Prediction model for major bleeding in anticoagulated patients with cancer-associated venous thromboembolism using machine learning and natural language processing.

PURPOSE: We developed a predictive model to assess the risk of major bleeding (MB) within 6 months of primary venous thromboembolism (VTE) in cancer patients receiving anticoagulant treatment. We also sought to describe the prevalence and incidence of VTE in cancer patients, and to describe clinical characteristics at baseline and bleeding events during follow-up in patients receiving anticoagulants. METHODS: This observational, retrospective, and multicenter study used natural language processing and machine learning (ML), to analyze unstructured clinical data from electronic health records from nine Spanish hospitals between 2014 and 2018. All adult cancer patients with VTE receiving anticoagulants were included. Both clinically- and ML-driven feature selection was performed to identify MB predictors. Logistic regression (LR), decision tree (DT), and random forest (RF) algorithms were used to train predictive models, which were validated in a hold-out dataset and compared to the previously developed CAT-BLEED score. RESULTS: Of the 2,893,108 cancer patients screened, in-hospital VTE prevalence was 5.8% and the annual incidence ranged from 2.7 to 3.9%. We identified 21,227 patients with active cancer and VTE receiving anticoagulants (53.9% men, median age of 70 years). MB events after VTE diagnosis occurred in 10.9% of patients within the first six months. MB predictors included: hemoglobin, metastasis, age, platelets, leukocytes, and serum creatinine. The LR, DT, and RF models had AUC-ROC (95% confidence interval) values of 0.60 (0.55, 0.65), 0.60 (0.55, 0.65), and 0.61 (0.56, 0.66), respectively. These models outperformed the CAT-BLEED score with values of 0.53 (0.48, 0.59). CONCLUSIONS: Our study shows encouraging results in identifying anticoagulated patients with cancer-associated VTE who are at high risk of MB.

Development of a predictive model of venous thromboembolism recurrence in anticoagulated cancer patients using machine learning.

INTRODUCTION: Patients with cancer and venous thromboembolism (VTE) show a high risk of VTE recurrence during anticoagulant treatment. This study aimed to develop a predictive model to assess the risk of VTE recurrence within 6 months at the moment of primary VTE diagnosis in these patients. MATERIALS AND METHODS: Using the EHRead® technology, based on Natural Language Processing (NLP) and machine learning (ML), the unstructured data in electronic health records from 9 Spanish hospitals between 2014 and 2018 were extracted. Both clinically- and ML-driven feature selection were performed to identify predictors for VTE recurrence. Logistic regression (LR), decision tree (DT), and random forest (RF) algorithms were used to train different prediction models, which were subsequently validated in a hold-out data set. RESULTS: A total of 16,407 anticoagulated cancer patients with diagnosis of VTE were identified (54.4 % male and median age 70). Deep vein thrombosis, pulmonary embolism and metastases were observed in 67.2 %, 26.6 %, and 47.7 % of the patients, respectively. During the study follow-up, 11.4 % of the patients developed a recurrent VTE, being more frequent in patients with lung cancer. Feature selection and model training based on ML identified primary pulmonary embolism, deep vein thrombosis, metastasis, adenocarcinoma, hemoglobin and serum creatinine levels, platelet and leukocyte count, family history of VTE, and patients' age as predictors of VTE recurrence within 6 months of VTE diagnosis. The LR model had an AUC-ROC (95 % CI) of 0.66 (0.61, 0.70), the DT of 0.69 (0.65, 0.72) and the RF of 0.68 (0.63, 0.72). CONCLUSIONS: This is the first ML-based predictive model designed to predict 6-months VTE recurrence in patients with cancer. These results hold great potential to assist clinicians to identify the high-risk patients and improve their clinical management.

Estrategia de tratamiento de la dislipemia y consecución de objetivos en la población española con diabetes tipo 2 sin enfermedad cardiovascular / Therapeutic approach to dyslipidemia and goal achievement in a spanish population with type 2 diabetes without cardiovascular disease

Objetivo: Evaluar la estrategia terapéutica y el grado de consecución de los objetivos lipídicos en la población española con diabetes y alto riesgo cardiovascular. Sujetos y métodos: Estudio descriptivo, transversal y multicéntrico con inclusión mediante muestreo consecutivo de los 10 primeros pacientes que acudieron a consulta de Atención Primaria y que hubieran sido visitados durante los 12 meses previos al estudio. Se incluyeron pacientes con diabetes tipo 2 sin enfermedad cardiovascular, concentraciones de colesterol LDL(cLDL) ≤ 160 mg/dl y triglicéridos ≤ 600 mg/dl, y al menos uno de los siguientes: retinopatía, albuminuria, tabaquismo actual o hipertensión. Resultados: Se evaluaron 2.412 pacientes (edad: 61,3±8,3 años; 46,8% mujeres, duración de la diabetes de 8,6±7,4 años). En comparación con la visita previa (8,1±5 meses antes), en el momento de la evaluación, la proporción de pacientes con cLDL < 100 mg/dl (22,7 vs 28,6%),c-noHDL < 130 mg/dl (27,7 vs 33,8%) y la combinación de ambos (17,6 vs 22,1%) aumentaron significativamente. Las estatinas eran los fármacos hipolipemiantes más prescritos (65,5%) y,desde la visita previa, en el 38,7% de los pacientes se cambió el fármaco hipolipemiante, en el17,3% se aumentó la dosis y en un 5% se añadió otro fármaco. Conclusión: La utilización de estatinas de mayor potencia y el incremento de la dosis es la estrategia terapéutica más utilizada para mejorar el control de la dislipemia en los pacientes con diabetes tipo 2, pero estos cambios resultan claramente insuficientes para alcanzar los objetivos lipídicos en la mayoría de los pacientes con diabetes tipo 2 (AU)

Objective: To assess the therapeutic approach and lipid goal achievement in a spanish diabetic population at high cardiovascular risk. Subjects and methods: A multicenter, descriptive, cross-sectional study consecutively recruited the first 10 patients who attended the primary care office and had been seen in the12 months prior to the study visit. Inclusion criteria were type 2 diabetes without cardiovascular disease, LDL cholesterol levels ≤160 mg/dL, triglyceride levels ≤600 mg/dL, and at least one of the following: retinopathy, albuminuria, current smoking, or hypertension. Results: A total of 2412 patients were evaluated (aged 61.3±8.3 years, 46.8% women, diabetes duration 8.6±7.4 years). As compared to the previous visit (8.1±5 months before), the proportion of patients who achieved LDL-C levels <100 mg/dL (22.7% vs 28.6%), non-HDL-C levels<130 mg/dL (27.7% vs 33.8%) and both goals (17.6% vs 22.1%) significantly increased at the time of assessment. Statins were the most widely prescribed lipid-lowering drugs (65.5%) and the lipid-lowering drug was changed from the previous visit in 38.7% of patients, drug dosage was increased in 17.3%, and another drug was added in 5%.Conclusion: Use of more potent statins and higher statin doses were the most commonly used therapeutic strategies for improving control of dyslipidemia in patients with type2 diabetes, but these changes were clearly inadequate to achieve lipid goals in most patients with type 2 diabetes (AU)

[Therapeutic approach to dyslipidemia and goal achievement in a Spanish population with type 2 diabetes without cardiovascular disease]. / Estrategia de tratamiento de la dislipemia y consecución de objetivos en la población española con diabetes tipo 2 sin enfermedad cardiovascular.

OBJECTIVE: To assess the therapeutic approach and lipid goal achievement in a spanish diabetic population at high cardiovascular risk. SUBJECTS AND METHODS: A multicenter, descriptive, cross-sectional study consecutively recruited the first 10 patients who attended the primary care office and had been seen in the 12 months prior to the study visit. Inclusion criteria were type 2 diabetes without cardiovascular disease, LDL cholesterol levels ≤160mg/dL, triglyceride levels ≤600mg/dL, and at least one of the following: retinopathy, albuminuria, current smoking, or hypertension. RESULTS: A total of 2412 patients were evaluated (aged 61.3±8.3 years, 46.8% women, diabetes duration 8.6±7.4 years). As compared to the previous visit (8.1±5 months before), the proportion of patients who achieved LDL-C levels <100mg/dL (22.7% vs 28.6%), non-HDL-C levels <130mg/dL (27.7% vs 33.8%) and both goals (17.6% vs 22.1%) significantly increased at the time of assessment. Statins were the most widely prescribed lipid-lowering drugs (65.5%) and the lipid-lowering drug was changed from the previous visit in 38.7% of patients, drug dosage was increased in 17.3%, and another drug was added in 5%. CONCLUSION: Use of more potent statins and higher statin doses were the most commonly used therapeutic strategies for improving control of dyslipidemia in patients with type 2 diabetes, but these changes were clearly inadequate to achieve lipid goals in most patients with type 2 diabetes.

Therapeutic interventions and success in risk factor control for secondary prevention of stroke.

OBJECTIVE: We sought to evaluate the success rates in achieving preventive therapeutic goals in patients who experienced an ischemic stroke (IS) and compare them with those achieved in patients with coronary artery disease (CAD). METHODS: This was an observational multicenter case-control study (3 patients with IS and one control subject with CAD) performed in 1444 primary health centers in Spain. Preventive therapeutic objectives according to American Heart Association guidelines were predefined. Demographic data, vascular risk factors, and success/failure in achievement of objectives were recorded and compared between patients with IS and CAD. RESULTS: A total of 5458 patients were included, 4098 (75.1%) had IS and 1360 (24.9%) had CAD. Although more than 90% of patients with hypertension, diabetes, or dyslipidemia were under specific drug regimens, only about 25% achieved the recommended therapeutic objective for each risk factor. Success rate was especially low among patients with IS compared with CAD: hypertension (23.8% v 27.2%; P = .028); dyslipidemia (13.6% v 20.3%; P < .001); and abdominal obesity (49.1% v 54.6%; P = .002). The only objective widely achieved in both groups was the use of antithrombotic drugs in atrial fibrillation (97.2% v 94.7%; P = .125). Only 3.3% of patients with IS had all risk factors under control, compared with 5.6% of those with CAD (P = .006). For all patients, multivariate logistic regression model showed that independent predictors of full risk factor control were: presence of CAD as compared with IS (odds ratio [OR] 2.11; 95% confidence interval [CI] 1.35-3.29; P = .001), older age (OR 1.02; 95% CI 1.00-1.04; P = .028), and having less than 3 risk factors (OR 16.98; 95% CI 9.02-31.97; P < .001). CONCLUSIONS: Success in achieving preventive therapeutic objectives for secondary prevention of vascular events is low, especially among patients with IS. There is an urgent need to devise strategies to improve risk factor control.

Effect of doxazosin gastrointestinal therapeutic system on patients with uncontrolled hypertension: the ASOCIA Study.

OBJECTIVE: To evaluate extended-release doxazosin gastrointestinal therapeutic system (GITS) as add-on therapy in patients with treated, but uncontrolled hypertension. METHODS: A 16-week, open, noncomparative, multicenter, prospective study of patients with hypertension (> or = 140/> or = 90 mm Hg). Doxazosin GITS 4 mg/d was added to entry medication and increased to 8 mg/d at Week 4 in cases of inadequate blood pressure (BP) control. RESULTS: A total of 3631 patients (40% women) with mean age of 62.4 +/- 0.2 years were included. Proportion of patients reaching goal (< 140/< 90 mm Hg) after 4 weeks of add-on therapy with doxazosin GITS was 39% and increased to 61% at Week 16. Systolic and diastolic BP (mean +/- SEM) decreased, respectively, from 161.6 +/- 0.2 and 95.1 +/- 0.1 mm Hg at baseline to 142.2 +/- 0.2 and 84.1 +/- 0.1 mm Hg at Week 4 (P < 0.0001) and 136.8 +/- 0.2 and 80.6 +/- 0.2 mm Hg at Week 16 (P < 0.0001). Adverse events occurred in 108 patients (3.0%), with 57 (1.6%) related to the study treatment. In 17 patients (0.5%), serious adverse events were described, but only one was related to the study drug. CONCLUSIONS: Doxazosin GITS as add-on therapy achieved target blood pressure and was well tolerated in patients with hypertension uncontrolled by previous regimens. Doxazosin GITS efficacy and tolerability was achieved in combination with all classes of antihypertensives tested.

Possible role of parathyroid hormone-related protein as a proinflammatory cytokine in atherosclerosis.

BACKGROUND AND PURPOSE: Parathyroid hormone-related protein (PTHrP) is a vasodilator peptide. In addition, PTHrP appears to affect vascular growth and to be a mediator of inflammation in rheumatic and brain disorders. We examined the possible role of PTHrP in the inflammatory process in atherosclerosis METHODS: We immunohistochemically analyzed the cellular localization of PTHrP, the type 1 PTH/PTHrP receptor (PTH1R), and monocyte chemoattractant protein-1 (MCP-1) in 26 human carotid atherosclerotic plaques. RESULTS: The inflammatory region of plaques was characterized by high PTHrP, PTH1R, and MCP-1 immunostaining in relation to the cap (0.75+/-0.1 versus 0.29+/-0.04, 0.5+/-0.1 versus 0.25+/-0.05, 0.72+/-0.2 versus 0.29+/-0.05, respectively; P<0.05). PTHrP and MCP-1 were colocalized in both resident and inflammatory cells in the plaque. Moreover, in cultured vascular smooth muscle cells (VSMC), PTHrP(1-36) increased MCP-1 mRNA (3-fold at 6 hours) and MCP-1 protein (2.5-fold at 24 hours). This effect was inhibited by either PTHrP(7-34) or various protein kinase A inhibitors and by the nuclear factor-kappaB (NF-kappaB) inhibitor parthenolide. Furthermore, PTHrP(1-36) elicited an increase in NF-kappaB activation in VSMC. The 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitor simvastatin inhibited the PTHrP(1-36) induction of both NF-kappaB activity and MCP-1 overexpression, and this was reversed by mevalonate. CONCLUSIONS: PTHrP appears to be a novel proinflammatory mediator in the atheroma lesion and may contribute to the instability of carotid atherosclerotic plaques. Our data provide a new rationale to understand the mechanisms involved in the beneficial effects of statins in atherosclerosis.

Simvastatin reduces NF-kappaB activity in peripheral mononuclear and in plaque cells of rabbit atheroma more markedly than lipid lowering diet.

OBJECTIVE: To study whether simvastatin reduces inflammation in atherosclerosis beyond its hypolipidemic effects. METHODS: Twenty-four rabbits with induced femoral injury and on an atherogenic diet were randomized to normolipidemic diet (n=9), or to continue the atherogenic diet while receiving simvastatin 5 mg/kg/day (n=9) or no treatment (n=6) for 4 weeks. RESULTS: As compared with no treatment, the normolipidemic diet significantly reduced lipid levels, while simvastatin produced nonsignificant reductions. In spite of this, NF-kappaB binding activity in peripheral mononuclear cells was reduced in the simvastatin group [2,958+/-5,123 arbitrary units (a.u.)] as compared with no treatment (49,267+/-20,084 a.u.; P<0.05) and normolipidemic groups (41,492+/-15,876 a.u.; P<0.05) (electrophoretic mobility shift assay). NF-kappaB activity in the atherosclerotic lesions was also reduced by simvastatin as compared to nontreated animals (4,108+/-3,264 vs. 8,696+/-2,305 nuclei/mm(2); P<0.05), while the normolipidemic diet induced only a nonsignificant diminution (P>0.05) (Southwestern histochemistry). Similarly, simvastatin decreased macrophage infiltration (4.6+/-12 vs. 19+/-12% of area staining positive; P<0.05) and the expression of interleukin-8 (24+/-12 vs. 63+/-21%; P<0.05) and metalloproteinase-3 (16+/-3 vs. 42+/-28%; P<0.05) (immunohistochemistry), while the reduction achieved by normolipidemic diet in all these parameters was again nonsignificant (P>0.05). CONCLUSIONS: These findings suggest that simvastatin reduces inflammation in atherosclerotic plaques and in blood mononuclear cells more than expected for the lipid reduction achieved.

3-Hydroxy-3-methylglutaryl coenzyme A reductase inhibitors increase the binding activity and nuclear level of Oct-1 in mononuclear cells.

3-Hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) are drugs very effective to decrease low-density lipoprotein (LDL) cholesterol. In addition, a number of studies suggest that statins have other beneficial clinical effects beyond cholesterol lowering. We recently reported that statins decrease nuclear factor kappa B (NF-kappaB) binding activity in monocytes and vascular smooth muscle cells. We now explored the effect of two different statins, simvastatin and atorvastatin, in the activation of the octamer transcription factor Oct-1 on the monocytic cell line THP-1. Oct-1 is a nuclear factor that represses the transcription of proinflammatory genes such as interleukin-8, CD11c/CD18, vascular cell adhesion molecule-1 (VCAM-1) and platelet endothelial cell adhesion molecule-1 (PECAM-1). Low concentrations of both statins increased Oct-1 DNA binding activity (electrophoretic mobility shift assay) that was resolved into two specific bands. The upper one was supershifted by preincubation of nuclear extracts with anti-Oct-1 antibody. The lower one was supershifted by preincubation of nuclear extracts with an anti-Oct-2 antibody, also partially competed with 100 mol/l excess of cold activator protein-1 (AP-1) and attenuated by anti-c-Jun antibody. Both statins increased Oct-1 and Oct-2 nuclear protein levels (Western blot). In contrast, neither had any effect on PMA-differentiated cells, suggesting a distinct sensitivity between circulating monocytes and resident tissular macrophages. In addition, statins did not increase Oct-lipoprotein lipase binding activity that contains an Oct-1 binding element. The mRNA expression of interleukin-8, a chemokine containing Oct sites in its promoter, was diminished by statin pretreatment. Our results indicate that simvastatin and atorvastatin increase the activity of the transcriptional repressor Oct-1 in mononuclear cells, and could thus contribute to decrease the activation of these cells. These data suggest a possible novel mechanism supporting a certain anti-inflammatory effect of these two 3-hydroxy-3-methylglutaryl-CoA reductase inhibitors.

3-Hydroxy-3-methyl-glutaryl coenzyme A reductase inhibitors, atorvastatin and simvastatin, induce apoptosis of vascular smooth muscle cells by downregulation of Bcl-2 expression and Rho A prenylation.

The mechanism by which 3-hydroxy-3-methyl-glutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) induce apoptosis in vascular smooth muscle cells (VSMCs) is unknown. In this work, we demonstrate that treatment of VSMCs with simvastatin and atorvastatin inhibited Bcl-2 expression in a time and dose-dependent manner, while Bax expression was not modified. This effect was reversed by mevalonate (100 micromol/l), farnesylpyrophosphate (5 micromol/l) or geranylgeranylpyrophosphate (5 micromol/l), suggesting the involvement of protein prenylation. The treatment of VSMCs with lipophilic statins was associated with decreased prenylation of p-21 Rho A and mevalonate, farnesyl pyrophosphate (F-PP) and geranylgeranyl pyrophosphate (G-PP) reversed prenylation to basal levels. In addition, overexpression of constitutively active Q63L Rho A prevented, at least in part, apoptosis induced by statins and downregulation of Bcl-2. We also investigated the participation of caspases (proteases) in the apoptosis induced by statins. The treatment of VSMCs with lipophilic statins induced activation of the caspase 9, the first caspase of the mitochondrial pathway. Coincubation of VSMCs with the caspase inhibitor ZVAD-fmk (100 micromol/l) significantly inhibited lipophilic statin-induced apoptosis. These findings indicate that the downregulation of Bcl-2 by Rho GTPases mediates statin-induced apoptosis and suggest a new potential mechanism of action for these drugs on the regulation of cell number in the atherosclerotic lesions.

Atorvastatin reduces the expression of cyclooxygenase-2 in a rabbit model of atherosclerosis and in cultured vascular smooth muscle cells.

Inflammation is involved in the genesis and rupture of atherosclerotic plaques. We assessed the effect of atorvastatin (ATV) on the expression of cyclooxygenase-2 (COX-2) and other proinflammatory molecules in a rabbit model of atherosclerosis. Fourteen animals underwent injury of femoral arteries and 2 weeks of atherogenic diet. Afterwards, they were randomized to receive either 5 mg/kg per day of ATV (n=8) or no treatment (NT, n=6) during 4 weeks, and were finally killed. ATV reduced lipid levels, neointimal size (0.13 (0.03-0.29) mm(2) vs 0.65 (0.14-1.81) mm(2), P=0.005) and the percentage of neointimal area positive for macrophages (1% (0-3) vs 19% (5-32), P=0.001), COX-2 (32% (23-39) vs 60% (37-81) P=0.019), interleukin-8 (IL-8) (23% (3-63) vs 63% (25-88) P=0.015), and metalloproteinase-3 (19% (12-34) vs 42% (27-93), P=0.010), without significant differences in COX-1 expression (immunohistochemistry). In situ hybridization confirmed a decreased expression of COX-2 mRNA (22% (5-40) vs 43% (34-59) P=0.038). The activity of nuclear factor-kappaB, which controls many proinflammatory genes including COX-2, was reduced in atherosclerotic lesions (3538 (2663-5094) vs 8696 (5429-11312)) positive nuclei per mm(2), P=0.001) and circulating mononuclear cells (2966 vs 17130 arbitrary units). In cultured vascular smooth muscle cells, ATV reduced the expression of COX-2 mRNA induced by IL-1beta and TNF-alpha without affecting COX-1 expression. In conclusion, ATV, besides decreasing a number of inflammatory mediators in the atherosclerotic lesion, significantly downregulates COX-2 both in vivo and in vitro. These anti-inflammatory actions could partially account for the reduction of acute coronary events achieved by statins.