Questioning the use of an age‐adjusted D‐dimer threshold to exclude venous thromboembolism: comment

Abstract

D‐dimer testing is a cornerstone of the diagnostic management of deep vein thrombosis (DVT) and pulmonary embolism (PE). Over the past decade, the threshold used to exclude venous thromboembolism (VTE) has transitioned from a fixed threshold to an age‐adjusted threshold 1.Righini M. Van Es J. Den Exter P.L. Roy P.‐.M. Verschuren F. Ghuysen A. Rutschmann O.T. Sanchez O. Jaffrelot M. Trinh‐Duc A. Le Gall C. Moustafa F. Principe A. Van Houten A.A. Ten Wolde M. Douma R.A. Hazelaar G. Erkens P.M.G. Van Kralingen K.W. Grootenboers M.J.J.H. et al.Age‐adjusted D‐dimer cutoff levels to rule out pulmonary embolism: the ADJUST‐PE study.JAMA. 2014; 311: 1117-24Crossref PubMed Scopus (582) Google Scholar and, more recently, to a clinical probability dependent threshold 2.van der Hulle T. Cheung W.Y. Kooij S. Beenen L.F.M. van Bemmel T. van Es J. Faber L.M. Hazelaar G.M. Heringhaus C. Hofstee H. Hovens M.M.C. Kaasjager K.A.H. van Klink R.C.J. Kruip M.J.H.A. Loeffen R.F. Mairuhu A.T.A. Middeldorp S. Nijkeuter M. van der Pol L.M. Schol‐Gelok S. et al.Simplified diagnostic management of suspected pulmonary embolism (the YEARS study): a prospective, multicentre, cohort study.Lancet. 2017; 390: 289-97Abstract Full Text Full Text PDF PubMed Scopus (285) Google Scholar. This transition intended to increase the proportion of patients in whom imaging can be withheld without jeopardizing safety. In a post‐hoc analysis of two prospective diagnostic management studies, Takach Lapner and colleagues found no differences in diagnostic accuracy between an age‐adjusted, an increased fixed and an inverse age‐adjusted D‐dimer threshold 3.Takach Lapner S. Julian J.A. Linkins L.‐.A. Bates S.M. Kearon C. Questioning the use of an age‐adjusted D‐dimer threshold to exclude venous thromboembolism: analysis of individual patient data from two diagnostic studies.J Thromb Haemost. 2016; 14: 1953-9Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar. The authors concluded that the improvement in efficiency with the age‐adjusted threshold is due to a non‐specific increase in mean D‐dimer threshold rather than an increasing D‐dimer threshold in older patients. To evaluate the value of the age‐adjusted D‐dimer threshold, we replicated the analyses from Takach Lapner and colleagues by combining data from several large prospective diagnostic management studies in patients with suspected PE. Individual patient data from six studies in which diagnostic management of PE was guided by the dichotomous Wells rule and D‐dimer testing was used. These studies were selected from a previously published systematic review 4.Van Es N. Der V.a.n. Hulle T. Van Es J. Den Exter P.L. Douma R.A. Goekoop R.J. Mos I.C.M. Galipienzo J. Kamphuisen P.W. Huisman M.V. Klok F.A. Buller H.R. Bossuyt P.M. Wells rule and d‐dimer testing to rule out pulmonary embolism a systematic review and individual‐patient data meta‐ Analysis.Ann Intern Med. 2016; 165: 253-61Crossref PubMed Scopus (102) Google Scholar. D‐dimer testing was performed with the INNOVANCE® (Siemens, Erlangen, Germany), VIDAS® (bioMérieux, Marcy‐l'Étoile, France), Tina‐quant® (Roche Diagnostics, Basel, Switzerland), and STA‐Liatest® (Stago, Asnières sur Seine, France) assays. We compared the efficiency and failure rate of four D‐dimer thresholds: (i) the conventional threshold of 500 μg L−1; (ii) an age‐adjusted threshold, defined as 500 μg L−1 in patients of 50 years or younger, and patient age multiplied by 10 μg L−1 in those older than 50 years; (iii) a mean D‐dimer threshold (600 μg L−1), which is the mean of the individual age‐adjusted thresholds in the present study; and (iv) an inverse age‐adjusted threshold, defined as patient age multiplied by 10 μg L−1 in those younger than 59 years, and as 500 μg L−1 in patients of 59 years or older. The latter represents the exact opposite of the age‐adjusted threshold, yet with the same overall mean D‐dimer threshold as the age‐adjusted strategy. Efficiency was defined as the proportion of patients in whom imaging could have been withheld based on a ‘PE unlikely’ Wells score and a D‐dimer below the threshold. The failure rate was defined as the proportion of patients with a ‘PE unlikely’ Wells score and a D‐dimer below the threshold in whom VTE was confirmed at baseline or during 3‐month follow‐up. Efficiency and failure rates for each strategy were computed as proportions with 95% confidence intervals (CI) based on the Wilson score method. Efficiency was compared between strategies using the McNemar test for correlated proportions. The number of patients reclassified by the mean and the inverse age‐adjusted strategy compared with the age‐adjusted strategy was evaluated. As the age‐adjusted threshold targets patients older than 50 years, subgroup analyses were performed in those 50 years or older and those younger than 50 years. All analyses were performed in R (R Foundation for Statistical Computing, Vienna, Austria, https://www.R-project.org). A P‐value below 0.05 was regarded as indicating statistical significance. Data from 7268 patients were available, of whom 908 were excluded for the present analysis because they were inpatients (n = 813), had missing age (n = 7) or had a ‘PE unlikely’ Wells score in combination with a missing D‐dimer result (n = 88). Of the remaining 6360 patients, 4732 (74%) were classified as ‘PE unlikely’ according to the Wells rule. The mean age was 56 years. Overall, 1269 (20%) were diagnosed with PE at baseline or VTE during 3‐month follow‐up. The efficiency and failure rates of the conventional, age‐adjusted, mean and inverse age‐adjusted thresholds are shown in Table 1, overall and in the subgroups of patients of 50 years or older and those younger than 50 years, as well as the difference in efficiency between the age‐adjusted and mean and inverse age‐adjusted strategies.Table 1Efficiency and failure rate of four different D‐dimer thresholdsConventional % (95% CI)Age‐adjusted % (95% CI)Mean % (95% CI)Inverse age‐adjusted % (95% CI)Absolute difference age‐adjusted and mean strategy % (95% CI; P‐value)Absolute difference age‐adjusted and inverse age‐adjusted strategy % (95% CI; P‐value)All patientsEfficiency32 (31–33)37 (36–39)37 (36–38)35 (34–37)0.2 (−1.5 − 1.9; P = 0.83)2.1 (0.4–3.8; P = 0.01)Failure rate0.58 (0.33–1.0)0.84 (0.54–1.3)1.1 (0.71–1.6)1.0 (0.68–1.5)––Age ≥ 50 yearsEfficiency22 (21–23)30 (29–31)27 (26–29)23 (21–24)2.8 (0.8–4.8; P = 0.006)7.4 (5.4–9.3; P < 0.001)Failure rate0.45 (0.17–1.1)0.98 (0.56–1.7)1.1 (0.62–1.9)0.54 (0.23–1.3)––Age <50 yearsEfficiency51 (49–53)51 (49–53)55 (53–58)59 (56–61)–4.2 (−1.2 to −7.1; P = 0.005)−7.2 (−4.2 to −10; P < 0.001)Failure rate0.69 (0.35–1.4)0.69 (0.35–1.4)1.0 (0.61–1.8)1.4 (0.86–2.1)––CI, confidence interval. Efficiency and failure rate of four D‐dimer thresholds in all patients, and in the subgroups ≥ 50 years and < 50 years: conventional threshold (500 μg L−1), age‐adjusted threshold (≤ 50 years, 500 μg L−1; > 50 years, age × 10 μg L−1), mean D‐dimer threshold (600 μg L−1), and inverse age‐adjusted threshold (< 59 years, 1090 μg L−1 – age × 10 μg L−1; ≥ 59 years, 500 μg L−1). Open table in a new tab CI, confidence interval. Efficiency and failure rate of four D‐dimer thresholds in all patients, and in the subgroups ≥ 50 years and < 50 years: conventional threshold (500 μg L−1), age‐adjusted threshold (≤ 50 years, 500 μg L−1; > 50 years, age × 10 μg L−1), mean D‐dimer threshold (600 μg L−1), and inverse age‐adjusted threshold (< 59 years, 1090 μg L−1 – age × 10 μg L−1; ≥ 59 years, 500 μg L−1). Reclassifications of the different D‐dimer strategies in the overall population are detailed in Tables S1–3. The mean D‐dimer threshold reclassified 142 patients with a positive age‐adjusted D‐dimer to a negative test result, of whom seven (4.9%) were diagnosed with VTE. Conversely, 155 patients were reclassified from a negative to a positive result, of whom two (1.3%) were diagnosed with VTE. The inverse age‐adjusted threshold reclassified 180 patients with a positive age‐adjusted D‐dimer as negative, of whom 11 (6.1%) were diagnosed with VTE. Conversely, 315 patients were reclassified from a negative to a positive result, of whom eight were diagnosed (2.5%) with VTE. In the subgroup of patients of 50 years or older the mean D‐dimer threshold reclassified 46 patients with a positive age‐adjusted D‐dimer to a negative result, of whom two (4.3%) were diagnosed with VTE. Conversely, 160 patients were reclassified from a negative to a positive test result, of whom two (1.3%) were diagnosed with VTE. The inverse age‐adjusted threshold reclassified 17 patients with a positive age‐adjusted D‐dimer to a negative result, of whom one (5.9%) was diagnosed with VTE. Conversely, 320 were reclassified from a negative to a positive test result, of whom eight were diagnosed with VTE (2.5%). In those younger than 50 years the age‐adjusted threshold is a fixed threshold of 500 μg L−1. In these patients the mean D‐dimer threshold reclassified 95 patients with a positive age‐adjusted D‐dimer to a negative result, of whom five (5.3%) were diagnosed with VTE. The inverse age‐adjusted threshold reclassified 162 patients with a positive age‐adjusted threshold to a negative result, of whom 10 (6.2%) were diagnosed with VTE. As the age‐adjusted threshold in those younger than 50 years is lower than the mean and inverse age‐adjusted thresholds by definition, none of the patients were reclassified from a negative age‐adjusted D‐dimer to a positive result. The present analysis shows that, overall, the efficiency of the age‐adjusted and mean D‐dimer thresholds is comparable and higher than with the inverse age‐adjusted strategy. In patients older than 50 years, an age‐adjusted D‐dimer threshold yielded a 3% to 7% higher efficiency than the mean D‐dimer and inverse age‐adjusted thresholds. In those younger than 50 years the age‐adjusted threshold is identical to the conventional threshold of 500 μg L−1. In this group the age‐adjusted threshold yielded a 4% to 7% lower efficiency than the mean and inverse age‐adjusted thresholds but was associated with the lowest failure rate. In the overall study group and in patients 50 years or older, a 3‐ to 4‐fold larger proportion of patients were diagnosed with VTE in those with a positive age‐adjusted D‐dimer but negative mean or inverse age‐adjusted D‐dimer than in those reclassified from negative to positive. Yet, the overall failure rate was acceptable for all D‐dimer thresholds, with the upper limit of the 95% CI well below the predefined safety margin of 3%. Takach Lapner and colleagues found no difference in the efficiency, failure rate and proportions of missed diagnoses between reclassifications with the three strategies 3.Takach Lapner S. Julian J.A. Linkins L.‐.A. Bates S.M. Kearon C. Questioning the use of an age‐adjusted D‐dimer threshold to exclude venous thromboembolism: analysis of individual patient data from two diagnostic studies.J Thromb Haemost. 2016; 14: 1953-9Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar. There are several possible explanations for the discrepancies between their findings and the results of the present analysis. First, VTE prevalence was considerably higher in the present analysis (20%) than in the study by Tackach Lapner and colleagues (6.6%). This may have led to differences in performance of the different thresholds, as the efficiency and failure rate are highly dependent on the proportion of patients with VTE 5.Dronkers C.E.A. van der Hulle T. Le Gal G. Kyrle P.A. Huisman M.V. Cannegieter S.C. Klok F.A. Towards a tailored diagnostic standard for future diagnostic studies in pulmonary embolism: communication from the SSC of the ISTH.J Thromb Haemost. 2017; 15: 1040-3Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar. Second, Takach Lapner and colleagues included patients with both suspected DVT and suspected PE. Thus far, the age‐adjusted threshold has only been prospectively validated in the management of PE. An ongoing study is evaluating age‐adjusted D‐dimer testing in the diagnostic management of DVT (ClinicalTrials.gov; NCT02384135). Until publication of the results, the performance of age‐adjusted D‐dimer testing in suspected DVT remains unclear. Third, the use of different D‐dimer assays may have influenced the performance of the different D‐dimer thresholds as the diagnostic accuracy may differ between assays 6.Di Nisio M. Squizzato A. Rutjes A.W.S. Büller H.R. Zwinderman A.H. Bossuyt P.M.M. Diagnostic accuracy of D‐dimer test for exclusion of venous thromboembolism: a systematic review.J Thromb Haemost. 2007; 5: 296-304Crossref PubMed Scopus (350) Google Scholar. All D‐dimer measurements in the study by Takach Lapner and colleagues were performed with the STA‐Liatest, whereas in the present analysis, four different D‐dimer assays were used. At present, there are scant data on the performance of different D‐dimer assays using the age‐adjusted threshold 7.Goodwin A.J. Higgins R.A. Moser K.A. Smock K.J. Chandler W.L. Kottke‐Marchant K. Hartman S.K. Volod O. Brown A.F. Johari V.P. Burr S. Polyakov N. Chen D. Issues surrounding age‐adjusted D‐dimer cutoffs that practicing physicians need to know when evaluating patients with suspected pulmonary embolism.Ann Intern Med. 2017; 5: 296-304Google Scholar, although one recent prospective VTE management study suggested comparable accuracy of four D‐dimer assays 8.Farm M. Siddiqui A.J. Onelöv L. Järnberg I. Eintrei J. Maskovic F. Kallner A. Holmström M. Antovic J.P. Age‐adjusted D‐dimer cut‐off leads to more efficient diagnosis of venous thromboembolism in the emergency department: a comparison of four assays.J Thromb Haemost. 2018; 16: 866-75Abstract Full Text Full Text PDF PubMed Scopus (22) Google Scholar. Last, the study by Takach Lapner and colleagues may have been underpowered to detect differences in performance of the three D‐dimer strategies. Two studies included in the present analysis were previously used to externally validate age‐adjusted D‐dimer testing 1.Righini M. Van Es J. Den Exter P.L. Roy P.‐.M. Verschuren F. Ghuysen A. Rutschmann O.T. Sanchez O. Jaffrelot M. Trinh‐Duc A. Le Gall C. Moustafa F. Principe A. Van Houten A.A. Ten Wolde M. Douma R.A. Hazelaar G. Erkens P.M.G. Van Kralingen K.W. Grootenboers M.J.J.H. et al.Age‐adjusted D‐dimer cutoff levels to rule out pulmonary embolism: the ADJUST‐PE study.JAMA. 2014; 311: 1117-24Crossref PubMed Scopus (582) Google Scholar, 9.Van Belle A. Büller H. Huisman M. Huisman P. Kaasjaer K. Kamphuisen P. Effectiveness of managing suspected pulmonary embolism using an algorithm combining clinical probability, D‐dimer testing, and computed tomography.JAMA. 2006; 295: 172Crossref PubMed Scopus (895) Google Scholar, which may have influenced the results in favor of the age‐adjusted threshold. Current guidance recommends the use of a diagnostic algorithm in patients with suspected PE, with the aim to safely exclude the diagnosis without additional imaging. The age‐adjusted threshold was proposed and validated to increase the specificity of D‐dimer testing in patients older than 50 years, thereby reducing the number of false positive D‐dimer results and avoiding unnecessary imaging tests 1.Righini M. Van Es J. Den Exter P.L. Roy P.‐.M. Verschuren F. Ghuysen A. Rutschmann O.T. Sanchez O. Jaffrelot M. Trinh‐Duc A. Le Gall C. Moustafa F. Principe A. Van Houten A.A. Ten Wolde M. Douma R.A. Hazelaar G. Erkens P.M.G. Van Kralingen K.W. Grootenboers M.J.J.H. et al.Age‐adjusted D‐dimer cutoff levels to rule out pulmonary embolism: the ADJUST‐PE study.JAMA. 2014; 311: 1117-24Crossref PubMed Scopus (582) Google Scholar, 10.Douma R.A. le Gal G. Söhne M. Righini M. Kamphuisen P.W. Perrier A. Kruip M.J.H.A. Bounameaux H. Büller H.R. Roy P.‐.M. Potential of an age adjusted D‐dimer cut‐off value to improve the exclusion of pulmonary embolism in older patients: a retrospective analysis of three large cohorts.BMJ. 2010; 340: c1475Crossref PubMed Scopus (249) Google Scholar. In this group, the age‐adjusted D‐dimer strategy was associated with a higher efficiency than strategies using a mean or inverse age‐adjusted D‐dimer threshold. Therefore, the present analysis supports an age‐adjusted D‐dimer threshold over a nonspecific increased threshold in the diagnostic management of PE. All authors were responsible for the concept and design of the letter. N. Kraaijpoel and N. van Es were responsible for data acquisition and statistical analysis. All authors interpreted the data, drafted the manuscript, critically revised the manuscript for important intellectual content and gave final approval of the manuscript. The authors state that they have no conflict of interest. Download .docx (.02 MB) Help with docx files Table S1. Reclassifications of age‐adjusted strategy and mean D‐dimer strategy.Table S2. Reclassifications of age‐adjusted strategy and inverse age‐adjusted D‐dimer strategy.Table S3. Reclassifications of age‐adjusted strategy and conventional D‐dimer strategy.