Numerous studies have examined whether plasma D-dimer (DD) can be used to identify patients with acute aortic dissection (AAD). These studies have been inconclusive because of their limited sample sizes and the different cut-off values employed. We aimed to conduct a systematic review and meta-analysis to examine the utility of plasma DD as a screening tool for AAD. We systematically searched EMBASE and MEDLINE and hand-searched relevant articles to identify studies investigating plasma DD as a screening tool for AAD. A value of 500 ng/ml was defined as the threshold for a positive plasma DD finding because it is widely used for ruling out pulmonary emboli. Using DerSimonian–Laird random-effects models we pooled data across studies to estimate sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios (LRs). We identified 7 studies involving 298 subjects with AAD and 436 without. When data were pooled across studies, sensitivity (0.97, 95% confidence interval [CI] 0.94 to 0.99) and negative predictive value (0.96, 95% CI 0.93 to 0.98) were high. Specificity (0.56, 95% CI 0.51 to 0.60) and positive predictive value (0.60, 95% CI 0.55 to 0.66) were low. Negative LR showed an excellent discriminative ability (0.06, 95% CI 0.03 to 0.12), whereas positive LR did not (2.43, 95% CI 1.89 to 3.12). In conclusion, our meta-analysis suggests that plasma DD <500 ng/ml is a useful screening tool to identify patients who do not have AAD. Plasma DD may thus be used to identify subjects who are unlikely to benefit from further aortic imaging.
Recent studies have suggested that plasma D-dimer (DD) may be a useful screening tool to rule out acute aortic dissection (AAD) and thus avoid further imaging. However, these studies were inconclusive because of limited sample sizes and different DD cut-off values employed (range 100 to 900 ng/ml). Some studies have also lacked a control group, preventing assessment of important diagnostic test measurements such as specificity, positive predictive value (PPV) and negative predictive value (NPV), and positive and negative likelihood ratios (LRs). Therefore, we conducted a meta-analysis of previous studies to examine the utility of plasma DD as a screening test for AAD.
Methods
We systematically searched EMBASE (1980 through 2009) and MEDLINE (1964 through 2009) to identify all clinical studies investigating plasma DD as a screening tool for AAD. The Medical Subject Heading search string was (“fibrin fragment D”[Substance Name] or “fibrin fragment D”[All Fields] or “d dimer”[All Fields]) and (“aorta”[Medical Subject Heading Terms] or “aorta”[All Fields] or “aortic”[All Fields]) and (“dissection”[Medical Subject Heading Terms] or “dissection”[All Fields]). Our literature search was limited to studies conducted in humans and published in peer-review journals in English. Editorial comments, reviews, and reference lists of retrieved articles were hand-searched for additional data. A threshold of 500 ng/ml was used in our meta-analysis to define a positive plasma DD finding because this value is widely used for ruling out pulmonary emboli. When original studies reported data using different plasma DD cutoffs, we used supplemental data reported in previous reviews. Retrieved studies were examined to eliminate potential duplicates or overlapping data.
We included a study in our meta-analysis if (1) it reported AAD as occurring within 2 weeks of onset of symptoms and was confirmed by standard imaging techniques including computed tomography, transesophageal echocardiography, magnetic resonance imaging, aortography, autopsy examination, or pathologic examination; (2) plasma DD was measured by standardized assays; (3) it included a control group (healthy subjects or patients with other confirmed diagnoses) not having an AAD; and (4) absolute numbers of true positive, false positive, true negative, and false negative were reported or could be derived. Studies without a control group including case reports and case series were excluded, as were studies that involved nonhuman subjects. In addition, conference abstracts were excluded because these typically do not undergo rigorous peer review and their results may not be final.
Two reviewers (A.S. and T.D.) independently extracted data from each trial. Results were compared and disagreements were resolved by consensus. Data extracted included first author, year of publication, study period, study location, study design, AAD diagnostic techniques, and plasma DD assay employed. We extracted AAD population characteristics including gender and mean ± SD for age or median with interquintile range when available, symptoms at onset, time of enrollment, mortality, AAD classification according to Stanford or de Bakey classification systems, and mean/median level of plasma DD. In addition, we extracted diagnoses of control subjects and their mean plasma DD level. We extracted absolute numbers of true positive, false positive, true negative, and false negative to generate pooled estimates of diagnostic test characteristics.
This meta-analysis was performed according to the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) checklist for reporting meta-analyses of observational studies. Grading of quality was performed according to the Quality Assessment of Diagnostic-Accuracy Studies instrument, which is an established, evidence-based tool for systematic reviews of diagnostic studies. This tool uses a list of 14 questions, which are answered as “yes,” “no,” or “unclear,” to examine the potential for bias in a study.
To assess the ability of plasma DD to discriminate between patients with and without AAD we examined the following test characteristics: sensitivity, specificity, PPV and NPV, and positive and negative LRs. We examined these diagnostic test characteristics for each study and then pooled data across all studies using the random-effects models of DerSimonian and Laird. Zero cells were handled using a 0.5 continuity correction. Statistical heterogeneity was assessed using the Q statistic (p <0.1 considered statistically significant). We also calculated I 2 statistics to estimate the proportion of variation attributable to between-study heterogeneity. We used Excel (Office 2007, Microsoft, Redmond, Washington), Meta-DiSc 1.4 (Unidad de Biostatistica Clinica, Hospital Ramon y Cajal, Madrid, Spain), and Meta-Analyst (Tufts Medical Center, Boston, Massachusetts) for data handling and statistical analyses.
Results
In total 77 studies (excluding duplicates) were identified using the search strategy outlined earlier ( Figure 1 ). After the exclusion of nonrelevant studies, case reports, and reviews by title and abstracts, 12 studies were retrieved for full text evaluation. Five further studies were excluded at this stage because of lack of a control group, leaving 7 that met the inclusion criteria. These studies included 298 subjects with AAD and 436 without. Two of the 7 studies reported data using different plasma DD cutoffs. Consequently, we used previously reported supplemental data, which reported data using a cutoff of 500 ng/ml.
Tables 1 through 6 present summaries of the quality and characteristics of the included studies and patients. The number of subjects with AAD in each study was 16 to 94, and their mean/median age was 53 to 70 years. Patients with AAD were predominantly men in all studies apart from 1 12 (range 46% to 94%). Control patients varied and included healthy subjects and patients with other confirmed diagnoses such as atrial fibrillation, gastroesophageal disorders, heart failure, myocardial infarction, pulmonary emboli, unstable angina, and nonspecific causes of chest pain. Five studies reported in-hospital mortality of patients with AAD ( Table 3 ). Mortality rate in these studies was 13% to 50%. Four studies evaluated the prognostic value of plasma DD with respect to mortality. However, only 1 study found a significant correlation between plasma DD level and in-hospital mortality in patients with AAD (area under the curve 0.65 ± 0.07, p = 0.04), whereas the other studies did not.
| QUADAS Tool Items | Suzuki et al (2009) | Sbarouni et al (2007) | Ohlmann et al (2006) | Hazui et al (2005) | Akutsu et al (2005) | Eggebrecht et al (2004) | Weber et al (2003) |
|---|---|---|---|---|---|---|---|
| 1. Was the spectrum of patients representative of patients who will receive the test in practice? | + | + | + | + | + | + | + |
| 2. Were selection criteria clearly described? | + | + | + | + | + | + | + |
| 3. Is the reference standard likely to correctly classify the target condition? | + | + | + | + | + | + | + |
| 4. Is the period between the reference standard and index test short enough to be reasonably sure that the target condition did not change between the 2 tests? | + | + | + | + | + | + | + |
| 5. Did the entire sample or a random selection of the sample receive verification using a reference standard of diagnosis? | + | + | + | + | + | + | + |
| 6. Did patients receive the same reference standard regardless of index test result? | unclear | 0 | 0 | + | + | 0 | 0 |
| 7. Was the reference standard independent of the index test (i.e., index test did not form part of the reference standard)? | + | + | + | + | + | + | + |
| 8. Was execution of the index test described in sufficient detail to permit replication of the test? | 0 | 0 | + | unclear | + | + | + |
| 9. Was execution of the reference standard described in sufficient detail to permit its replication? | 0 | + | + | + | + | + | + |
| 10. Were index test results interpreted without knowledge of results of the reference standard? | unclear | unclear | unclear | unclear | unclear | unclear | unclear |
| 11. Were reference standard results interpreted without knowledge of results of the index test? | unclear | unclear | unclear | unclear | unclear | unclear | unclear |
| 12. Were the same clinical data available when test results were interpreted as would be available when the test is used in practice? | + | + | + | + | + | + | + |
| 13. Were uninterruptable/intermediate test results reported? | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 14. Were withdrawals from the study explained? | 0 | 0 | + | 0 | 0 | 0 | 0 |
| Study | Publication Year | Study Period | Study Location | Study Design | AAD Diagnostic Technique | Plasma DD Assay |
|---|---|---|---|---|---|---|
| Suzuki et al | 2009 | NR | Europe, United States, Japan | cohort | CT | whole blood immunoassay-triage D dimer |
| Sbarouni et al | 2007 | NR | Greece | cohort | CT, Echo | ELISA |
| Ohlmann et al | 2006 | 1997–2003 | France | cohort | CT, MRI, Echo | turbidimetric |
| Hazui et al | 2005 | 2001–2003 | Japan | case–control | CT | latex agglutination |
| Akutsu et al | 2005 | 2002–2004 | Japan | cohort | CT | latex agglutination |
| Eggebrecht et al | 2004 | 2002–2004 | Germany | cohort | CT, MRI, Echo, aortography | latex agglutination |
| Weber et al | 2003 | NR | Austria | cohort | CT, MRI, Echo, aortography | turbidimetric |
| Study | Sample Size | Age (years), Mean ± SD | Men (%) | Symptoms at Onset | Time of Enrollment | Mortality (%) |
|---|---|---|---|---|---|---|
| Suzuki et al | 87 | 61 ± 14 | 53 (61%) | NR | within first 24 hours of symptom onset | NR |
| Sbarouni et al | 18 | 53 ± 17 | 17 (94%) | chest pain (n = 11), syncope (n = 3), TIA (n = 2) | NR | 27.8 |
| Ohlmann et al | 94 | 64 ± 12 | 62 (66%) | chest pain (n = 66), syncope (n = 16), TIA (n = 24) | within first 15 days of symptom onset | 23.4 |
| Hazui et al | 29 | 65 ± 11 | 16 (55%) | NR | within first 4 hours of symptom onset | NR |
| Akutsu et al | 30 | 70 (60–76) ⁎ | 19 (63%) | NR | NR | 13.3 |
| Eggebrecht et al | 16 | 65 ± 15 | 11 (69%) | chest pain (n = 16) | within first 48 hours of symptom onset | 50 |
| Weber et al | 24 | 65 ± 14 | 11 (46%) | NR | NR | 41.7 |
| Study | All Patients With AAD | Stanford (%) | De Bakey (%) | DD Level (ng/ml), Mean ± SD | |||||
|---|---|---|---|---|---|---|---|---|---|
| A | B | I | II | III | Total | Stanford A | Stanford B | ||
| Suzuki et al | 87 | 64 (74%) | 23 (26%) | NR | NR | NR | 3,308 ± 1,455 | 3,213 ± 1,465 | 3,754 ± 1,430 |
| Sbarouni et al | 18 | 13 (72%) | 5 (28%) | NR | NR | NR | 4,630 ± 3,010 | NR | NR |
| Ohlmann et al | 94 | 67 (71%) | 27 (29%) | 48 (51%) | 19 (20%) | 27 (29%) | 8,610 (2,982–20,000) ⁎ | 9,260 ⁎ | 3,975 ⁎ |
| Hazui et al | 29 | NR | NR | 23 (79%) | 6 (21%) | 0 | 45,300 ± 68,000 | NR | NR |
| Akutsu et al | 30 | 12 (40%) | 18 (60%) | 8 (27%) | 3 (10%) | 19 (63%) | 1,800 (1,070–2,730) ⁎ | 1,250 (980–2,750) ⁎ | 1,800 (1,280–2,880) ⁎ |
| Eggebrecht e al | 16 | 6 (38%) | 10 (62%) | NR | NR | NR | 2,238 ± 1,765 | 2,872 ± 2,244 | 1,857 ± 1,401 |
| Weber et al | 24 | 12 (50%) | 12 (50%) | 4 (17%) | 8 (33%) | 12 (50%) | 9,400 | 8,800 ± 14,500 | 10,100 ± 14,800 |
| Reference | Number of Controls | Type of Subjects | DD level (ng/ml), Mean ± SD |
|---|---|---|---|
| Suzuki et al | 133 | angina (n = 37), MI (n = 46), PE (n = 5) | 402 ± 181 in patients with MI |
| Sbarouni et al | 29 | chronic AA (n = 21), healthy subjects (n = 8) | 203 ± 109 in healthy subjects, 1,278 ± 1,457 in patients with MI |
| Ohlmann et al | 94 | AA (n = 15), gastroesophageal (n = 6), HF (n = 2), MI (n = 21), musculoskeletal (n = 17), neuroradicular (n = 2), PE (n = 12), pericarditis (n = 7), shock (n = 5) | 625 (257–1,542) ⁎ |
| Hazui et al | 49 | MI (n = 49) | 400 ± 40 |
| Akutsu et al | 48 | AA (n = 7), AF (n = 3), gastroesophageal (n = 4), MI (n = 5), musculoskeletal (n = 8), other (n = 9), PE (n = 2), UA (n = 10) | 420 (20–1,380) ⁎ |
| Eggebrecht e al | 48 | MI (n = 16), other chest pain (n = 16), PE (n = 16) | 171 ± 100 in patients with MI |
| Weber et al | 35 | AS (n = 2), arrhythmia (n = 2), atypical chest pain (n = 5), HF (n = 5), MI (n = 10), PE (n = 1), UA (n = 10) | NR |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
