Safety of dobutamine or adenosine stress cardiac magnetic resonance imaging in patients with left ventricular thrombus

The clinical database was searched for patients undergoing stress CMR (adenosine, dobutamine, or regadenoson) with concomitant LV thrombus between February 2015 and September 2022.

If available, it was noted whether the transthoracic echocardiography reports at the time of CMR (or up to 30 days before or after) detected the LV thrombus. Cardiovascular risk factors such as arterial hypertension, current or former smoking habit, hypercholesterolemia, diabetes mellitus, obesity, and family history of cardiovascular disease were assessed using medical reports. Comorbidities including prior myocardial infarction, prior venous thrombosis, prior arterial embolism, active malignant disease, and known thrombophilia were also assessed. For arterial embolisms, the timing of occurrence relative to the CMR was recorded: > 12 months prior to CMR (no assumed correlation to LV thrombus) or < 12 months before CMR (assumed correlation to LV thrombus). As a control, a sex-, age- and LV ejection fraction-matched control group of patients undergoing a stress-free CMR with concomitant LV thrombus was retrospectively identified in our database. Likewise, prior arterial embolisms < 12 months before CMR (assumed correlation to LV thrombus) were assessed for this group.

CMR imaging was performed at a 1.5 Tesla or a 3 Tesla MRI scanner (Ingenia and Ingenia CX, Philips Healthcare, Best, The Netherlands). Cine images were obtained with a steady-state free precession sequence using retrospective electrocardiographic gating in at least three long-axis planes (two, three, and four-chamber views), as well as a short-axis stack of the whole left ventricle (slice thickness 8 mm) during breath-hold with 35 phases per cardiac cycle. For dobutamine stress, CMR, three long axis and three short-axis planes (basal, midventricular, apical) were acquired before and during dobutamine infusion starting at 10 µg/kg/min and increased by 10 µg/kg/min every step to a maximum of 40 µg/kg/min (20 µg/kg for low-dose vitality testing). In case of an insufficient heart rate response, additional medication with up to 2 mg atropine intravenously was used in the absence of contraindications. For vasodilator stress CMR, an intravenous infusion of 140 or 210 µg/kg/min adenosine for three minutes or a 0,4 mg single injection of regadenoson was followed by a gadolinium-based contrast agent bolus and a three-slice turbo field gradient echo-echo-planar imaging sequence. Patients were able to communicate with the technician or doctor during the exam via intercom and were regularly asked about the occurrence of symptoms like shortness of breath or angina. Late gadolinium enhancement images were acquired 10 min after the contrast agent injection. We used the contrast agent Gadovist^™ (Bayer AG, Leverkusen, Germany) at a dose of 0.1–0.2 mmol/kg. Image analysis was performed using the IntelliSpace Portal (Philips Healthcare) and cvi42 (Circle Cardiovascular Imaging, Calgary, Canada). Images were verified for the presence of an LV thrombus and analyzed by two independent readers with more than 3 years of clinical CMR experience each. LV wall motion including the presence of a local akinesia or aneurysm (defined qualitatively as an abnormality in the diastolic contour with systolic dyskinesis) was assessed [41]. Each thrombus was classified as either mural (flat and parallel to the endocardial surface) or protruding (projecting into the left ventricular cavity) and mobile (motion of the thrombus independently of surrounding myocardial wall) or non-mobile as described earlier [41]. Global and regional (at the site of thrombus) longitudinal strain was calculated using the feature tracking module of cvi42.

This retrospective study was approved by the ethical commission of our institution (S-151/2019) and followed the declaration of Helsinki.

During the CMR scans, the occurrence of moderate or severe symptoms as well as the occurrence of minor or major adverse events was noted. A minor adverse event was defined as persisting severe symptoms or arrhythmias needing therapy. Major adverse cardiovascular events (MACE) were defined as cardiovascular death or non-fatal thromboembolic event such as stroke, transient ischemic attack, myocardial infarction, or other arterial embolism.

Patients’ follow-up was conducted by telephone interview, hospital follow-up or at their outpatient cardiologist for the occurrence of death from any cause and, limited to 12 months, for the occurrence of MACE. Controls were followed up for 12 months for the occurrence of MACE. Medication with oral anticoagulant before and after CMR was registered for both patients and controls.

Analyses were carried out using the R language and environment for statistical computing (version 4.2.1) with the user interface R Studio (version 2023.06.0/421) [42].

Normal distribution was assessed by using the Shapiro–Wilk test. Parametric variables are given as mean ± standard deviation and non-parametric variables as median with interquartile range.

19,312 patients in our databank were screened and 95 (78 male/17 female, 65 ± 10.7 years) fulfilled the inclusion criteria. Cardiovascular risk factors were frequent, 40 patients (42.1%) exhibited ≥ 3 risk factors. 93 patients (97.8%) had a history of a prior myocardial infarction, which occurred in the last 6 months before the CMR in 35 (36.8%) of those. In 68 of those patients (73.1%), initial reperfusion therapy had been successful (TIMI 2 in 7 patients, TIMI 3 in 61 patients). In another 15 (16.1%), reperfusion had been attempted but ultimately failed (TIMI 0 in 14 patients, TIMI 1 in 1 patient). After the myocardial infarction diagnosis, of those 83 patients 61 were treated in the form of dual therapy and 22 in the form of triple therapy. Ultimately, in 11 patients (11.8%) a myocardial infarction had not been noticed or diagnosed before, but invasive coronary angiography showed a total (9 patients, TIMI 0) or near total (2 patients, TIMI 1) occlusion of a coronary artery and CMR showed late gadolinium enhancement suggestive of past myocardial infarction. Since troponin values were inconspicuous, no reperfusion therapy was attempted, and no new medication was started in those 11 patients. A relevant portion of patients had an active malignant disease (10.5%), a known thrombophilia (3.2%), or a history of previous arterial (24.2%) or venous (12.6%) thrombosis. For all patients, a complete medical history with all events prior to and up to 48 h after the CMR exam was available. For MACE, a 30 day and 12 month follow-up was available for 94 (98.9%) and 89 (93.7%) patients, respectively. All patient characteristics are given in Table 1.

For the matched control group, 89 patients were identified to match the 89 patients of the study group for whom a 12 month follow-up was available. A 12-month follow-up after CMR and details about embolic events prior to the CMR were available for all controls. As per matching, sex [72 male, 17 female], age (62 ± 13.3 years) and LV ejection fraction (36.7 ± 13.4%) did not differ between the study group and controls. Neither did the rate of anticoagulation prior to (p = 0.84) or after the CMR exam (p = 0.77). The underlying cardiac pathology was an ischemic cardiomyopathy in 61 (68.5%), a dilated cardiomyopathy in 14 (15.7%) and another non-ischemic cardiomyopathy in 14 (15.7%). Cardiovascular risk factors were equally common with hypertension in 51 (57.3%), a smoking habit in 50 (56.2%), hypercholesterolaemia in 35 (39.3%), diabetes in 15 (16.9%), family history in 15 (16.9%) and obesity in 18 patients (20.2%), respectively.

CMR showed a reduced ejection fraction in most patients (89.5%), with 46.3% of patients exhibiting an ejection fraction of < 40%, a known precipitator of thrombus development [43]. Global longitudinal strain was − 9.3 ± 2.2%, whereas regional longitudinal strain at the thrombus location was significantly lower (− 2.8 ± 2.6%, p < 0.001). In 87.4% of patients, the thrombus was located in the apex and local akinesia was present at the site of the thrombus in nearly all patients (98.9%). 83.2% had an aneurysm at the site of the thrombus. Multiple thrombi were present in 12.6% and protruding thrombi in 51.6%. The mean thrombus size (2D) was 1.6 ± 1.6 cm². 43 patients underwent dobutamine stress (36 high-dose, 7 low-dose) whereas 52 underwent vasodilator stress CMR. Heart rate increased in all groups, most notably during high-dose dobutamine stress (Δ 68/min). The dobutamine stress CMR was positive in 3 cases and vasodilator stress CMR in 9 cases. The positive stress CMR lead to a successful revascularization of the affected coronary via PCI in 7 patients and via Bypass in 2 patients. In 1 patient, a coronary artery bypass surgery was initiated but ultimately failed. In 4 patients (2 of those with negative stress CMR but persisting symptoms) a revascularization via PCI was planned but estimated as technically infeasible during invasive diagnostic angiography. CMR data is given in Tables 2, 3 and exemplary cases are shown in Fig. 1.

Only three patients (3.1%) experienced moderate or severe symptoms during high-dose dobutamine stress, which completely resolved without medication. There were no minor or major complications during and up to 48 h after the CMR stress exam, and no examination had to be aborted. No patient in the control group experienced symptoms or complications during CMR and up to 48 h after the exam. Details on periprocedural safety data are given in Table 4.

In 61 patients, multimodality imaging with an additional transthoracic echocardiography within 30 days was available. The LV thrombus was only found in 17 cases (27.9%) by echocardiography. As a results of the newly diagnosed LV thrombus, anticoagulation was initiated in 65 patients. Of those, only 11 LV thrombi would have been diagnosed with echocardiography alone and the diagnosis and treatment decision were solely dependent on CMR imaging in the other 54.

Follow-up imaging (echocardiography or CMR) was conducted for 74 patients at least 3 months after the initial CMR exam. A thrombus resolution was observed in 55 cases (74.3%). In the subgroup of patients undergoing CMR as follow-up, thrombus resolution was observed in 27 of 42 patients (64.3%). There was no significant correlation between the type of anticoagulant and thrombus resolution (p = 0.299). Data is displayed in Fig. 2.

The thrombus characteristics size, mobility and protrusion did not correlate with the occurrence of embolic events in the study group (p = 0.625, p = 0.988, p = 0.848) or death (p = 0.329, p = 1, p = 0.236). Likewise, embolic events were independent of the chosen anticoagulant (p = 0.883 before CMR, p = 0.146 after CMR) and stress agent (p = 0.437) for the study group. During a median follow-up of 33.7 months, 6 deaths (6.3%) occurred in the study group with no correlation observed to the choice of stress agent or thrombus characteristics. The outcomes of the study group and controls are displayed in Fig. 3.