of hospital stay did not include the stay of patients who died during hospitalization. |
S. maltophilia, Stenotrophomonas maltophilia; ICU, intensive care unit; BSI, blood stream infection; IQR, interquartile range |
Antibiotic therapy for BSI due to S. maltophilia
The prescription of antibiotic therapy is summarized in Table 2. After the onset of S. maltophilia BSI, the median number of classes of the antibiotic therapy empirically used in each infected patient was 2 (1–2), and 146 of 214 (68.2%) patients received at least two classes of empiric antibiotics. The most prescribed empiric antibiotic for S. maltophilia BSI was carbapenem (50.9%). Overall, 133 of 214 (62.1%) patients received appropriate antibiotic therapy after BSI. Only 29 of 214 (13.6%) patients received appropriate antibiotics within 48 h of the BSI onset. Among 133 patients with appropriate antibiotic therapy, 62 received levofloxacin, 67 received trimethoprim–sulfamethoxazole, and 4 received a combination therapy with levofloxacin and trimethoprim–sulfamethoxazole. For patients receiving trimethoprim–sulfamethoxazole, the median dosage of trimethoprim component was 6.2 mg/kg (IQR 4.49–8.94) regardless of renal function, and 6.6 mg/kg (IQR 5.3–12) for those with normal renal function (n = 35).
Table 2
Antibiotic therapy related to BSI due to S. maltophilia
Variable | Total (n = 214) |
No. of different antibiotic classes used empirically in each infected patient, median (IQR) | 2 (1–2) |
Empiric antibiotics used after BSI due to S. maltophilia, n (%) | |
Third-generation cephalosporin | 19 (8.9) |
Fourth-generation cephalosporin | 41 (19.2) |
Carbapenem | 109 (50.9) |
Aminoglycoside | 9 (4.2) |
Colistin | 29 (13.6) |
Daptomycin | 5 (2.3) |
Fluoroquinolone | 50 (23.4) |
Glycopeptide | 60 (28.0) |
Linezolid | 10 (4.8) |
β-lactam/β-lactamase inhibitor | 52 (24.3) |
Trimethoprim-sulfamethoxazole | 24 (11.2) |
Tigecycline | 16 (7.8) |
Colimycin | 29 (13.6) |
Fosfomycin | 3 (1.4) |
Appropriate antibiotic therapy after BSI, n (%) | 133 (62.1) |
Empiric antibiotic therapy | 48 (36.1) |
Definite antibiotic therapy | 85 (63.9) |
S. maltophilia, Stenotrophomonas maltophilia; BSI, blood stream infection; IQR, interquartile range |
Primary Outcome Analysis
A comparison of clinical characteristics and clinical outcomes among patients with- and without appropriate antibiotic therapy before and after propensity score matching is shown in Table 3. For original unmatched included patients (n = 214), there were no significant differences in age, sex, hospital stay before BSI, comorbidities, and source of BSI (all p > 0.05). With regard to disease severity, patients receiving an appropriate antimicrobial therapy had a significantly lower APACHE score than those who did not receive an appropriate antimicrobial therapy (22 vs. 25, p < 0.001). Comparison of clinical outcomes revealed that patients receiving an appropriate antibiotic therapy had a significantly lower mortality rate than those who did not receive an appropriate antibiotic therapy (10.5% vs. 46.9%, p < 0.001) at day 14. The median hospital stay after BSI showed no significant difference for the groups (45 days vs. 37 days, p = 0.382). After propensity score matching, we were able to match 61 couples to patients who received or did not receive an appropriate antibiotic therapy after BSI (Table 3). Overall, the groups were well balanced, with standardized differences of the variables involved in PS being < 0.1 after matching except for the variables, liver cirrhosis and acute kidney injury (standardized differences being 0.128 and 0.10, respectively, Table S1). Comparison of clinical outcomes between groups revealed consistent results as for the original unmatched patients. There were 7 (11.5%) and 24 (39.3%) deaths within the 14-day follow-up period in patients with and without appropriate antimicrobial therapy, respectively (p < 0.001). Hospital stay after BSI was not significantly different for the groups (37 days vs. 36 days, p = 0.430)
Table 3
Clinical characteristics for patients with- and without appropriate antibiotic therapy before and after propensity score matching analysis
Variable | Unmatched cohort | Propensity-matched cohort |
With appropriate antibiotic therapy (n = 133) | Without appropriate antibiotic therapy (n = 81) | p value | With appropriate antibiotic therapy (n = 61) | Without appropriate antibiotic therapy (n = 61) | p value |
Age, median (IQR) | 74 (58.5–83) | 74 (56–84) | 0.977 | 73 (58–83) | 74 (56–84) | 0.890 |
Sex, n (%) | 88 (66.2) | 52 (64.2) | 0.769 | 38 (62.3) | 38 (62.3) | 1.000 |
Length of hospital stay before BSI, median (IQR) | 23 (15–36) | 26 (15–46) | 0.608 | 24 (16–37) | 26 (16–45) | 0.535 |
Co-morbidity, n (%) | | | | | | |
Heart failure | 26 (19.5) | 18 (22.2) | 0.639 | 18 (29.5) | 15 (24.6) | 0.541 |
Old cerebrovascular accident | 23 (17.3) | 19 (23.5) | 0.271 | 9 (14.8) | 12 (19.7) | 0.472 |
Chronic obstructive pulmonary disease | 15 (11.3) | 17 (21.0) | 0.053 | 9 (14.8) | 10 (16.4) | 0.803 |
Diabetes mellitus | 49 (36.8) | 29 (35.8) | 0.878 | 24 (39.3) | 21 (34.4) | 0.573 |
Liver cirrhosis | 8 (6.0) | 9 (11.1) | 0.181 | 5 (8.2) | 3 (4.9) | 0.717 |
Chronic kidney disease | 24 (18.0) | 15 (18.5) | 0.931 | 12 (46.2) | 14 (53.8) | 0.658 |
Malignancies | 34 (25.6) | 23 (28.4) | 0.650 | 13 (21.3) | 14 (23.0) | 0.827 |
Charlson Comorbidity Index, median (IQR) | 3 (2–5) | 3 (2–6) | 0.279 | 3 (2–5) | 3 (2–5) | 0.729 |
Table 3
Clinical characteristics for patients with- and without appropriate antibiotic therapy before and after propensity score match (continued)
Variable | Unmatched cohort | Propensity-matched cohort |
With appropriate antibiotic therapy (n = 133) | Without appropriate antibiotic therapy (n = 81) | p value | With appropriate antibiotic therapy (n = 61) | Without appropriate antibiotic therapy (n = 61) | p valuea |
Source of BSI, n (%) | | | | | | |
Respiratory tract | 75 (56.4) | 44 (54.3) | 0.768 | 38 (62.3) | 37 (60.7) | 0.852 |
Urinary tract | 2 (1.5) | 1 (1.2) | 1.000 | 1 (1.6) | 1 (1.6) | 1.000 |
Catheter related | 6 (4.5) | 5 (6.2) | 0.751 | 2 (3.3) | 3 (4.9) | 1.000 |
Intra-abdominal | 2 (1.5) | 1 (1.2) | 1.000 | 0 (0) | 0 (0) | - |
Primary | 48 (36.1) | 29 (35.8) | 0.966 | 19 (31.1) | 20 (32.8) | 0.846 |
Skin and soft tissue | 0 (0) | 1 (1.2) | 0.379 | 0 (0) | 0 (0) | - |
Disease severity, n (%) | | | | | | |
APACHE score | 22 (16–26) | 25 (20–30) | < 0.001 | 23 (19–28) | 22 (19–28) | 0.760 |
Acute kidney injury | 28 (21.1) | 23 (28.4) | 0.221 | 17 (27.9) | 13 (21.3) | 0.400 |
Shock | 43 (32.3) | 31 (38.3) | 0.459 | 21 (34.4) | 20 (32.8) | 0.848 |
Mechanical ventilation | 84 (63.2) | 58 (71.6) | 0.205 | 44 (72.1) | 44 (72.1) | 1.000 |
Clinical outcome, n (%) | | | | | | |
14-day mortality | 14 (10.5) | 38 (46.9) | < 0.001 | 7 (11.5) | 24 (39.3) | < 0.001 |
Hospital stay after BSI, median (IQR) | 45 (25–67) | 37 (24.8–60) | 0.382 | 37 (22–67) | 36 (21–53) | 0.430 |
The assessment of hospital stays did not include patients who died during hospitalization. |
BSI, blood stream infection; IQR, interquartile range |
In Table 4, in univariate analysis of Cox proportional hazards model for original included patients, the variables significantly associated with 14-day mortality were APACHE score (HR 1.166, 95% CI 1.098–1.237, p < 0.001), acute kidney injury (HR 3.148, 95% CI 1.616– 6.133, p = 0.001), whereas appropriate antibiotic therapy was a protective factor (HR 0.501, 95% CI 0.253–0.991, p = 0.047) (Table 4). In multivariate analysis, the APACHE score (HR 1.151, 95% CI 1.090–1.216, p < 0.001), acute kidney injury (HR 2.392, 95% CI 1.223–4.675, p = 0.011), and appropriate antibiotic therapy (HR 0.482, 95% CI 0.243–0.957, p = 0.037) were still significantly associated with 14-day mortality (Table 4). Similarly, for the propensity-matched cohort, the APACHE score was significant associated with 14-day mortality in univariate and multivariate analysis (HR 1.114 and 1.118, 95% CI 1.057–1.174 and 1.056–1.184, both p < 0.001). Appropriate antibiotic therapy remained a significant protective factor in appropriate antibiotic therapy for 14-day mortality in univariate and multivariate analysis (HR 0.241 and 0.222, 95% CI 0.104–0.558 and 0.095–0.518, p = 0.001 and 0.001, respectively).
Table 4
Cox regression analysis of factors associated with 14-day mortality in patients with BSI due to S. maltophilia
Variable | Unmatched cohort | Propensity-matched cohort |
Unadjusted HR (95% CI) | p value | Adjusted HR (95% CI) | p value | Unadjusted HR (95%CI) | p value | Adjusted HR (95%CI) | p value |
Age | 0.992 (0.976–1.008) | 0.337 | | | 0.984 (0.966–1.001) | 0.068 | | |
Length of hospital stay before BSI | 1.007 (0.996–1.019) | 0.211 | | | 1.012 (0.999–1.024) | 0.066 | | |
Charlson Comorbidity Index | 1.050 (0.922–1.195) | 0.461 | | | 1.068 (0.931–1.225) | 0.345 | | |
Source of BSI | | | | | | | | |
Respiratory tract | 0.865 (0.446–1.678) | 0.667 | | | 0.737 (0.352–1.541) | 0.415 | | |
Catheter related | 1.693 (0.527–5.445) | 0.374 | | | 2.046 (0.432–9.700) | 0.364 | | |
Primary | 0.845 (0.415–1.724) | 0.642 | | | 0.949 (0.434–2.074) | 0.895 | | |
Disease severity at BSI | | | | | | | | |
APACHE score | 1.166 (1.098–1.237) | < 0.001 | 1.151 (1.090–1.216) | < 0.001 | 1.114 (1.057–1.174) | < 0.001 | 1.118 (1.056–1.184) | < 0.001 |
Acute kidney injury | 3.148 (1.616–6.133) | 0.001 | 2.392 (1.223–4.675) | 0.011 | 1.826 (0.836–3.989) | 0.130 | | |
Shock | 1.591 (0.812–3.116) | 0.174 | | | 1.365 (0.631–2.950) | 0.426 | | |
Mechanical ventilation | 2.112 (0.907–4.919) | 0.083 | | | 2.594 (0.881–7.637) | 0.083 | | |
Appropriate antibiotic therapy | 0.501 (0.253–0.991) | 0.047 | 0.482 (0.243–0.957) | 0.037 | 0.241 (0.104–0.558) | 0.001 | 0.222 (0.095–0.518) | 0.001 |
S. maltophilia, Stenotrophomonas maltophilia; HR, hazard ratio; CI, confidence interval; BSI, blood stream infection |
Secondary Outcome Analysis
Table 5 shows the comparison of 14-day mortality according to treatment regimens. In univariate analysis, factors associated with 14-day mortality were acute kidney injury (HR 2.999, 95% CI 1.040–8.648, p = 0.042) and APACHE score (HR 1.138, 95% CI 1.045–1.239 p = 0.003). Compared with the use of trimethoprim–sulfamethoxazole, the use of levofloxacin was associated with a lower mortality (HR 0.167, 95% CI 0.037–0.748, p = 0.019). In the multivariable analysis after adjustment, the use of levofloxacin was still associated with a lower mortality than achieved with the use of trimethoprim–sulfamethoxazole, but the difference was not significant (HR 0.233, 95% CI 0.050–1.084, p = 0.063).
Table 5
Clinical predictors for 14-day after BSI by S. maltophilia among patients with appropriate antibiotic therapy
Variable | Unadjusted HR (95% CI) | p value | Adjusted HR (95% CI) | p value |
Age | 0.994 (0.967–1.021) | 0.664 | | |
Length of hospital stay before BSI | 1.005 (0.991–1.019) | 0.512 | | |
Charlson Comorbidity Index | 1.017 (0.820–1.261) | 0.881 | | |
Source of blood stream infection | | | | |
Respiratory tract | 1.434 (0.480–4.278) | 0.518 | | |
Primary | 0.649 (0.204–2.069) | 0.465 | | |
Disease severity at BSI | | | | |
APACHE score | 1.138 (1.045–1.239) | 0.003 | 1.092 (0.996–1.197) | 0.061 |
Acute kidney injury | 2.999 (1.040–8.648) | 0.042 | 2.190 (0.733–6.542) | 0.160 |
Shock | 1.651 (0.573–4.758) | 0.353 | | |
Mechanical ventilation | 1.443 (0.453–4.601) | 0.535 | | |
Antibiotic use | | | | |
Trimethoprim-sulfamethoxazole containing regimen | Ref variable | Ref variable | Ref variable | Ref variable |
Levofloxacin containing regimen | 0.167 (0.037–0.748) | 0.019 | 0.233 (0.050–1.084) | 0.063 |
129 patients were analyzed, 4 patients were excluded because they received combination therapy for BSI |
HR, hazard ratio; CI, confidence interval; BSI, blood stream infection |