Development of interpretable machine learning models to predict in-hospital prognosis of acute heart failure patients.

AIMS: In recent years, there has been remarkable development in machine learning (ML) models, showing a trend towards high prediction performance. ML models with high prediction performance often become structurally complex and are frequently perceived as black boxes, hindering intuitive interpretation of the prediction results. We aimed to develop ML models with high prediction performance, interpretability, and superior risk stratification to predict in-hospital mortality and worsening heart failure (WHF) in patients with acute heart failure (AHF). METHODS AND RESULTS: Based on the Kyoto Congestive Heart Failure registry, which enrolled 4056 patients with AHF, we developed prediction models for in-hospital mortality and WHF using information obtained on the first day of admission (demographics, physical examination, blood test results, etc.). After excluding 16 patients who died on the first or second day of admission, the original dataset (n = 4040) was split 4:1 into training (n = 3232) and test datasets (n = 808). Based on the training dataset, we developed three types of prediction models: (i) the classification and regression trees (CART) model; (ii) the random forest (RF) model; and (iii) the extreme gradient boosting (XGBoost) model. The performance of each model was evaluated using the test dataset, based on metrics including sensitivity, specificity, area under the receiver operating characteristic curve (AUC), Brier score, and calibration slope. For the complex structure of the XGBoost model, we performed SHapley Additive exPlanations (SHAP) analysis, classifying patients into interpretable clusters. In the original dataset, the proportion of females was 44.8% (1809/4040), and the average age was 77.9 ± 12.0. The in-hospital mortality rate was 6.3% (255/4040) and the WHF rate was 22.3% (900/4040) in the total study population. In the in-hospital mortality prediction, the AUC for the XGBoost model was 0.816 [95% confidence interval (CI): 0.815-0.818], surpassing the AUC values for the CART model (0.683, 95% CI: 0.680-0.685) and the RF model (0.755, 95% CI: 0.753-0.757). Similarly, in the WHF prediction, the AUC for the XGBoost model was 0.766 (95% CI: 0.765-0.768), outperforming the AUC values for the CART model (0.688, 95% CI: 0.686-0.689) and the RF model (0.713, 95% CI: 0.711-0.714). In the XGBoost model, interpretable clusters were formed, and the rates of in-hospital mortality and WHF were similar among each cluster in both the training and test datasets. CONCLUSIONS: The XGBoost models with SHAP analysis provide high prediction performance, interpretability, and reproducible risk stratification for in-hospital mortality and WHF for patients with AHF.

Clinical outcomes after discontinuing anticoagulant therapy in patients with first unprovoked venous thromboembolism.

BACKGROUND: The duration of anticoagulation for a first episode of unprovoked venous thromboembolism (VTE) should balance the likelihood of VTE recurrence against the risk of major bleeding. OBJECTIVES: Analyze rates and case-fatality rates (CFRs) of recurrent VTE and major bleeding after discontinuing anticoagulation in patients with a first unprovoked VTE after at least 3 months of anticoagulation. PATIENTS/METHODS: We compared the rates and CFRs in patients of the RIETE and COMMAND-VTE registries. We used logistic regression models to identify predictors for recurrent pulmonary embolism (PE) and major bleeding. RESULTS: Of 8,261 patients with unprovoked VTE in RIETE, 4,012 (48.6%) had isolated deep vein thrombosis (DVT) and 4,250 had PE. Follow-up (median, 318 days) showed 543 recurrent DVTs, 540 recurrent PEs, 71 major bleeding episodes, and 447 deaths. The COMMAND VTE registry yielded similar results. Corresponding CFRs of recurrent DVT, PE, and major bleeding, were 0.4%, 4.6%, and 24%, respectively. On multivariable analyses, initial PE presentation (hazard ratio (HR): 3.03; 95%CI: 2.49-3.69), dementia (HR: 1.47; 95%CI: 1.01-2.13), and anemia (HR: 0.72; 95%CI: 0.57-0.91) predicted recurrent PE, whereas older age (HR: 2.11; 95%CI: 1.15-3.87), inflammatory bowel disease (HR: 4.39; 95%CI: 1.00-19.3) and anemia (HR: 2.24; 95%CI: 1.35-3.73) predicted major bleeding. Prognostic scores were formulated, with c-statistics of 0.63 for recurrent PE and 0.69 for major bleeding. CONCLUSIONS: Recurrent DVT and PE were frequent but had low CFRs (0.4% and 4.6%, respectively) after discontinuing anticoagulation. Contrary, major bleeding was rare, but had high CFR (24%). A few clinical factors may predict these outcomes.

Statins use and recurrent venous thromboembolism in the direct oral anticoagulant era: insight from the COMMAND VTE Registry-2.

Statins were reported to have a potential effect of primary prevention of venous thromboembolism (VTE), although that of secondary prevention remains uncertain. To investigate the association between statins use and recurrent VTE in the current era. The COMMAND VTE Registry-2 is a multicenter registry enrolling 5,197 consecutive VTE patients among 31 centers in Japan between January 2015 and August 2020. We divided the entire cohort into 2 groups according to statins use at the time of discharge; the statins (N = 865) and no statins groups (N = 4332). The statins group was older (72.9 vs. 66.7 years, P < 0.001), and less often had active cancer (22.0% vs. 30.4%, P < 0.001). The cumulative incidence of discontinuation of anticoagulation was significantly lower in the statins group (60.3% vs. 52.6%, Log-rank P < 0.001). The cumulative 5-year incidence of recurrent VTE was significantly lower in the statins group (6.8% vs. 10.1%, Log-rank P = 0.01). Even after adjusting for the confounders, the lower risk of the statins group relative to the no statins group remained significant for recurrent VTE (HR 0.65, 95% CI 0.45-0.91, P = 0.01). The cumulative 5-year incidence of major bleeding was significantly lower in the statins group (12.2% vs. 14.1%, Log-rank P = 0.04), although, after adjusting for the confounders, the risk of the statins group relative to the no statins group turned to be insignificant (HR 0.77, 95% CI 0.59-1.00, P = 0.054). In this large real-world VTE registry, statins use was significantly associated with a lower risk for the recurrent VTE in the current era.

Prophylactic Anticoagulation and Thrombosis in Hospitalized Patients with Clinically Stable COVID-19 at Admission: From the Practice-Based Observational Study.

Objectives: The potential benefit of routine prophylactic anticoagulation for all hospitalized patients with clinically stable coronavirus disease 2019 (COVID-19) is still controversial. Method: The CLOT-COVID Study was a multicenter observational study enrolling 2894 consecutive hospitalized patients with COVID-19. The current study population consisted of 1738 hospitalized patients with mild COVID-19 at admission not requiring oxygen administration, who were divided into 2 groups: patients with prophylactic anticoagulation (n = 326) and those without (n = 1412). Results: Patients with prophylactic anticoagulation had more severe status of the worst severity of COVID-19 during hospitalization compared with those without (mild: 38% versus 82%, moderate: 55% versus 17%, and severe or death at discharge: 6.4% versus 0.7%, P <0.001). During hospitalization, 8 patients (0.5%) developed thrombosis, and the incidences of thrombosis were numerically higher in patients with more severe status of worst severity of COVID-19 during hospitalization (mild: 0.2%, moderate: 1.2%, and severe or death at discharge: 3.2%). Conclusions: Among hospitalized patients with clinically stable COVID-19 at admission, patients who did not worsen in COVID-19 severity after admission rarely developed thrombosis, although patients with worsening of COVID-19 severity after admission more often received prophylactic anticoagulation and might have a higher risk of thrombosis.

Subclass phenotypes in patients with unprovoked venous thromboembolisms using a latent class analysis.

BACKGROUND: Patients with unprovoked venous thromboembolisms (VTEs) can be sub-classified based on the different phenotypes using a latent class analysis (LCA), which might be useful for selecting individual management strategies. METHODS: In the COMMAND VTE Registry-2 database enrolling 5197 VTE patients, the current derivation cohort consisted of 1556 patients with unprovoked VTEs. We conducted clustering with an LCA, and the patients were classified into subgroups with the highest probability. We compared the clinical characteristics and outcomes among the developed subgroups. RESULTS: This LCA model proposed 3 subgroups based on 8 clinically relevant variables, and classified 592, 813, and 151 patients as Class I, II, and III, respectively. Based on the clinical features, we named Class I the younger, Class II the older with a few comorbidities, and Class III the older with many comorbidities. The cumulative 3-year anticoagulation discontinuation rate was highest in the older with many comorbidities (Class III) (39.9 %, 36.1 %, and 48.4 %, P = 0.02). There was no significant difference in the cumulative 5-year incidence of recurrent VTEs among the 3 classes (12.8 %, 11.1 %, and 4.0 % P = 0.20), whereas the cumulative 5-year incidence of major bleeding was significantly higher in the older with many comorbidities (Class III) (7.8 %, 12.7 %, and 17.8 %, P = 0.04). CONCLUSION: The current LCA revealed that patients with unprovoked VTEs could be sub-classified into further phenotypes depending on the patient characteristics. Each subclass phenotype could have different clinical outcomes risks especially a bleeding risk, which could have a potential benefit when considering the individual anticoagulation strategies. CLINICAL TRIAL REGISTRATION: URL: http://www.umin.ac.jp/ctr/index.htm COMMAND VTE Registry-2: Unique identifier, UMIN000044816 COMMAND VTE Registry: Unique identifier, UMIN000021132.

Potential treatment option of rivaroxaban for breastfeeding women: A case series.

The use of direct oral anticoagulants (DOACs) in breastfeeding women is currently challenging due to limited safety data for breastfeeding infants, and there have been no previous studies on the drug concentration in breastfeeding infants. We treated 2 patients (one case was twin pregnancy) with venous thromboembolisms in breastfeeding women administered rivaroxaban at our institution. Blood samples from the mothers and breastmilk samples were collected at time 0 and 2 h after the rivaroxaban administration, breastfeeding was conducted 2 h after the rivaroxaban administration, and blood samples from the infants were collected 2 h after breastfeeding (4 h after maternal rivaroxaban administration). The milk-to-plasma (M:P) ratios were 0.27 in Case 1 and 0.32 in Case 2. The estimated relative infant dose (RID) was 0.82 % in Case 1 Children 1 and 2, and 1.27 % in Case 2. The rivaroxaban concentration in the infant plasma was below the lower limit of quantification in all infants. In addition, even in the high-exposure case simulation based on 5 days of breastfeeding in Case 2, the infant plasma concentration level was below the lower limit of quantification. At 3 months of follow-up, breastfeeding was continued, and all infants grew and developed without any health problems including bleeding events. The current case series showed that there were no pharmacokinetic or clinical concerns for breastfeeding women or breastfed infants, and provides support for rivaroxaban as a safe treatment option for these patients.

Edoxaban, Rivaroxaban, or Apixaban for Cancer-Associated Venous Thromboembolism in the Real World: Insights from the COMMAND VTE Registry-2.

BACKGROUND: Real-world data on clinical characteristics and outcomes related to the use of different direct oral anticoagulants (DOACs) for cancer-associated venous thromboembolism (VTE) is lacking. METHODS: The COMMAND VTE Registry-2 is a multicenter registry enrolling 5,197 consecutive patients with acute symptomatic VTE from 31 centers in Japan from January 2015 to August 2020. Our study population comprised 1,197 patients with active cancer who were divided into the edoxaban (N = 643, 54%), rivaroxaban (N = 297, 25%), and apixaban (N = 257, 22%) groups. RESULTS: The cumulative 5-year incidence of recurrent VTE (9.3, 10.2, and 8.5%, respectively, p = 0.82) and all-cause death (67.5, 66.8, and 63.8%, respectively, p = 0.22) did not differ among the groups. Despite adjusting for confounders, the risks of recurrent VTE and all-cause death did not differ significantly among the groups. The cumulative 5-year incidence of major and clinically relevant bleeding was significantly lower in the rivaroxaban group than those in the other groups (22.6, 14.0, and 22.8%, p = 0.04; and 37.6, 26.8, and 38.3%, p = 0.01, respectively). After adjusting for confounders, in the rivaroxaban group, the risk for major bleeding was numerically lower (hazard ratio [HR]: 0.65, 95% confidence interval [CI]: 0.40-1.01) and that of clinically relevant all bleeding was significantly lower (HR: 0.67, 95% CI: 0.48-0.92) than those in the edoxaban group. CONCLUSION: The risks of recurrent VTE and all-cause death did not differ significantly among the different DOACs ; however, the risk of bleeding events could differ, with a potentially lower risk of bleeding with rivaroxaban.

Edoxaban for 12 Versus 3 Months in Cancer-associated Isolated Distal Deep Vein Thrombosis According to Different Doses: Insights from the ONCO DVT study.

BACKGROUND: The ONCO DVT study revealed the superiority of 12-month relative to 3-month edoxaban treatment for cancer-associated isolated distal deep vein thrombosis (DVT) regarding the thrombotic risk. METHODS: In this pre-specified subgroup analysis of the ONCO DVT study, we stratified the patients into those with a standard edoxaban dose (60 mg/day; N=151) and those with a reduced edoxaban dose (30 mg/day; N=450) and evaluated the clinical outcomes for the 12-month and 3-month treatments. RESULTS: The cumulative 12-month incidence of symptomatic recurrent venous thromboembolism was lower in the 12-month than 3-month group for both the 60 mg (1.3% vs. 11.6%, P=0.02; odds ratio [OR], 0.12; 95% CI, 0.01-0.97) and 30 mg (1.1% vs. 7.6%, P=0.002; OR, 0.14; 95% CI, 0.03-0.60) edoxaban subgroups, which was consistent across the edoxaban doses without a significant interaction (P =0.90). The 12-month cumulative incidence of major bleeding was higher in the 12-month group than 3-month group for the 60 mg edoxaban subgroup (14.3% vs. 4.4%, P=0.046; OR, 3.61; 95% CI, 0.97-13.52), whereas it did not significantly differ between the two groups for the 30 mg edoxaban subgroup (8.7% vs. 8.6%, P=0.89; OR, 0.97; 95% CI, 0.49-1.91), signaling there was a potential interaction (P=0.07). CONCLUSIONS: A 12-month edoxaban regimen for cancer-associated isolated distal DVT was consistently superior to a 3-month regimen, across the edoxaban doses for the thrombotic risk. However, caution was suggested for the standard dose of edoxaban due to the potential for an increased risk of bleeding with prolonged anticoagulation therapy.

Direct oral anticoagulant-associated bleeding complications in patients with gastrointestinal cancer and venous thromboembolism.

BACKGROUND: Direct oral anticoagulants (DOACs) have become widely used for cancer-associated venous thromboembolism (VTE). However, DOAC-associated bleeding complications remain challenging, especially in patients with gastrointestinal (GI) cancer. This study aimed to compare the bleeding outcomes between patients with upper or lower GI cancers and those without GI cancer. METHODS: Using the COMMAND VTE Registry-2 database, which is a multicenter registry enrolling 5197 consecutive acute symptomatic VTE patients among 31 centers in Japan between January 2015 and August 2020, we identified 1149 active cancer patients with DOACs (upper GI cancer: N = 88; lower GI cancer: N = 114; non-GI cancer: N = 947). The primary outcome was major bleeding during anticoagulation therapy, which was evaluated in the competing risk regression model. RESULTS: The upper GI cancer group had a lower mean body weight, and most often had anemia. The cumulative 5-year incidence of major bleeding was higher in the upper GI cancer group (upper GI cancer: 22.4 %, lower GI cancer: 15.4 %, and non-GI cancer: 11.6 %, P = 0.015). The most frequent major bleeding site in the upper GI cancer group was the upper GI (53 %), followed by the lower GI (24 %). After adjusting for the confounders, the excess risk in upper GI cancer relative to non-GI cancer remained significant for major bleeding (adjusted subhazard ratio, 2.25; 95 %CI, 1.31-3.87, P = 0.003), but that in lower GI cancer was insignificant. CONCLUSIONS: Upper GI cancer, but not lower GI cancer, as compared to non-GI cancer was associated with a higher risk for major bleeding during anticoagulation therapy with DOACs. CLINICAL TRIAL REGISTRATION: URL: http://www.umin.ac.jp/ctr/index.htm Unique identifier: UMIN000044816.

Home Treatment for Active Cancer Patients With Low-Risk Pulmonary Embolism　- A Predetermined Companion Report From the ONCO PE Trial.

BACKGROUND: Patients with appropriately selected low-risk pulmonary embolism (PE) can be treated at home, although it has been controversial whether applies to patients with cancer, who are considered not to be at low risk.Methods and Results: The current predetermined companion report from the ONCO PE trial evaluated the 3-month clinical outcomes of patients with home treatment and those with in-hospital treatment. The ONCO PE trial was a multicenter, randomized clinical trial among 32 institutions in Japan investigating the optimal duration of rivaroxaban treatment in cancer-associated PE patients with a score of 1 using the simplified version of the Pulmonary Embolism Severity Index (sPESI). Among 178 study patients, there were 66 (37%) in the home treatment group and 112 (63%) in the in-hospital treatment group. The primary endpoint of a composite of PE-related death, recurrent venous thromboembolism (VTE) and major bleeding occurred in 3 patients (4.6% [0.0-9.6%]) in the home treatment group and in 2 patients (1.8% [0.0-4.3%]) in the in-hospital treatment group. In the home treatment group, there were no cases of PE-related death or recurrent VTE, but major bleeding occurred in 3 patients (4.6% [0.0-9.6%]), and 2 patients (3.0% [0.0-7.2%]) required hospitalization due to bleeding events. CONCLUSIONS: Active cancer patients with PE of sPESI score=1 could be potential candidates for home treatment.

Fragility and long-term clinical outcomes in patients with venous thromboembolism receiving direct oral anticoagulants: From the COMMAND VTE REGISTRY-2.

INTRODUCTION: There is limited data on the safety of direct oral anticoagulants (DOACs) in fragile patients with venous thromboembolism (VTE). MATERIALS AND METHODS: We used the COMMAND VTE Registry-2 enrolling patients with acute symptomatic VTE. The study population consisted of 3928 patients receiving DOACs, who were divided into fragile (2136 patients) and non-fragile groups (1792 patients). Fragility was defined as patients of age ≥ 75 years, creatinine clearance level ≤ 50 ml/min, and/or body weight ≤ 50 kg. RESULTS: The fragile group significantly more often received reduced doses of DOACs compared to the non-fragile group (51 % and 19 %, P < 0.001). The cumulative 5-year incidence of major bleeding was numerically higher in the fragile group than the non-fragile group (15.0 % and 11.1 %, P = 0.052), even with no significant excess risk after adjusting for confounders (HR 1.03, 95%CI 0.81-1.31, P = 0.78). The cumulative 5-year incidence of clinically relevant bleeding was significantly higher in the fragile group than the non-fragile group (28.6 % and 19.6 %, P < 0.001), even after adjusting for confounders (HR 1.28, 95%CI 1.08-1.53, P = 0.005). There was no significant difference in cumulative 5-year incidence of recurrent VTE between the groups (9.6 % and 8.9 %, P = 0.68), which was consistent after adjusting for confounders (HR 1.13, 95%CI 0.84-1.51, P = 0.41). CONCLUSIONS: Among VTE patients receiving DOACs, fragile patients were associated with a numerically higher rate of major bleeding and a significantly increased risk of clinically relevant bleeding, but not an increased risk of recurrent VTE.

CRISPR-Cas9-guided amplification-free genomic diagnosis for familial hypercholesterolemia using nanopore sequencing.

Familial hypercholesterolemia is an inherited disorder that remains underdiagnosed. Conventional genetic testing methods such as next-generation sequencing (NGS) or target PCR are based on the amplification process. Due to the efficiency limits of polymerase and ligase enzymes, these methods usually target short regions and do not detect large mutations straightforwardly. This study combined the long-read nanopore sequencing and CRISPR-Cas9 system to sequence the target DNA molecules without amplification. We originally designed and optimized the CRISPR-RNA panel to target the low-density lipoprotein receptor gene (LDLR) and proprotein convertase subtilisin/kexin type 9 gene (PCSK9) from human genomic DNA followed by nanopore sequencing. The average coverages for LDLR and PCSK9 were 106× and 420×, versus 1.2× for the background genome. Among them, continuous reads were 52x and 307x, respectively, and spanned the entire length of LDLR and PCSK9. We identified pathogenic mutations in both coding and splicing donor regions in LDLR. We also detected an 11,029 bp large deletion in another case. Furthermore, using continuous long reads generated from the benchmark experiment, we demonstrated how a false-positive 670 bp deletion caused by PCR amplification errors was easily eliminated.

The impact of renal function on clinical outcomes of patients with cancer-associated isolated distal deep vein thrombosis: Insights from the ONCO DVT study.

BACKGROUND: The multicenter, open-label, randomized clinical trial ONCO DVT compared 3-month and 12-month edoxaban treatment regimens for isolated distal deep vein thrombosis (DVT) and suggested potential benefits of prolonged edoxaban treatment in terms of thrombotic risk. However, the risk-benefit balance of prolonged edoxaban treatment in patients with renal function remains unclear. OBJECTIVES: To compare the safety and efficacy of 3-month and 12-month edoxaban treatment regimens in patients with cancer-associated isolated distal DVT and different renal functions. METHODS: This pre-specified subgroup analysis of the ONCO DVT study included 601 patients divided into subgroups according to renal function using a 50 mL/min creatinine clearance (Ccr) cutoff. The primary endpoint was symptomatic recurrent venous thromboembolism (VTE) and VTE-related death at 12 months and the major secondary endpoint was major bleeding at 12 months. RESULTS: Among the 601 patients, 131 (21.8 %) comprised the renal dysfunction subgroup. The primary endpoint occurred in 6 (9.7 %) and 1 (1.4 %) patients in the 3-month and 12-month edoxaban groups in the renal dysfunction subgroup, respectively, and in 16 (6.6 %) and 2 (0.9 %) patients in the no renal dysfunction subgroup, respectively. The major secondary endpoint occurred in 9 (14.5 %) and 7 (10.1 %) patients in the 12-month and 3-month edoxaban groups in the renal dysfunction subgroup, and in 13 (5.3 %) and 21 (9.3 %) patients in the no renal dysfunction subgroup, respectively. CONCLUSIONS: A 12-month edoxaban regiment was superior to a 3-month treatment in terms of thrombotic risk irrespective of renal function. A higher bleeding risk was not identified in patients with renal dysfunction who received prolonged edoxaban treatment.

Transition of management strategies and long-term outcomes in cancer-associated venous thromboembolism from the warfarin era to the direct oral anticoagulant era.

BACKGROUND: There have been still limited data on the transition of management strategies and clinical outcomes after introduction of direct oral anticoagulant (DOAC) for cancer-associated venous thromboembolism (VTE) in the real-world clinical practice. METHODS: Using the 2 series of multicenter COMMAND VTE registries in Japan enrolling consecutive patients with acute symptomatic VTE, we compared 695 patients with cancer-associated VTE in the Registry-1 of the warfarin era and 1507 patients in the Registry-2 of the DOAC era. RESULTS: Regarding oral anticoagulation therapy, 576 patients (82.9 %) in the Registry-1 received warfarin, whereas 1119 patients (79.6 %) in the Registry-2 received DOACs. The cumulative 3-year incidence of discontinuation of anticoagulation was not significantly different between the 2 registries (56.7 % vs. 62.7 %, P = 0.11). The cumulative 5-year incidence of recurrent VTE was significantly lower in the Registry-2 than in the Registry-1 (17.7 % vs. 10.1 %, P < 0.001). The cumulative 5-year incidence of major bleeding was significantly lower in the Registry-2 than in the Registry-1 (26.6 % vs. 20.4 %, P = 0.045). The proportion of gastrointestinal bleeding numerically increased from the Registry-1 to the Registry-2 (46.7 % and 49.5 %), whereas that of intracranial bleeding numerically decreased from the Registry-1 to the Registry-2 (17.1 % and 14.1 %). CONCLUSIONS: In the current historical comparison of cancer-associated VTE between the 2 large real-world registries, there was a striking change in the treatment strategies with decreased risks of recurrent VTE and major bleeding in the DOAC era compared with those in the warfarin era, while there seemed to be unmet needs of DOAC-related gastrointestinal bleeding. CLINICAL TRIAL REGISTRATION: URL: http://www.umin.ac.jp/ctr/index.htm UNIQUE IDENTIFIER: UMIN000044816.

A Predictive Model for Cancer-Associated Thrombosis in Japanese Cancer Patients: Findings from the J-Khorana Registry.

Background Although the close relationship between cancer and venous thromboembolism (VTE) has been identified, risk stratification for VTE in Japanese patients with cancer remains unclear. Objectives This study aimed to validate the Khorana VTE risk assessment score (KRS) for VTE diagnosis and establish an optimal predictive model for VTE in Japanese patients with cancer. Methods A total of 7,955 Japanese patients with cancer were subdivided into low- (0), intermediate- (1-2), and high-score (3) groups according to the KRS. Using 37 explanatory variables, a total of 2,833 patients with cancer were divided into derivation and validation cohorts (5:5). A risk model for Japanese participants was developed using the derivation cohort data. Results The prevalence of VTE in low-, intermediate-, and high-score patients was 1.2, 2.5, and 4.3%, respectively. Logistic regression analysis demonstrated that cancer stage (III-IV) and KRS ≥ 2 were independent and significant predictors of VTE onset. The risk model for VTE assigned 1 point to body mass index ≥25 kg/m 2 and 2 points each to the prevalence of osteochondral cancer and D-dimer level ≥1.47 µg/mL. The areas under the curve of the risk model were 0.763 and 0.656 in the derivation and validation cohorts, respectively. Conclusion The KRS was useful in Japanese patients, and our new predictive model may be helpful for the diagnosis of VTE in Japanese patients with cancer.

Cancer-associated venous thromboembolism in the direct oral anticoagulants era: Insight from the COMMAND VTE Registry-2.

BACKGROUND: There is a paucity of data on real-world management strategies and clinical outcomes of cancer-associated venous thromboembolism (VTE) in the direct oral anticoagulants (DOACs) era. OBJECTIVES: To investigate the status of cancer-associated VTE in the DOAC era. METHODS: This multicenter, retrospective cohort study among 31 centers in Japan between 2015 and 2020 enrolled 5197 consecutive patients with acute symptomatic VTE, who were divided into 1507 patients (29 %) with active cancer and 3690 patients (71 %) without. RESULTS: The cumulative 3-year rate of anticoagulation discontinuation was significantly higher in patients with active cancer than in those without (62.7 % vs. 59.1 %, P < 0.001). The cumulative 5-year incidence of recurrent VTE was higher in patients with active cancer than in those without (10.1 % vs. 9.1 %, P = 0.01), however, after adjusting for the confounders and competing risk of mortality, the excess risk of the active cancer group relative to the no active cancer group was no longer significant (HR: 0.95, 95 % CI: 0.73-1.24). The cumulative 5-year incidence of major bleeding was much higher in the active cancer group (20.4 % vs. 11.6 %, P < 0.001). Even after adjusting for the confounders and competing risk of mortality, the risk of the active cancer group relative to the no active cancer group remained significant (HR: 1.36, 95 % CI: 1.11-1.66). CONCLUSIONS: The current large real-world registry revealed that the risk of major bleeding was still higher in patients with active cancer than in those without, leading to the frequent anticoagulation discontinuation, which has been still a huge challenge to overcome in the DOAC era.