Background
The aim of this study was to investigate the ability of transthoracic echocardiography (TTE) to detect vegetations and the relationship between blood cultures and transesophageal echocardiography (TEE).
Methods
Five hundred eleven TTE and TEE pairs performed to evaluate endocarditis were retrospectively analyzed. Vegetation on TTE, prosthetic valve, change in regurgitation, and blood cultures were correlated with vegetation on TEE.
Results
TTE detected 45% of vegetations seen on TEE. There was no difference for prosthetic valves. Prosthetic valves (odds ratio, 1.7; P = .03) and increased regurgitation (odds ratio, 1.7; P = .01) were associated with vegetations on TEE; staphylococcal bacteremia and fungemia were not. Negative blood cultures were associated with negative results on TEE ( P < .0001), but 27% of patients with prosthetic valves had culture-negative endocarditis or nonbacterial thrombotic endocarditis, and 6% had abscesses missed by TTE.
Conclusion
This study demonstrates a limited capacity of TTE to detect vegetations. TEE may be an appropriate initial study to evaluate prosthetic valves. TEE for culture-negative endocarditis deserves further study.
The incidence of infective endocarditis (IE) is estimated to be between 2 to 6 cases per 100,000 patient-years. Even with advances in imaging technology, diagnostic protocols, and surgery, annual mortality approaches 40%. The prompt recognition and diagnosis of IE are essential to improving mortality rates and decreasing complications such as heart failure, periannular extension of infection, and embolization. Echocardiography and blood culture data play crucial roles in the diagnosis of IE.
The Duke criteria, currently the standard by which the diagnosis of IE is made, involve two major findings: (1) the presence of positive blood cultures and (2) evidence of endocardial involvement, primarily diagnosed using echocardiography. Transthoracic echocardiography (TTE) is often performed first because it is readily available and noninvasive. Multiple studies, most of which were performed before harmonic imaging and other modern echocardiographic advances, suggested that TTE has low sensitivity for the diagnosis of IE. The studies that have been published subsequently have been relatively small (≤50 patients), have excluded patients with prosthetic valves, and have reported conflicting results. Transesophageal echocardiography (TEE) is a more sensitive imaging modality for the detection of vegetations but is a semi-invasive test with a low but definite potential for complications such as aspiration and esophageal injury. In 2007, the echocardiographic appropriateness criteria deemed TEE an appropriate initial imaging test (instead of TTE) to diagnose and manage endocarditis with moderate or high pretest probability (eg, in the presence of staphylococcal bacteremia or fungemia).
In this study, we investigated the use of TTE and TEE in the diagnosis of IE by addressing 2 specific aims. We sought to compare the ability of TTE to detect vegetations visualized on TEE in the era of modern imaging techniques in a large sample of paired transthoracic and transesophageal studies. The sample included a large number of studies imaging prosthetic valves. We also investigated the association between blood culture data and vegetations found on TEE.
Methods
We retrospectively analyzed 511 consecutive TTE and TEE pairs performed for the indication of endocarditis (as defined by the ordering physician) at the Hospital of the University of Pennsylvania between 2003 and 2008. In this same time period, nearly 2500 transthoracic echocardiographic studies were performed for the indication of endocarditis, without subsequent TEE; these were excluded from the study. We restricted the analysis to only TTE and TEE pairs performed within 1 week. Fifteen pairs with TTE performed >1 week prior to TEE were excluded from the analysis comparing TTE and TEE. We then determined the sensitivity and positive predictive value (PPV) of findings consistent with vegetation on TTE compared with TEE.
Vegetations were defined as lesions that had a typical appearance and that did not have the characteristics of valvular degeneration, Lambl’s excrescence, or fibroelastoma. “Possible vegetations” were echodensities that could not be adequately defined and for which vegetations could not be excluded. Regurgitant jets were characterized on a semiquantitative scale (0 = none, 1 = mild, 1.5 = mild-moderate, 2 = moderate, 2.5 = moderate-severe, and 3 = severe) as assessed by a level 3 certified echocardiographer. Change in regurgitant jet was defined as an increase of ≥1 full grade (eg, from 1 to 2 but not from 1.5 to 2) from previous TTE (if available). Other valve abnormalities were noted, including sclerosis, calcification, thickening, flattening, prolapse, or flail leaflets.
We reviewed blood culture data drawn within 5 days of TEE. For those patients who were found to have vegetations on TEE and negative blood cultures, charts were reviewed to determine whether antibiotics were administered prior to the cultures’ being drawn. Ten transesophageal echocardiographic studies had no associated blood cultures drawn and were excluded from the analysis.
The χ 2 test was used to compare valve characteristics seen on TTE and clinical variables with the presence of vegetation on TEE, except as specifically noted. A 2-tailed P value < .05 was used to define statistical significance. This study was approved by the local institutional review board.
Results
Vegetations and Findings on TTE
Of 486 TTE and TEE pairs, there were 130 in which TTE demonstrated vegetations. There were 14 transthoracic studies with “possible vegetations.” Of the 144 transthoracic studies with vegetations or possible vegetations, 81 (56%) had vegetations seen on TEE. Of 342 pairs in which TTE had no vegetation, 98 (28%) had vegetations seen on TEE ( Figure 1 ). TTE was performed before TEE in all cases. The mean time interval between TTE and TEE was 2.4 days (range, 0-7 days). TTE detected 45% of the vegetations seen on TEE (including “possible” vegetations on TTE). The PPV was 56%. TTE performed ≤2 days before TEE detected 45% of the vegetations found on TEE. TTE performed >2 days before TEE detected 48% of vegetations found on TEE.
There were 258 transthoracic studies in which prior TTE was available for comparison. The mean time interval between index TTE and prior TTE was 313 days (range, 8-2160 days). There were increases in regurgitant jets by ≥1 grade in 138, which was associated with vegetations on TEE (odds ratio [OR], 1.7; 95% confidence interval [CI], 1.1-2.5; P = .014). The presence of any other valve abnormality on TTE (sclerosis, calcification, thickening, etc; OR, 1.5; 95% CI, 0.8-2.3; P = .26) was not significantly associated with vegetation on TEE.
Vegetations, Prosthetic Valves, and Findings on TTE
There were 114 patients with prosthetic valves. Vegetations were found in 51 (45%) on TEE. There were 8 abscesses (6%). The presence of a prosthetic valve was associated with vegetation on TEE (OR, 1.7; 95% CI, 1.1-2.6; P = .031). TTE identified 45% of the vegetations seen on TEE. Combining the findings of vegetation on TTE, prosthetic valve, and increased regurgitation as an indication for vegetation on TEE, TTE provided sensitivity of 64% and PPV of 38%.
Vegetations and Blood Cultures
Of the 501 pairs with blood cultures drawn within 5 days, 186 (37%) had vegetations on TEE. Of the transesophageal studies with vegetations, 100 of 186 (54%) had positive blood cultures, of which 53 grew staphylococcal species (48 Staphylococcus aureus and 5 S epidermidis ), 27 grew streptococcal species, 2 grew candidal species, and 18 grew other species, including gram-negative and anaerobic organisms. Of the transesophageal studies without vegetations, 172 of 315 (55%) had positive blood cultures, of which 102 (59%) grew staphylococcal species (83 S aureus and 19 S epidermidis ), 31 (18%) grew streptococcal species, 15 (9%) grew candidal species, and 24 (14%) grew other species ( Figure 2 ). For 13 of the transesophageal studies with vegetations but with negative blood cultures, it was unclear from the charts whether antibiotics were given prior to cultures’ being drawn. Of the remaining 73 transesophageal studies with vegetations but negative cultures, 47 (64%) had received antibiotics prior to cultures’ being drawn, and 26 had not. Therefore, 26 of 186 (14%) vegetations appeared to be culture-negative endocarditis (CNE) or nonbacterial thrombotic endocarditis (NBTE).
Excluding those patients who received antibiotics prior to cultures’ being drawn and those in whom it was unclear from the charts whether antibiotics had been given, the PPV of positive blood cultures for vegetation on TEE in this sample was 58%. None of the specific species of bacteria was associated with vegetation on TEE ( S aureus , P = .92; S epidermidis , P = .35; Streptococcus spp, P = .11). Candida fungemia was associated with negative results on TEE (OR, 0.2; 95% CI, 0.05-0.95; Fisher’s exact P = .03), although the number of patients with fungemia was small. Overall, positive blood cultures were not associated with evidence of IE on TEE (OR, 0.97; 95% CI, 0.67-1.4; P = .92). However, excluding those patients who received antibiotics prior to cultures’ being drawn and those for whom it was unclear from the charts whether antibiotics had been given, negative blood cultures were strongly associated with negative results on TEE (OR, 3.2; 95%, CI 2.0-5.2; P < .0001).
Prosthetic Valves and Blood Cultures
Of the patients with prosthetic valves and vegetations on TEE, 25 of 49 (51%) had positive blood cultures, while 24 of 49 (49%) had negative blood cultures. Ten of the patients with negative blood cultures had received antibiotics prior to cultures’ being drawn, and for 1 patient, it was unclear from the chart whether the patient had received antibiotics prior to culturing. Therefore, 13 of 49 patients (27%) with prosthetic valves had CNE or NBTE.
Overall, in the subset of patients with prosthetic valves, positive blood cultures were not associated with vegetations on TEE (OR, 1.6; 95% CI, 0.8-3.4; P = .30). The PPV of a blood culture for vegetation on TEE in this sample was 51%. However, excluding those patients who did receive or may have received antibiotics prior to cultures’ being drawn, negative blood cultures were associated with negative results on TEE (OR, 3.0; 95% CI, 1.3-6.9; P = .019). The PPV of a blood culture for TEE with vegetation in this sample rose to 62%.
Vegetations and Blood Cultures With Inconclusive Results on TTE
Restricting the analysis to the 371 transesophageal studies performed for the indication of endocarditis after inconclusive results on TTE (negative for vegetation or demonstrating a “possible vegetation”), 103 (28%) had vegetations identified. Of the transesophageal studies with vegetations, 33 of 103 (32%) had negative blood cultures, and 70 of 103 (68%) had positive cultures. Positive blood cultures were not statistically associated with the presence of vegetation on TEE (OR, 1.5; 95% CI, 0.9-2.4; P = .14). For 3 of the transesophageal studies associated with negative blood cultures, it was unclear from the charts whether the patients had received antibiotics prior to cultures’ being drawn. Of the remaining 30 transesophageal studies with vegetations but negative blood cultures, 19 patients (63%) had received antibiotics prior to cultures’ being drawn, and 11 (37%) had not ( Figure 3 ). Therefore, 11% of transesophageal studies (11 of 103) with vegetations after negative or inconclusive findings on TTE demonstrated CNE or NBTE. Of these, 2 of 11 patients (18%) died during their hospital stays of presumed sepsis, 1 (9%) had a history of rheumatologic disease, 1 (9%) was diagnosed with lung cancer during hospitalization, and 7 (64%) were treated and discharged with no other diagnoses made to account for the vegetations.
Of the 268 transesophageal studies without vegetations, 110 (41%) had negative blood cultures and 158 (59%) had positive blood cultures. In this sample of TTE and TEE pairs, the PPV of a blood culture for vegetation in this sample was 34%. Excluding those patients who had received antibiotics prior to cultures’ being drawn and those for whom it was unclear from the charts whether antibiotics had been given, the PPV rose to 57%. A negative blood culture was strongly associated with negative result on TEE (OR, 4.4; 95% CI, 2.2-8.8; P < .0001).
Results
Vegetations and Findings on TTE
Of 486 TTE and TEE pairs, there were 130 in which TTE demonstrated vegetations. There were 14 transthoracic studies with “possible vegetations.” Of the 144 transthoracic studies with vegetations or possible vegetations, 81 (56%) had vegetations seen on TEE. Of 342 pairs in which TTE had no vegetation, 98 (28%) had vegetations seen on TEE ( Figure 1 ). TTE was performed before TEE in all cases. The mean time interval between TTE and TEE was 2.4 days (range, 0-7 days). TTE detected 45% of the vegetations seen on TEE (including “possible” vegetations on TTE). The PPV was 56%. TTE performed ≤2 days before TEE detected 45% of the vegetations found on TEE. TTE performed >2 days before TEE detected 48% of vegetations found on TEE.
There were 258 transthoracic studies in which prior TTE was available for comparison. The mean time interval between index TTE and prior TTE was 313 days (range, 8-2160 days). There were increases in regurgitant jets by ≥1 grade in 138, which was associated with vegetations on TEE (odds ratio [OR], 1.7; 95% confidence interval [CI], 1.1-2.5; P = .014). The presence of any other valve abnormality on TTE (sclerosis, calcification, thickening, etc; OR, 1.5; 95% CI, 0.8-2.3; P = .26) was not significantly associated with vegetation on TEE.
Vegetations, Prosthetic Valves, and Findings on TTE
There were 114 patients with prosthetic valves. Vegetations were found in 51 (45%) on TEE. There were 8 abscesses (6%). The presence of a prosthetic valve was associated with vegetation on TEE (OR, 1.7; 95% CI, 1.1-2.6; P = .031). TTE identified 45% of the vegetations seen on TEE. Combining the findings of vegetation on TTE, prosthetic valve, and increased regurgitation as an indication for vegetation on TEE, TTE provided sensitivity of 64% and PPV of 38%.
Vegetations and Blood Cultures
Of the 501 pairs with blood cultures drawn within 5 days, 186 (37%) had vegetations on TEE. Of the transesophageal studies with vegetations, 100 of 186 (54%) had positive blood cultures, of which 53 grew staphylococcal species (48 Staphylococcus aureus and 5 S epidermidis ), 27 grew streptococcal species, 2 grew candidal species, and 18 grew other species, including gram-negative and anaerobic organisms. Of the transesophageal studies without vegetations, 172 of 315 (55%) had positive blood cultures, of which 102 (59%) grew staphylococcal species (83 S aureus and 19 S epidermidis ), 31 (18%) grew streptococcal species, 15 (9%) grew candidal species, and 24 (14%) grew other species ( Figure 2 ). For 13 of the transesophageal studies with vegetations but with negative blood cultures, it was unclear from the charts whether antibiotics were given prior to cultures’ being drawn. Of the remaining 73 transesophageal studies with vegetations but negative cultures, 47 (64%) had received antibiotics prior to cultures’ being drawn, and 26 had not. Therefore, 26 of 186 (14%) vegetations appeared to be culture-negative endocarditis (CNE) or nonbacterial thrombotic endocarditis (NBTE).
