During the study period, 39 patients were diagnosed of definite E. faecalis IE. Subsequently, in the same period, 82 patients with E. faecalis bacteremia were identified, in whom the diagnosis of IE had been ruled out. Table 1 shows the characteristics and clinical evolution of both groups of patients.
Table 1
Clinical and epidemiological characteristics of patients with E. faecalis bacteremia
|
No IE
(n = 82)
|
IE
(n = 39)
|
p
|
Age, years (IQR)
|
75 (63–85)
|
78 (73–84)
|
0.094
|
Male sex, n (%)
|
60 (73)
|
29 (74)
|
0.999
|
Charlson index, (IQR)
|
2 (1–4)
|
2 (1–3)
|
0.419
|
Nosocomial-acquired, n (%)
|
40 (49)
|
12 (31)
|
0.078
|
Comorbidities, n (%)
Heart failure
Ischemic heart disease
Chronic liver diseases
Diabetes mellitus
Chronic renal failure
Tumour
|
15 (18)
16 (20)
5 (10)
22 (27)
13 (16)
26 (32)
|
15 (39)
12 (31)
4 (11)
12 (31)
1 (3)
9 (23)
|
0.024
0.176
0.999
0.670
0.035
0.394
|
Previous invasive procedures, n (%)
Urologic
Digestive
Others
|
17 (21)
2 (2)
7 (9)
|
2 (5)
1 (3)
7 (18)
|
0.032
0.999
0.141
|
Intracardiac devices, n (%)
Prosthetic valve
Pacemaker
|
14 (17)
13 (16)
2 (2)
|
30 (77)
24 (62)
8 (21)
|
< 0.001
< 0.001
0.002
|
DENOVA scale, (IQR)
|
1 (1–2)
|
3 (3–5)
|
< 0.001
|
Treatment duration, days (IQR)
|
13 (10–15)
|
46 (42–50)
|
< 0.001
|
TTP, hours (IQR)
TTP ≤12 hours, n (%)
TTP ≤ 8 hours, n (%)
|
11 (8.6–13.7)
52(63)
16 (20)
|
10 (8.0–13.0)
28 (72)
10 (26)
|
0.175
0.416
0.482
|
Outcome, n (%)
Clinical cure
Recurrence at 90 days
Mortality at 30 days
|
72 (88)
6 (7)
10 (12)
|
31 (80)
2 (5)
2 (5)
|
0.230
0.999
0.334
|
Note: IE, infective endocarditis; IQR, interquartile range; TTP, time to positive. |
The median age was slightly higher in the IE group (IE, 78 years vs. non-IE, 75 years, p = 0.094). Although there were no differences in the Charlson index score (IE 2 vs. non-IE 2, p = 0.419), there was a higher proportion of patients with heart failure in the IE group (IE, 39% vs. non- IE, 18%, p = 0.024) and a lower percentage of patients with chronic renal failure (IE, 3% vs. non-IE, 16%, p = 0.035). In both groups, a high number of nosocomial bacteremia was observed (IE, 31% vs. non-IE, 49%, p = 0.078). Previous urological procedures were more common in the group without IE (IE, 5% vs. non-IE, 21%, p = 0.032). On the other hand, the presence of cardiac devices (prosthetic valve or pacemaker) was more frequent in patients with IE (IE, 77% vs. non-IE, 17%, p < 0.001).
Transthoracic echocardiography (TTE) was performed in 100% of patients with IE and in 31% of patients without IE (p < 0.001). Transoesophagic echocardiography (TEE) was performed in 38% of patients.
The DENOVA score was higher in the group with IE (IE, 3 [3–5] vs. non-IE, 1 [1–2], p < 0.001) and the TTP was slightly lower in patients with IE (IE, 10 hours [8.0-13-0] vs. non-IE 11 hours [8.6–13.7], p = 0175). However, TTP ≤ 12 hours (IE, 72% vs. non-IE, 63%, p = 0.416) and TTP ≤ 8 hours (IE, 26% vs. non-IE, 20%, p = 0.482)
The clinical outcome was similar among the two groups, both in clinical cure (IE, 80% vs. non-IE 88%, p = 0.230), recurrence at 90 days (IE, 5% vs. non-IE, 7%, p = 0.999), as in 30-day mortality (IS, 5% vs. non-IS, 12%, p = 0.334).
E. faecalis community-acquired bacteremia
A sub-analysis of patients with E. faecalis community bacteraemia was performed. Again, the characteristics of both groups of patients were similar (Table 2). A greater number of urological procedures were performed in patients without IE (IE, 4% vs. non-IE, 26%, p = 0.021). A higher percentage of intracardiac devices (IE, 67% vs. no -IE, 14%, p < 0.001) and a higher score on the DENOVA scale (IE 3 vs. non-IE 1, p < 0.001) were observed in patients with IE.
Table 2
Clinical and epidemiological characteristics of patients with community-acquired E. faecalis bacteremia
|
No IE
(n = 42)
|
IE
(n = 27)
|
p
|
Age, years (IQR)
|
77 (65–86)
|
77 (71–80)
|
0.810
|
Male sex, n (%)
|
32 (76)
|
20 (74)
|
0.999
|
Charlson index, (IQR)
|
3 (1–6)
|
2 (1–3)
|
0.198
|
Comorbidities, n (%)
Heart failure
Ischemic heart disease
Chronic liver disease
Diabetes mellitus
Chronic renal failure
Tumour
|
5 (17)
9 (22)
3 (7)
14 (33)
9 (21)
16 (38)
|
9 (33)
6 (22)
4 (15)
9 (33)
0
7 (24)
|
0.146
0.999
0.420
0.999
0.010
0.433
|
Previous invasive procedures, n (%)
Urologic
Digestives
Others
|
11 (26)
0 (0)
2 (5)
|
1 (4)
1 (4)
3 (11)
|
0.021
0.391
0.373
|
Intracardiac devices, n (%)
Prosthetic valve
Pacemarker
|
6 (14)
5 12)
2 (2)
|
18 (67)
13 (52)
5 (19)
|
< 0.001
0.001
0.002
|
DENOVA scale, (IQR)
|
1 (1–2)
|
3 (3–5)
|
< 0.001
|
TTP, hours (IQR)
TTP ≤12 hours, n (%)
TTP ≤8 hours, n (%)
|
12 (10–14)
23 (55)
3 (7)
|
10 (8–13)
18 (67)
7 (26)
|
0.061
0.452
0.041
|
Outcome, n (%)
Clinical cure
Recurrence at 90 days
Mortality at 30 days
|
39 (93)
5 (12)
4 (10)
|
22 (82)
1 (4)
2 (7)
|
0.253
0.392
0.999
|
Note: IE, infective endocarditis; IQR, interquartile range; TTP, time to positive. |
In this sub-analysis, a lower TTP (IE, 10 hours [8–13] vs. non-IE, 12 hours [10–14], p = 0.061), and a TTP ≤ 8 hours (IE, 26% vs. non-IE 7%, p = 0.041) was found in IE patients. However, TTP ≤ 12 hours was similar among both groups (IE, 67% vs. non-IE, 55%, p = 0.452). In both groups the clinical course was similar.
DENOVA score ± TTP sensitivity analysis
Using ROC curves, the AUC of the DENOVA scale and the AUC of this scale combined with different TTP cut-off points (TTP ≤ 12 hours, TTP ≤ 10 hours and TTP ≤ 8 hours) were calculated. The best value of AUC was obtained by the DENOVA scale, with an AUC of 0.896 (Graph 1). The DENOVA scale, in our population, showed a sensitivity of 82%, a specificity of 85%, a PPV 73% and NPV 91%.
In the community bacteremia subgroup, it was observed that the addition of a TTP ≤ 8 hours to the DENOVA scale improved the accuracy in the prediction of IE, with an AUC of 0.914 (Graph 2). In this subgroup, the sensitivity of the DENOVA score was 85%, the specificity 81%, PPV 82% and NPV 90%. The DENOVA-8 score presented a sensitivity of 89%, a specificity of 88%, a PPV of 83% and a NPV 93%.
Factors associated with poor prognosis
No association between a lower TTP and a worse prognosis was found (Table 3). A Charlson index ≥ 3 was the only factor associated with poor clinical outcome (OR 3.4, 95% CI [1.21–9.62], p = 0.020).
Table 3
Factors related with poor clinical prognosis
|
Poor prognosis
(n = 20)
|
RR (95% CI)
|
p
|
OR (95% CI)
|
p
|
Age ≥ 80 years
Yes
No
|
6 (13)
14 (18)
|
0.7 (0.30–1.75)
|
0.614
|
|
|
IE
Yes
No
|
4 (10)
16 (20)
|
0.5 (0.19–1.47)
|
0.295
|
|
|
Charlson index ≥ 3
Yes
No
|
14 (26)
6 (9)
|
2.8 (1.15–7.85)
|
0.025
|
3.4 (1.21–9.62)
|
0.020
|
Dementia
yes
No
|
5 (36)
15 (14)
|
1.5 (1.09–5.92)
|
0.055
|
|
|
TTP ≤ 12 hours
Yes
No
|
13 (16)
7 (17)
|
1.0 (0.41–2.20)
|
0.999
|
|
|
TTP ≤ 10 hours
Yes
No
|
4 (9)
16 (21)
|
0.4 (0.15–1.18)
|
0.127
|
|
|
TDP ≤ 8 hours
Yes
No
|
2 (8)
18 (19)
|
0.4 (0.10–1.64)
|
0.238
|
|
|
Note: IE, infective endocarditis; OR, odds ratio; RR, relative risk; TTP, time to positive. |