|Common Sense - AAEM Member Magazine|
Edited By: Jay Khadpe, MD FAAEM and Michael C. Bond, MD FAAEM
Originally Published: November/December 2016 Common Sense
The D-dimer test is an important and widespread tool to assess for venous thromboembolic disease (VTE) in low risk patients. The test has a high sensitivity and negative predictive value; however it is also prone to false positives. Additionally, as D-dimer levels naturally rise with age, the test may lead to more frequent false positives in the elderly than the general population. Recently several investigations examined age related refinements to the interpretation of D-dimer results to rule out VTE. For this month’s resident journal review, we review two retrospective studies and one prospective study that evaluate using age-adjusted D-dimer levels to increase its specificity while retaining its sensitivity. Verification of the results of studies could reduce the use of expensive imaging studies, reduce patient exposure to radiation and contrast, and prevent unnecessary hospital admissions and anticoagulation. These issues are particularly pertinent for the elderly population.
Gupta A, Raja A, Ip I, Khorasani R. Assessing the 2 D-dimer age-adjusted strategies to optimize computer tomographic use in ED evaluation of pulmonary embolism. American Journal of Emergency Medicine. 2014;(32):1499-1502.
The authors set out to validate a prospective European study by Righini et al., which showed that using higher D-dimer cutoffs for patients older than 50 years were more specific for diagnosis pulmonary embolus (PE) and did not increase the false negative rate.1 This retrospective study took place from 2011-2013 in a single urban academic ED and used an automated quantitative D-dimer assay with a positive cutoff value of 500ng/mL. The study included 1,055 adult patients with D-dimer levels and computed tomography pulmonary angiograms (CTPAs). Using the traditional cutoff of 500ng/mL, sensitivity of D-dimer for PE was 100% (95% CI, 94.2%-100%) and specificity was 7.4% (95% CI, 5.8%-9.2%).
The authors then adjusted the D-dimer positive cutoff value based on age. Data was analyzed for individual ages (age in years x 10ng/mL) and in decade cohorts (e.g., patients 61-70 years old had a cutoff of 600ng/mL and 71-80 year olds had a cutoff of 700ng/mL). Using these cutoffs, decade-adjusted sensitivity was 98.7% (95% CI 92.1-99.9%) and specificity was 13.5% (95% CI 12.2-16.8%). Yearly-age-adjusted D-dimer had a sensitivity of 97.4% (95% CI 90.2%-99.6%) and a specificity of 16.7% (95% CI 14.4%-19.2%). The authors concluded that using decade-adjusted cutoffs would have avoided 37 CTPAs (or 19.6% of patients older than 60 with a Wells score ≤4) and using yearly-adjusted cutoffs would have avoided 52 (18%) CTPAs. They note that a 52-year-old and an 87-year-old with a PE would have been missed reducing the decade-adjusted sensitivities to 93.3% and 85.7%, respectively. However, these are not statistically different than the sensitivities calculated for the traditional cutoff.
This study is limited because it is a retrospective review at a single hospital center using a single D-dimer assay which severely limits its generalizability. Also the authors do not comment about the severity of the Pes in the patients who were missed by increasing the D-dimer threshold. Even so this study suggests that for patients with a low Wells score and D-dimer below an age-adjusted cutoff, PE may be safely ruled-out most of the time.
Adams D, Welch J, Kline J. Clinical utility of an age-adjusted D-dimer in the diagnosis of venous thromboembolism. Annals of Emergency Medicine. 2014;64(3):232-234.
Adams et al., performed a systematic review and meta-analysis on the use of age-adjusted D-dimer to diagnose VTE. Among the five studies included, there were 13 patient groups. Six of these groups had suspected deep vein thrombosis (DVT) and the other seven groups had suspected PE.
The prevalence of VTE was lowest in patients less than 50 years of age and highest in patients 71-80 years of age. The sensitivity of D-dimer values was similar between ages in the conventional and age-adjusted cutoffs. The specificity of conventional D-dimer testing decreased with age. For example, specificity was 66.8% in patients less than 50 years of age compared to 14.7% in patients older than 80 years of age. The specificity of age-adjusted D-dimer also decreased with increasing age, but to a smaller degree when compared to the traditional cutoff.
This systemic review determined that the use of an age-adjusted D-dimer cutoff in older patients with a non-high clinical probability of VTE improves specificity of the test without compromising sensitivity. In 12 of the 13 groups, the determination of non-high-risk patients was based on clinical decision rules such as the revised Geneva score or Wells score. In the remaining group, it was determined by clinician gestalt (estimated clinical probability of less than 80% for DVT).
There were several limitations to this review. Two different reference tests were used among the studies. In addition, the D-dimer assays were different in each study. Lastly varying D-dimer cutoff values were used in the different studies. Overall this systemic review supported the feasibility of using an age-adjusted D-dimer cutoff to avoid excessive testing for VTE in older patients.
Righini M, et al. Age-adjusted D-dimer cutoff levels to rule out pulmonary embolism; The ADJUST-PE Study. Journal of American Medical Association. 2014;311(11):1117-1124.
The ADJUST-PE trial was a multicenter, multinational, prospective study which evaluated if an age-adjusted D-dimer could improve the rate of PE exclusion without increasing false negative results in older patients with suspected PE. The primary outcome was the failure rate of a diagnostic strategy using age-adjusted D-dimer. This was defined as the rate of symptomatic VTE (proximal DVT or PE) during a three-month follow up period in those in whom PE was excluded based on a negative age-adjusted D-dimer and accordingly had anticoagulation withheld. Patients were enrolled if they presented to the ED with acute onset chest pain or dyspnea without an obvious explanation for their symptoms.
Patients were placed into groups based on their risk of PE by using either a dichotomized 2-level Wells score (likely versus unlikely) or the simplified, revised Geneva score (high versus non-high). D-dimer levels were measured in the latter groups and interpreted according to an age-adjusted cutoff. In patients <50 years old, the traditional 500mcg/L was used. The study initially risk-stratified 3,324 patients (median age of 63 years old). Of these patients, 2,898 (87.2%) were deemed as non-high risk and therefore underwent D-dimer measurements. 673 (23.2%) of these patients were older than 75. Of the non-high risk group, 817 (28.2%) were below the standard cutoff of 500mcg/L (95% CI, 26.6-29.9%), while 337 patients (11.6%) had a D-dimer above 500mcg/L but below the ageadjusted cutoff (95% CI, 10.5-12.9%). 1,744 (60%) were above the ageadjusted cutoff and went on to undergo diagnostic imaging with CTPA. Using the age-adjusted D-dimer level resulted in 11.6 % more patients being D-dimer negative. Per this pathway, these patients were considered to not have PE.
On three-month follow up of patients with D-dimer levels below 500mcg/mL, the rate of thromboembolism was 0.1% (95% CI, 0.0-0.7%); those with D-dimer levels between 500mcg/mL and the age-adjusted cutoff had a 0.3% rate of thromboembolism (95% CI, 0.1-0.7%). Rate of thromboembolism in those who had positive D-dimer values or who were risk stratified into a high-risk group and who had a negative CTPA was 0.5% (95% CI, 0.2-1.0%).
The authors also looked more specifically at patients above 75 years of age of whom 200 of 673 (29.7%) in total had negative D-dimer levels using the age-adjusted cutoffs (95% CI, 26.4-33.3%). None of these patients had thromboembolic events on three-month follow up. Onehundred fifty seven of the 673 patients older than 75 and in the non-high risk group had a D-dimer level below their age-adjusted cutoff (23.3%) while 43 were below the traditional 500mcg/mL cutoff, thus increasing the rule-out rate in this specific age group from 6.4% (95% CI, 4.8-8.5%) to 29.7% (95% CI, 26.4-33.3%).
This study demonstrates the clinical utility of using an age-adjusted D-dimer cutoff in patients above 50 years of age as compared to a set cutoff of 500mcg/mL. Use of the age-adjusted cutoff increases the proportion of patients that can effectively have PE excluded without significantly increasing false negative rates. This may significantly benefit older patients (>75 years of age) who would fall into a category of nearly five-fold increase in the PE exclusion rate with no significant increase in the false negative rate.
While this study demonstrates how an age-adjusted D-dimer cutoff may result in higher PE exclusion rates in older patient populations, it would be informative to compare an algorithm of this type with one in which a standard pre-selected cutoff (e.g., 500mcg/mL) is used independent of age was used. Failure rates between each arm, as well as important clinical outcomes such as mortality, could be potentially elucidated.
There is a growing body of evidence in the EM literature supporting the use of age-adjusted D-dimer levels to safely exclude VTE in patients at low- or intermediate-risk of VTE in the ED. There are several significant benefits of a higher cutoff value including saving the cost and resources of further imaging as well as protecting the patient from exposure to contrast, radiography, prolonged hospital stay, empiric anticoagulation, and evaluation of inconsequential incidental findings. The studies reviewed here, as well as one recently published in the Annals of Internal Medicine, support the implementation of an age-adjusted D-dimer in EDs.
1. Raja AS, Greenberg JO, Gaseem A, Denberg T, Fitterman N, Schuur, J. Evaluation of Patients With Supected Acute Pulmonary Embolism: Best Practice Advice From the Clinical Guidelines Committee of the American College of Physicians. Ann Intern Med 2015 [Epub ahead of print 29 September 2015] doi:10.7326/M14-1772.