Patients with atrial fibrillation (AF) for >48 hours who are a candidate for cardioversion should have transesophageal echocardiography (TEE) performed to exclude left atrial thrombus (LAT) that may cause systemic thromboembolism upon conversion to sinus rhythm. However, TEE facilities were limited mainly in developing countries and left atrial appendage may be a challenge to visualize by transthoracic echocardiography. In this article, we would like to discuss a patient that presented with AF >48 hours and is a candidate for cardioversion, however, there is no TEE at the facility. We performed a comprehensive search on whether d-dimer assay can be used to exclude LAT in this patient in order to perform safe cardioversion.

The clinical question is whether d-dimer assay can be used to exclude the diagnosis of LAT in patients presenting with AF >48 hours without the use of TEE or should patients be anticoagulated for weeks before attempting cardioversion. To answer this question, two independent authors (E.Y and V.C) performed a comprehensive and systematic search on studies on the use of D-Dimer assay to exclude the left atrial thrombus in patients with atrial fibrillation presenting >48 hours with keywords [“d-dimer”, “left atrial thrombus” and “atrial fibrillation”] and its synonym from inception up until June 2019 through PubMed, EBSCOhost, Cochrane CENTRAL, Proquest, EuropePMC, and hand-sampling from article references, discrepancies were resolved by discussion. Data extraction and quality assessment using appraisal tools from the Center of Evidence-Based Medicine, the University of Oxford was done by two independent authors (R.P and V.C). We included original articles and systematic reviews. We found a total of 182 unique results, 9 were relevant through screening of titles/abstract and the full-texts were assessed for eligibility [Figure 1] After excluding duplicate, applying inclusion criteria and performing critical appraisal using Oxford CEBM tools, we included seven studies [Table 1] and [Table 2] . All seven studies founded that D-dimer level significantly elevated in a group with LAT. Five were cross-sectional studies; one was a cohort and a systematic review. [Table 3]

Figure 1. Flow Chart of Search Strategy

---

---

Direct current cardioversion into sinus rhythm on patients with AF >48 hours may cause systemic thromboembolism due to the presence of LAT.[¹] Thrombus formation leads to increased d-dimer products which means that D-dimer assay can possibly be used to exclude LAT in AF patients. Five included studies showed that d-dimer assay was elevated in those with LAT.

For D-dimer to have clinical implication, it is important to have a cut-off point that has satisfying diagnostic performance. Habara et al. showed that d-dimer assay at 1150 ng/mL had a negative predictive value (NPV) of 97% for LAT.[²] Somloi et al. found 600 ng/mL has a sensitivity: 89% and specificity: 75%, positive predictive value (PPV): 33% and NPV: 98%.[³] Tarnowski et al. found that d-dimer were significantly elevated in patients with AF and detected LA thrombus (0.13 ± 0.02 vs.0.69 ± 0.21, p=0.015); however, it is not significant at 500 ng/mL, which is the conventional cut-off value.[⁴] In contrast, Bejinariu et al. revealed that d-dimer assay failed to predict the LAT.[⁵] Wan et al. showed that elevated D-dimer was associated with left atrial spontaneous echo contrast/ LAT with moderate sensitivity and specificity. In 4 out of 5 studies, AUC was > 0.70, sensitivity and specificity varied from 75.9% to 89% and 73.1% to 95% respectively.[⁶] However, there is no single cut-off point, due to the heterogeneity of cut-off points.

The findings are congruent to Yasaka et al. study who evaluate the use of d-dimer to detect LAT in patients with mitral stenosis, they found that D-dimer at 300ng/mL has sensitivity= 61%, specificity= 93%, PPV= 79%, and NPV= 86%. In their study, 44 out of 63 subjects have AF. Several patients were found to have elevated d-dimer despite the absence of intra-atrial coagulation on echocardiography which may be explained by mobile thrombi undetected on TEE. This is a possible additional benefit of D-dimer assay.[⁷]

Currently, there is no recommendation on D-dimer use to exclude LAT; however, the NPV of D-dimer assay is comparable to TEE in excluding left atrial thrombi. D-dimer’s effectiveness was shown to fall short to that of TEE (NPV= 98.7%; Hwang et al.) in studies performed by Habara et al. (NPV= 97%), Somloi et al. (NPV=98%), and Yasaka et al. (NPV= 86%). Thus D-dimer assay has the potential to be used to exclude LA thrombi[⁸]. Du et al. reported age-adjusted d-dimer assay has sensitivity= 100.0%, specificity= 86.7%, PPV= 51.7%, NPV= 100.0% for detection of LAT, the study included both valvular and non-valvular AF; an excellent diagnostic performance.[⁹] Du et al. study further combined cardiac computed tomography angiography to reach a sensitivity= 100.0%, specificity= 97.9%, PPV= 88.2%, and NPV= 100.0%. Hence, the combination may be valuable in the hospital without TEE but with cardiac computed tomography angiography.[⁹] The importance of d-dimer assay was further strengthened by a study by Milhem et al. whom suggested a d-dimer >270 ng/mL along with the history of stroke, heart failure, and hypertension formed a reliable atrial thrombus exclusion (ATE) score.[¹⁰] Milhem et al. demonstrated that an ATE score of 0 has a 0% atrial thrombus in 911 patients. While ATE score of 1 and 2 had 2.3% (out of 988 patients) and 2.9% (out of 481 patients) incidence of thrombus.[¹⁰] An ATE score of 3 and 4 had 9% (out of 111 patients) and 33.3% (out of 3 patients) incidence of atrial thrombus.[¹⁰] The ATE score is ideal in a rural hospital with limited facilities, needing only d-dimer assay and other variables that can be obtained by history taking and physical examination.

The patient in previous case illustration may benefit from cardioversion; however, those with AF >48 hours have a high risk for thromboembolism. The current recommendation is to rule out the LAT by the use of TEE. We conclude at that time (the year 2017) that there was not enough evidence to routinely recommend measurement of plasma D-dimer to rule-out LAT; hence, the safest option is to administer anticoagulation to patients for weeks before performing cardioversion in the absence of TEE facility nearby. However, with the advent of ATE score proposed by Milhem et al. in 2019; measurement of D-dimer assay in the evaluation of patient requiring exclusion of LAT is reasonable. This patient has an ATE score of 1 having a 2.3% risk of stroke, hence low but null risk of cardioversion. D-dimer assay (and ATE score) in addition to transthoracic echocardiography may be of use, especially in those with a risk score of 1 to 3, hence, we recommend further studies to evaluate this possibility.

Limitations in this study are selection bias in which only positive studies are published, hence, may lead to overestimation of diagnostic performance. These studies also have varying cut-off point, to have a clinical impact, there has to be a single cut-off point.

D-dimer assay is valuable in excluding LAT; especially combined other variables of ATE score. Previously, weeks of anticoagulation is more advisable before attempting cardioversion in the absence of nearby TEE facilities. With current evidence, a low D-dimer and ATE score of 0 is safe for cardioversion. Further investigations on diagnostic performance by the combined use of transthoracic echocardiography and ATE score further risk stratify those with a score of 1 to 3 may further delineate the risk of cardioversion in hospitals with limited facilities.