key: cord-0757652-e1yc09c3 authors: Nickel, Christian H.; Kellett, John; Cooksley, Tim; Lyngholm, Le E; Chang, Simon; Imfeld, Stephan; Bingisser, Roland; Brabrand, Mikkel title: The diagnoses and outcomes of emergency patients with an elevated D-dimer over the next 90 days date: 2020-07-12 journal: Am J Med DOI: 10.1016/j.amjmed.2020.06.009 sha: e13c2c3ccf1f41ab481e8e8b9f6f353759b4a4a8 doc_id: 757652 cord_uid: e1yc09c3 BACKGROUND: it is not known what diagnoses are associated with an elevated D-dimer in unselected patients attending emergency departments (ED), nor their associated outcomes. METHODS: prospective observational study of 1,612 unselected patients attending a Danish ED, with 100% follow up for 90 days after presentation. RESULTS: the 765 (47%) of ED patients with an elevated D-dimer level (i.e. >= 0.5 mg/L) were more likely to be admitted to hospital (p <0.0001), represent (p 0.02), and die within 90 days (i.e. 8.1% of patients, p <0.0001). Only ten patients with a normal D-dimer level (1.2%) died within 90 days: five had chronic obstructive pulmonary disease and infection, and five had cancer (four of whom also had infection). Venous thromboembolism, infection, neoplasia, anaemia, heart failure and unspecified soft tissue disorders were significantly associated with an elevated D-dimer. Of the 72 patients with venous thromboembolism 20 also had infection, 8 had cancer and 4 anaemia. None of the patients with heart failure, stroke or acute myocardial infarction with a normal D-dimer level died within 90 days. CONCLUSION: Nearly half of all patients attending an ED have an elevated D-dimer level and these patients were more likely to be admitted to hospital, and to represent and/or die within 90 days. While elevated D-dimer levels are significantly associated with venous thromboembolism, in this unselected ED patient population they are also associated with infection, cancer, heart failure and anaemia. Currently the main clinical use of D-dimer is to rule out venous thromboembolism. Elevated Ddimers occur in a variety of clinical scenarios including pneumonia, cardiac arrest and cancer. [1] [2] D-dimer is a non-specific biomarker that is immediately released by anything that causes the plasmin mediated proteolysis of fibrin [3, 4] and has been found to be increased in many conditions presenting to emergency departments [5, 6] . Although not diagnostic for any condition, an elevated D-dimer level is a powerful predictor of mortality. Elevated levels have been independently associated with an increased risk of death from any cause in an apparently healthy adult population [7] . However, it is not known what diagnoses are associated with a positive D-dimer in unselected emergency department (ED) patients, nor the outcomes associated with them. There is also concern that D-dimer's routine use as a risk-stratification tool on every ED patient might trigger futile expensive investigations for which there would otherwise be no clinical indication [8] . This study of unselected patients attending a Danish ED is a secondary analysis of previously published data that showed normal D-dimer levels identified patients at low risk of 30-day mortality [9] . It reports how many patients had an elevated D-dimer at presentation, what diagnoses were associated with elevated levels, and what happened to patients for up to 90 days after presentation. Study design: Secondary analysis of a prospective observational cohort study performed in an unselected population of adult medical patients attending an ED [9] . Setting: The Hospital of South West Jutland, a 450-bed regional teaching hospital in the region of Southern Denmark that serves approximately 220,000 inhabitants. Medical patients are referred to the ED by general practitioners (GP), outpatient clinics, out-of-hours GP service and emergency medical services. Participants: All non-trauma patients aged 18 years or older who required a blood sample for any clinical indication on arrival to the ED were eligible for inclusion in the study. As D-dimer levels can only be measured up to 10 h after the blood sample is initially collected, the small number of patients who arrived between 10 pm and 1 am could not be included in the study for logistic reasons. Blood tests, other than D-dimer, were requested at the discretion of the treating physician. Participants were asked for written informed consent before enrolment. Patients incapable of giving informed consent (e.g. language barriers or lacking mental capacity) were excluded. Patients could only be included in the study once, but all the representations to the health service over 90 days after ED presentation were considered. Patients also had to be registered in the Danish healthcare system so that their ICD10 codes and 90-day follow-up data could be obtained. Three trained research assistants performed the screening and inclusion process. All medical patients presenting to the ED were screened for eligibility between the 24 April 2017 and 19 August 2017. Data collection: D-dimers were measured in all included patients. Plasma D-dimer was quantitatively measured using a latex agglutination test (STA Liatest D-dimer (Diagnostica Stago, Asnieres-sur-Seine, France)). Citrate plasma for D-dimer estimation was obtained by centrifuging at 3500 rpm for 10 minutes. An elevated D-dimer level was defined as >=0.5 mg/L [10] . Outcome ascertainment: The final discharge diagnosis was obtained from the Danish National Patient Registry [11] . The discharge diagnoses of all patients were recorded according to The International Classification of Diseases and Related Health Problems, 10th Revision (ICD-10), produced by the World Health Organization [12] . There are more than 69,800 ICD-10 diagnosis codes and multiple codes may refer to similar conditions. ICD10 codes were, therefore, grouped as follows: all the codes that captured venous thromboembolism and cancer were agreed by consensus, arbitrated by an oncologist and angiologist (see supplemental data). A list previously published by Vest-Hansen et al [13] was used to identify all ICD10 codes associated with infection, and the remaining common conditions were identified accordingly: hypertension (ICD10 I10- Blinding: The treating physicians were unaware of the study during its implementation and were only given the D-dimer result if they had ordered it as part of the patients' care. This was done in order to avoid unnecessary investigations and treatment of potential venous thromboembolism that had not been suspected. All results were registered in a confidential research database which could only be accessed by the study investigators after its inclusion phase. The study design was approved by the Danish Regional Committee of Health Research Ethics (Identifier: S-20170005) and the Danish Data Protection Agency (Identifier: Region Syddanmark 2452). The study protocol was registered at ClinicalTrials.gov 3 April 2017, before enrolment of patients (ClinicalTrials.gov, Identifier: NCT03108807). The results are reported in accordance with the STROBE guidelines [14] . Statistics: Continuous data are presented as median (interquartile range) and categorical data as proportion (95% confidence intervals (CI)). The association between outcome (diagnoses made within 90 days) and D-dimer were presented as unadjusted odds ratios (OR) (95% CI), using Epi-Info version 6.0 (Centre for Disease Control and Prevention, USA). During the study period 1,612 patients registered in the Danish healthcare system presented to the ED and 995 (62%) were admitted for a mean length of stay of 4.7 SD 8.4 days. Patients who were admitted were significantly older than those discharged from the ED (65.9 SD 16.9 versus 58.1 SD 18.7 years, p <0.0001) and more likely to die within 90 days (6.5 versus 1.1%, p <0.0001). At the time presentation 765 of patients (47%) had a D-dimer level >= 0.5 mg/L: these patients were older, were assigned more ICD10 codes, were more likely to be admitted to hospital with a longer length of stay after admission and more likely to represent to the health services and/or die within 90 days than those with a D-dimer level <0.5 mg/L/ (Table 1) . After discharge from either the ED or the hospital 601 patients (37%) represented to the health services somewhere in Denmark within 90 days. Patients with an elevated D-dimer were 1.29 (95% CI 1.04-1.58, Chi square 5.68, p 0.02) times more likely to represent, and patients who represented were 3.14 (95% CI 1.86 -5.31, Chi square 21.64, p <0.00001) times more likely to die within 90 days (Figure 1 ). At the first presentation 5,257 ICD10 codes were recorded (3.3 per patient): 37% were "nonspecific factors" (ICD10 Chapter Z), and 28% were "disorders of the circulation" (ICD10 Chapter I). Within 90 days of ED presentation there were only thirteen common diagnostic groupings assigned to more than ten patients: infection was the commonest (24% of patients), followed by hypertension (9% of patients), chronic obstructive pulmonary disease (8% of patients) and neoplasia (7% of patients). Of all the other diagnostic groupings assigned both at presentation and at representation within 90 days only venous thromboembolism, infection, neoplasia, anaemia, heart failure and unspecified soft tissue disorders significantly associated with an elevated D-dimer level ( Table 2) . Many patients with a raised D-dimer level had more than one diagnostic grouping significantly associated with D-dimer elevation. For example, of the 72 patients with venous thromboembolism 20 also had infection, 8 had cancer and 4 anaemia (Table 3) . Cancer, infection, anaemia and heart failure were all associated with an increased 90-day mortality, whereas venous thromboembolism regardless of D-dimer was not (Table 4 ). Only ten patients with a normal D-dimer level (mean age 73.2 SD 10.5 years) died within 90 days of ED presentationall of them died between 24 and 73 days after presentation: five had chronic obstructive pulmonary disease and infection, and five had cancer (four of whom also had infection). None of the 12 patients diagnosed with venous thromboembolism who had a normal D-dimer level died within 90 days, nor did any of the patients with heart failure, stroke or acute myocardial infarction die if their D-dimer level was normal (Table 5) . This study shows that a low D-dimer at presentation to an ED makes death within 90 days, the need for hospital admission and the chance of subsequent re-presentation unlikely. An elevated D-dimer is associated with six diagnostic groupings, which ranked by prevalence are: infection, neoplasia, venous thromboembolism, anaemia, heart failure and unspecified soft tissue disorders: many patients have several of these diagnoses simultaneously. The diagnoses that were made in our study, both in the hospital and during 90-day follow-up, could not be scrutinized for accuracy. Therefore, we cannot be sure that venous thromboembolism and other diagnoses were not missed or overlooked in some patients, especially those with serious obvious disease such as metastatic cancer. Furthermore, we were not able to discern between active and inactive malignancy. Most of the patients in this single centre study were Caucasian. As D-dimer levels can vary in Afro-Caribbean and other racial groups [15, 16] , our findings need to be confirmed in an ethnically diverse population. Age influences the level of D-dimer for the diagnosis of venous thromboembolic [17] disease and pulmonary embolus [18] , but it is not known if this is true for other diagnoses or the prediction of mortality. Based on previous work on mortality risk we found that in to retain a good sensitivity and likelihood ratio that no age-adjustment should be performed [19] . Although we chose a standard cut-off for D-dimer levels of 0.50 mg/l, it is possible that this might not have been optimal for all the variables we examined. In addition, we did not control for factors that are known to be associated with elevated D-dimer levels such as heparin use and pregnancy [8] . The practice of medicine requires the formulation of a diagnosis, prognosis, and treatment. We examined all the diagnostic codes recorded and do not know which, if any, was a "cause" of the patients presenting illness to the ED: hypertension, for example, may just have been a commonly observed co-morbidity. The immediate treatment of a diagnosis that is not associated with imminent death or severe morbidity may not be required, especially if the treatment is risky and/or expensive. On the other hand, if there is a slight possibility of a diagnosis that carries a high risk of imminent mortality or morbidity treatment may be justified. This study confirms previous studies [19, 20] that reported D-dimer's major clinical benefit is its ability to identify patients in whom imminent death is highly unlikely, even in those patients with conditions usually associated with mortality. This initial study of only 1600 patients implies that D-dimer might be routinely measured on every patient presenting with an acute medical illness. The current coronavirus pandemic vindicates this suggestion, as COVID 19 patients with mild disease all had persistently normal D-dimer levels [21] . The objection to measuring D-dimer on every ED patient is that it would result in an increase in unnecessary investigations. D-dimer is a useful test to rule out venous thrombo-embolic disease in patients at low to intermediate risk, but a positive test must be interpreted with great caution as it could indicate venous thromboembolism but could also reflect the presence of a host of other conditions, either alone or in conjunction with venous thromboembolism. Since patients with an elevated D-dimer are at greater risk they urgently require clinical acumen and skill to address every possibility [22] , whereas a normal D-dimer level should allow the luxury of more time to make a diagnosis and consider appropriate treatment. Elevated D-dimer's diagnoses and outcomes 9 Nearly half of all patients attending an ED have an elevated D-dimer level and these patients were more likely to be admitted to hospital, to represent within 90 days and are seven times more likely to die during this time. While elevated D-dimer levels are associated with an increased risk for VTE, they are also associated with infection, cancer, heart failure and anaemia. JK, CHN, and MB contributed to the conceptualization of the study, data analysis, drafting and review of manuscript TC arbitrated the ICD10 code selection, and contributed to the conceptualization of the study, data analysis, drafting and review of manuscript LEL and SC collected and validated the data, and reviewed the manuscript SI arbitrated the ICD10 code selection and reviewed the manuscript RB contributed to the conceptualization of the study and review of manuscript All costs were borne by the authors. John Kellett is a major shareholder, director and chief medical officer of Tapa Healthcare DAC. The other authors have no potential conflicts of interest. Cancer codes VTE codes C139 I236B C159M I260 C160 I269 C169 I269A C178M I800 C179 I800B C180 I802 C182 I802B C183 I803 C183M I803B C184 I803C C184M I803E C185 I803F C187 I808 C189 I808A C189M I808B C209 I809 C209M I819 C220 I829 C220M I829B C221A Z921 C229 C240 C241 C249 C250 C250M C259 C259M C340A C341 C343 C343M C349 C349M C349X C412A C430 C438 C439 C439M C442 C443 C445 C447 C449 C499 C509 C509M C519 C519M C539 C539M C539X C549 C569 C579 C609 C619 C619M C649 C649M C649X C679 C699 C709X C711 C712 C713 C714 C718 C719 C739 C749 C770G C771 C771B C773 C779 C779A C780 C781 C782 C786 C786A C787 C790B C791I C793 C793A C795 C795B C795E C797 C798 C800M C809 C809M C810 C829 C830 C831 C833 C865 C880 C900 C910 C911 C914 C920 C920D C920F C921 C923 C929 C931 D032A D049 D095 D462A D462B D469 D630 E340 Z031 Z031A Z031B Z031BR Z031C Z031D Z031DA Z031DB Z031E Z031F Z031H Z031H1 Z031J Z031K1 Z031K2 Z031K3R Z031R Z031S Z031T Z031W Z031X Z031XAR Z031Y Z031YB Z031Z Z038E Z850D Z851 Z853 Z855 Z858 Z859 Z926 22. Nickel CH, Kuster T, Keil C, Messmer AS, Geigy N, Bingisser R. Risk stratification using D-dimers in patients presenting to the emergency department with nonspecific complaints. European journal of internal medicine 2016; 31: 20-4. Patients according to D-dimer level on presentation, numbers admitted to hospital, numbers representing to the Danish health service within 90 days of presentation, and mortality within 90 days. Table 4 Association between all diagnostic groupings associated and 90-day mortality. VTE = venous thrombo-embolism Plasma D-dimer levels correlate with outcomes in patients with community-acquired pneumonia Usefulness of the D-dimer concentration as a predictor of mortality in patients with out of hospital cardiac arrest Utility of cross-linked fibrin degradation products in the diagnosis of pulmonary embolism The Early Course of D-dimer Concentration following Pulmonary Artery Embolisation Role of fibrin D-dimer testing in emergency medicine Causes of elevated D-dimer in patients admitted to a large urban emergency department Association of D-dimer levels with all-cause mortality in a healthy adult population: findings from the MOLI-SANI study Appropriate Use of D-dimer in Hospital Patients A negative D-dimer identifies patients at low risk of death within 30 days: a prospective observational emergency department cohort study American Society of Hematology 2018 guidelines for management of venous thromboembolism: diagnosis of venous thromboembolism The Danish civil registration system Improving data quality: a guide for developing countries. Manila: WHO Regional Office for the Western Pacific Acute admissions to medical departments in Denmark: Diagnoses and patient characteristics Elevated D-dimer's diagnoses and outcomes 15 Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration Age, functional status, and racial differences in plasma D-dimer levels in community dwelling elderly persons New D-dimer threshold for Japanese patients with suspected pulmonary embolism: a retrospective cohort study Diagnostic Accuracy of Conventional or Age Adjusted D-dimer Cut-Off Values in Older Patients With Suspected Venous Thromboembolism: Systematic Review and Meta-Analysis Age-Adjusted D-Dimer Cutoff Levels to Rule Out Pulmonary Embolism. The ADJUST-PE Study Risk stratification using D-dimers in patients presenting to the emergency department with nonspecific complaints Combined use of the National Early Warning Score and D-dimer levels to predict 30-day and 365-day mortality in medical patients Clinical Characteristics of 138 Hospitalized Patients With Novel Coronavirus-Infected Pneumonia in Wuhan, China Table 5 : 90-day mortality of patients according to diagnostic grouping and D-dimer level, sorted by mortality of patients with normal D-dimer levels