In this study, we conducted a comparative cost-effectiveness analysis of six different diagnostic strategies for the exclusion of VTE by D-dimer testing, utilizing data from a study conducted by Koch et al. in 2022 [11]. Our study population consisted of 526 cancer patients with suspected VTE who were recruited in the Chest Pain Unit of the University of Heidelberg (Heidelberg, Baden-Wuerttemberg, Germany). It is important to note that the cancer diagnosis preceded the occurrence of thromboembolism and the subsequent referral to the emergency department. The comparative group of cancer patients without the final adjudicated diagnosis of VTE encompassed a diverse array of conditions, including vascular diseases such as peripheral arterial disease and cardiac disease (e.g., hypertensive crisis, acute myocardial infarction), respiratory diseases (e.g., pneumonia, chronic obstructive pulmonary disease, bronchial asthma), gastrointestinal diseases (e.g., gastritis, gastroesophageal reflux), and orthopedic conditions (e.g., joint pain, arthritis, trauma) [11].

According to current guidelines, the diagnostic strategy for the exclusion of VTE involved a series of sequential steps within the emergency department [1]. This approach typically included a physical examination, an electrocardiogram (ECG), CUS or CTPA, and laboratory tests such as a complete blood count, creatinine assessment, D-dimer testing, and prothrombin time evaluation [11].

Upon admission, comprehensive data were gathered, encompassing patient characteristics, medical history, physical examination findings, diagnostic test results, and treatment details. In this analysis, the patients were categorized into distinct groups based on their pretest probability (PTP) of having PE or DVT, using the three-level Wells scores (< 2 points: low probability; 2–6 points: medium probability; ≥ 7 points: high probability). D-dimer testing was performed at the discretion of the attending physician immediately after the patient’s admission to the emergency department. If the D-dimer test yielded a negative result, it was considered evidence of excluding VTE, and no imaging or anticoagulant treatment had been initiated [11].

Following the institutional protocol and latest guidelines, patients with elevated D-dimer values and suspected VTE were further evaluated using either CUS or CTPA, except for high-risk patients with hemodynamic compromise, who received immediate rescue reperfusion therapy. Patients were deemed positive for VTE if they exhibited confirmed cases of PE or proximal DVT, as identified through diagnostic imaging tests. Figures 1 and 2 illustrate the diagnostic algorithm used for diagnosing VTE in our study.

Diagnostic algorithm for suspected DVT (first event) using compression ultrasonography, adapted according to current guidelines [1]. Pricings for CUS and D-dimer testing were taken from the German scale of fees for physicians (MFS) 51. Abbreviations: CUS, compression ultrasound; DVT, deep vein thrombosis; MFS, Medical fee schedule

Diagnostic algorithm for suspected PE in hemodynamically stable patients, adapted according to current guidelines [1]. All prices were taken from the German scale of fees for physicians (MFS) 51. Abbreviations: CUS, compression ultrasound; CTPA, computed tomography pulmonary angiography; DVT, deep vein thrombosis; EchoCG, echocardiogram; MFS, Medical fee schedule; PE, pulmonary embolism; RH, right heart; sPESI, simplified Pulmonary Embolism Severity Index; VTE, venous thromboembolism; V/Q-scintigraphy, ventilation-perfusion scintigraphy

To determine the most cost-effective method for VTE exclusion, we compared a conventional diagnostic strategy using the commonly recommended D-dimer rule-out cut-off 0.5 mg/L [1] (Method 1) together with five other methods based on different approaches to establish an optimal cut-off level for VTE exclusion according to current literature:

Method 2 (Age-adjusted cut-off [patient’s age × 0.01 mg/L]) [1]: The determination of the age-adjusted cut-off level was performed by multiplying the patient’s age by 10 in individuals who were over 50 years old. In this age-adjusted strategy, patients up to 50 years of age were considered negative for PE if their D-dimer levels were below 0.5 mg/L, while patients over 50 years of age were considered negative if their D-dimer levels were below 10 times their age.

Method 3 (Inverse age-adjusted cut-off [0.5 + (66-age) × 0.01 mg/L]): The establishment of the cut-off level in patients below 66 years was based on an age-adjusted criterion that operated inversely [11, 29, 30].

Method 4 (Increased fixed cut-off [1 mg/L]): The selection of the cut-off level was based on a higher fixed threshold of 1 mg/L [11, 30].

Method 5 (95%-Specificity cut-off [4.9 mg/L]): The selection of this cut-off level was based on a specificity of 95% that yielded a diagnostic threshold of 4.9 mg/L [11].

Method 6 (Receiver operating characteristic [ROC]-optimal cut-off [9.9 mg/L]): Prespecified ROC-optimized cut-off at 9.9 mg/L balancing sensitivities and specificities to establish an optimal detection threshold [11].

D-dimers were assessed in the central laboratory using the Innovance D-dimer assay, which had a diagnostic threshold of 0.5 mg/L (Siemens Healthineers AG, Forchheim, Germany). This assay utilizes a particle-enhanced, immune-turbidimetric method to quantitatively measure D-dimers in human plasma on dedicated coagulation analyzers (CS-5100; Siemens Healthineers AG) [11].

The analysis was conducted on the assumption that D-dimer was ordered for all patients with non-high PTP, and imaging tests were performed only when the test result exceeded the specified cut-off level of the respective diagnostic strategy. All monetary values utilized in the calculations were reported in € for Germany and in $ for the United States of America. In Germany, these expenses are invoiced using the German scale of fees for physicians (MFS, Medical fee schedule), which establishes standard fees for specific services. Accordingly, basic costs for CUS were €18.89, for D-dimer testing €24.13, and for CTPA €209.83 [31, 32]. Regarding the imaging costs in the US based on the published literature, which reflects the average expenses for patients with statutory health insurance, the prices were as follows: $184 for CUS, $14 for D-dimer testing, and $648 for CTPA [25, 33, 34].

To estimate the cost savings over one year, we adopted the presumption of a patient cohort consisting of 5475 individuals with cancer and suspected VTE (1825 with PE and 3650 with DVT), considering the data of our department (estimated average of 5 patients/day with PE and 10 patients/day with DVT).

The statistical analysis was conducted using MedCalc software (MedCalc Software Ltd., Version 22.016, Ostend, Belgium) [11]. The normality of data distribution was evaluated using the Kolmogorov–Smirnov test. Based on this, continuous variables were presented as mean ± standard deviation (SD) or as median with 25th/75th percentiles (interquartile range, IQR). Categorical variables were reported as numbers with corresponding percentages. To compare the D-dimer test results within the same group of patients, we used the Wilcoxon signed-rank test for paired samples and evaluated the correlation between the D-dimer test results and other variables, such as patient characteristics, risk factors, or clinical outcomes using Spearman’s coefficient of rank correlation. P values < 0.05 were considered statistically significant. Each diagnostic strategy’s sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), negative likelihood ratio (NLR), positive likelihood ratio (PLR), and the proportion of patients exhibiting a negative D-dimer test result were calculated and compared to those of the recommended cut-off level of 0.5 mg/L for ruling out VTE. Areas under the ROC curve (AUCs) were calculated using the methodology of DeLong. Additionally, ROC-optimized cut-offs were determined to strike a balance between sensitivities and specificities for identifying the optimal detection threshold for VTE.