Background

Pregnant women have a higher risk of venous thromboembolism (VTE) and an increasing D-Dimer over the course of pregnancy. The majority of clinical guidelines recommend the D-dimer should not be used to assess the risk of VTE in pregnant women. However, CTPE presents a radiation risk to both the mother and fetus. Unfortunately, most VTE studies exclude pregnant women. Recent studies have identified D-dimer as accurate measure or risk of VTE in conjunction with a clinical decision rule to identify those at low risk of VTE.

Clinical Question

In pregnant women at low to moderate risk of venous thromboembolism, with signs and/or symptoms of pulmonary embolism or deep vein thrombosis, is d-dimer accurate in identifying those with and without venous thromboembolism?

Design

This was a systematic review and meta-analysis that followed PRISMA guidelines and pre-defined inclusion criteria to identify studies of D-dimer to rule out venous thromboembolism in pregnant women signs and symptoms of pulmonary embolism or deep vein thrombosis. Studies required a reference standard of diagnostic imaging or clinical follow-up and follow-up at 3 months with sufficient data to calculate test characteristics. The use of multiple reference standard with differing accuracy could result in verification bias. Pre-test probability (PTP) was assessed as unlikely or low-intermediate by a clinical decision rule (YEARS or Revised Geneva) or clinical judgement. One study tested all patients regardless of pre-test probability.

The authors conducted an extensive search but did not present an assessment for publication bias.

Included studies were of high quality using QUADAS-2 criteria. One study was at high-risk of bias in patient selection. Inter-rater reliability for study inclusion and quality were not presented. The primary outcomes were sensitivity and predictive value of a negative test. Secondary outcomes included d-dimer yield (the proportion of patients with a negative d-dimer) and occurrence of VTE within 3 months of follow-up. 3 authors of the meta-analysis were authors of the largest included study.

Outcomes

Primary Outcome: Test Characteristics

Sensitivity and Predictive Value of a negative test for PE or DVT with a PTP of unlikely (<10%) or low-intermediate (10-30%)

Secondary Outcomes:

Diagnostic yield: Proportion of patients with a negative d-dimer

  1. All patients
  2. In each trimester and puerperium (time of placenta delivery to first few weeks)

Planned sensitivity analyses (study characteristics):

  1. Method of pre-test probability (PTP) assessment
  2. D-dimer type and cutoff
  3. Design: Prospective vs retrospective
  4. Reference standard: Imaging or 3-months follow-up or both
  5. Stratified by median QUADAS-2 score (assess distortion by low quality studies)

Inclusion

Inclusion: Prospective or retrospective studies of:

Studies:

  1. Plasma D-dimer
  2. Rule out venous thromboembolism (pulmonary embolism or deep vein thrombosis)
  3. Reference standard of diagnostic imaging or clinical follow-up
  4. Follow-up at 3 months
  5. Data sufficient to calculate test characteristics

Patients:

  1. 18 years, pregnant in any trimester or puerperium
  2. Presenting to an ED or outpatient clinic
  3. Signs and symptoms suggestive of PE or DVT without an obvious explanation:
    1. PE: Acute onset of chest pain and/or dyspnea, hemoptysis, tachycardia, tachypnea, hypotension or fainting
    2. DVT: Lower limp pain and/or swelling
  4. Estimated disease pre-test probability (PTP) of unlikely or low-intermediate by a clinical decision rule (prespecified) or irrespective of PTP when no CDR used

Excluded

None specified

Primary Results

Four studies were identified. These studies differed significantly in study definitions and design. 3 studies were prospective and 1 study was retrospective. 3 studies assessed pulmonary embolism and 1 study assessed DVT. Studies used different assays to assess d-dimer including three that were quantitative and 1 that was qualitative. Different d-dimer cutoffs were used. 2 studies used clinical decision rules to assess pre-test probability (Revised Geneva (1), YEARS (1)), 1 study used physician judgement and one study tested all patients without an assessment of pre-test probability. Patients were at different trimesters of pregnancy and different reference standards were used. Planned sensitivity analyses to assess the impact of some of these differences could not be performed due to an inadequate sample size. It is likely, that these studies should not be combined in a meta-analysis given these differences in definition and design.

The prevalence of VTE was highly variable among the studies (range 4.2-15.1% with a weighted mean of 5.0%, 95% CI (1.1, 11.4%). The sensitivity and predictive value of a negative test were high. The sensitivity was 99.5% (68/69), 95% CI (95.0, 100.0%). The negative predictive value was 100% (354/355), 95% CI (99.1, 100.0%). The proportion of patients with a negative d-dimer was 34.2%, 95% CI (15.9, 55.2%). There is a potential to reduce imaging. However, the imaging rate in the included was not presented for comparison. The overall rate of VTE at 3 months follow-up was 0.32% (1/312), (0.06, 1.8%).

Strengths

This was a systematic review and meta-analysis that followed PRISMA guidelines and pre-defined inclusion criteria to identify studies of D-dimer to rule out venous thromboembolism in pregnant women signs and symptoms of pulmonary embolism or deep vein thrombosis. Studies required a reference standard of diagnostic imaging or clinical follow-up and follow-up at 3 months with sufficient data to calculate test characteristics. The use of multiple reference standard with differing accuracy could result in verification bias. Pre-test probability (PTP) was assessed as unlikely or low-intermediate by a clinical decision rule (YEARS or Revised Geneva) or clinical judgement. One study tested all patients regardless of pre-test probability.

Limitations

The authors conducted an extensive search but did not present an assessment for publication bias.  Included studies were of high quality using QUADAS-2 criteria. One study was at high-risk of bias in patient selection. Inter-rater reliability for study inclusion and quality were not presented. The primary outcomes were sensitivity and predictive value of a negative test. Secondary outcomes included d-dimer yield (the proportion of patients with a negative d-dimer) and occurrence of VTE within 3 months of follow-up. 3 authors of the meta-analysis were authors of the largest included study.

The d-dimer is an objective laboratory test and there should be no concern for reproducibility. Reproducibility for the pre-test probability assessment using the Revised Geneva and YEARS algorithm was not presented. One of the criteria for the YEARS algorithm “PE is the most likely diagnosis” is very subjective. None of the studies were completed in the US.

Author's Conclusions

“In conclusion, the results of our meta-analysis suggest that D-dimer may be a safe and useful diagnostic tool in the management of pregnant women with suspected venous thromboembolism (VTE). However, limited data exist and further trials are needed to derive/validate specific clinical decision rules (CDR), and to identify the optimal D-dimer cutoff during pregnancy.”

Potential Impact To Current Practice

The results of this study should not impact current clinical practice. The study authors conclude that “further studies are needed to develop specific scores derived and validated in pregnant women for pre-test probability evaluation”.

The LEFt rule, a clinical decision rule for pregnant women with suspected deep vein thrombosis, was recently derived (Chan, Annals Intern Med 2009, PubMed ID: 19620161) and externally validated (Rhigini, Haematologica 2013, PubMed ID: 3659985) and a prospective validation is ongoing.