Full Length Article
Updated clinical models for VTE prediction in hospitalized medical patients

https://doi.org/10.1016/j.thromres.2018.02.004Get rights and content

Highlights

  • RAM for VTE is important forthe improvement of VTE prevention in hospitalized acutely ill medical patients

  • The Padua Prediction Score, IMPROVE RAM and Geneva Risk Score accurately classify patients at high and low risk for VTE

  • The Bleeding Risk Score together with the IMPROVE RAM identifies patients at high risk of bleeding

Abstract

Venous thromboembolism (VTE) occurring in hospitalized medical patients is associated with increased length of hospitalization, high rate of acute care hospital transfer, longer inpatient rehabilitation and multiplication of health-care costs. Identification of acutely ill hospitalized medical patients eligible for thromboprophylaxis is a sophisticated process. Global VTE risk stems from the combination of predictors related with the acute medical illness, comorbidities, associated treatments and patients' intrinsic risk factors. Emerging clinical risk factors related to underlying pathologies should be considered when VTE risk is assessed. The Padua Prediction Score (PPS), the International Medical Prevention Registry on Venous Thromboembolism (IMPROVE-RAM) and the Geneva Risk Score are three robust risk assessment models (RAM) which underwent extensive external validation in cohorts of acutely ill hospitalized medical patients. The development of the IMPROVE bleeding risk assessment model and the identification of D-Dimer increase as a biomarker-predictor of VTE are some steps forward for personalized thromboprophylaxis. The beneficial impact of the RAMs in VTE prevention is already seen by the decrease of in-hospital VTE rates when RAMs are incorporated in electronic alert systems.

Introduction

VTE is a multifactorial disease. Patients may present a combination of multiple risk factors. Hospitalization figures among the major risk factors for VTE [1]. Studies performed in the '90s - before the introduction of pharmacological thromboprophylaxis in hospitalized non surgical patients - showed that the incidence of VTE is in the range of 10–30%, while up to 10% of deaths in hospitalized patients are related to pulmonary embolism (PE). Most importantly, 75% of these deaths occur in non-surgical hospitalized patients [2,3]. In addition, VTE occurring in hospitalized medical patients is associated with increased length of stay independently of the cause of admission, high rate of intensive care hospital transfer, longer inpatient rehabilitation and multiplication of health-care costs [4].

The large phase III clinical trials MEDENOX, PREVENT and ARTEMIS established that acutely ill medical patients are at increased risk of VTE. Pharmacological thromboprophylaxis during hospitalization using the low molecular weight heparins (LMWH) enoxaparin 40 mg, dalteparin 5000 IU or fondaparinux 2.5 mg s.c. o.d. has a favorable risk-benefit ratio in the prevention of VTE [[5], [6], [7]]. Moreover, the EXCLAIM trial showed that a subpopulation of patients hospitalized for acute medical illness continues to be at risk for VTE after hospital discharge and may benefit from prolonged prophylaxis with enoxaparin [8]. Identification of acutely ill hospitalized medical patients eligible for thromboprophylaxis is a sophisticated process since the overall VTE risk results from the association of risk factors related to the acute disease itself as well as to patients' intrinsic risk factors and comorbidities. The three major phase III clinical trials for VTE prevention in acutely ill hospitalized medical patients firmly established the beneficial effect of pharmacological thromboprophylaxis [[5], [6], [7]]. However, the inclusion criteria which defined patients as high risk and therefore eligible for thromboprophylaxis were different in the three studies (Table 1). This discrepancy revealed a major limitation in routine application of pharmacological thromboprophylaxis as the target population that could benefit from this intervention is not well defined.

Expert consensus statements and national guidelines strongly recommend the administration of pharmacological thromboprophylaxis during hospitalization for acute medical illness [9,10]. Nevertheless, awareness for VTE as well as guidelines adherence are suboptimal. Less than 40% of hospitalized medical patients receive the recommended thromboprophylaxis and this figure is significantly lower compared to the corresponding in the surgical milieu [[11], [12], [13]]. Although the American College of Chest Physicians (ACCP) introduced a new approach to patient selection, adherence to VTE prophylaxis guidelines did not improve [14].

Underuse of pharmacological thromboprophylaxis in hospitalized acutely ill medical patients results from (i) under-estimation of VTE risk, (ii) low perception that patients included in the trials are actually at risk of thromboembolism, (iii) confusion regarding the clinical relevance of VTE risk factors, (iv) high heterogeneity of hospitalized acutely ill medical patients (v) need for a simple-to-use and validated risk assessment tool, (vi) fear of treatment related bleeding risk, (vii) parenteral (subcutaneous) route of administration of pharmacological thromboprophylaxis, (viii) absence of an integral, multidisciplinary, institution-based strategy targeting the integration of VTE risk assessment and thromboprophylaxis into the existing process for healthcare delivery [15,16].

Several tools for predicting thrombotic risk are developed. These risk assessment models (RAMs) incorporate clinical factors as well as biomarkers. The improvement of RAMs is a subject of ongoing research.

An essential step for stratification of medical patients risk groups is the identification of the intrinsic or predisposing risk factors and exposing or triggering risk factors [17,18]. Insights into VTE risk factors in hospitalized patients with acute medical illness can be obtained from large epidemiological studies and from the randomized controlled trials comparing groups of patients receiving placebo (or no-prophylaxis) versus those who are on pharmacological thromboprophylaxis.

Restricted mobility and immobilization as well as recent previous hospitalization are important risk factors for VTE in medical patients [19,20]. Hospitalization in medical intensive care unit further increases the risk of VTE [[21], [22], [23], [24]]. Established risk factors for VTE are part of the inclusion criteria in the three major clinical trials assessing the efficacy and safety of thromboprophylaxis with LMWH or fondaparinux in acutely ill hospitalized medical patients (Table 1).

However some additional underlying pathologic conditions are also related with the risk of VTE and should be evaluated when risk assessment is performed on individual basis.

Patients with aortic aneurysms are associated with a 1.88-fold higher risk of DVT and 1.90-fold higher risk of PE compared to those without aneurysm [25]. The risk of VTE is increased across the spectrum of chronic kidney disease (CKD), from mild to more advanced CKD, and typically characterizes the nephrotic syndrome [26]. In patients with nephrotic syndrome, low levels of serum albumin appear to be a predictor of VTE risk. The relative risk of VTE in relation to albumin levels increases as follows: albumin levels of 3–3.99 g/dL: OR = 1.51; [95% CI: 1.01–2.26]; albumin 2.5–2.99 g/dL: OR = 2.24, [95% CI: 1.24–4.05] and albumin <2.5 g/dL: OR = 2.79; [95% CI: 1.45–5.37] [27]. Sickle cell disease as well as other hemoglobinopathies are independent risk factors for VTE [28,29]. It should be noted that sickle cell disease carriers are at increased risk of VTE (OR 1.78, [95% CI: 1.18 to 2.69], p = 0.006), with the greatest risk for PE (OR 2.27, [95% CI: 1.17 – 4.39], p = 0.011) [30].

Acute medical conditions such as stroke, congestive heart failure, respiratory disease, infections, or myocardial infarction are well established risk factors for VTE and are present in the inclusion criteria list of the major phase III trials in this field (Table 1) [31,32]. VTE and atherothrombosis share common pathophysiology, which includes inflammation, hypercoagulability, and endothelial injury [33]. Atherothrombosis risk factors such as smoking, hypertension, diabetes, and obesity, overlap with VTE [34,35]. In patients weighing >100 kg or having a BMI of 25 kg/m2 or greater, the incidence of VTE increases incrementally with increasing obesity [36]. In the Atherosclerosis Risk In Communities (ARIC) Study, C-reactive protein levels above the 90th percentile were associated with a marked increase in VTE risk compared with lower percentiles [37]. The risks of unprovoked VTE and PE severity are associated with clustering of cardiovascular risk factors (CVRF). The evolution of VTE risk according to the number of cumulative CVRF is as follows: 1 CVRF: OR = 3 (95% CI: 1.44–6.52), 2 CVRF = OR 4.33 (95% CI: 2.07–9.49) and ≥3 CVRF: OR = 4.58 (95% CI: 2.27–9.7) [38]. Finally, obstructive sleep apnea figures among the important risk factors for both VTE and arterial thrombosis [39].

Chronic autoimmune diseases are independent risk factors for VTE [40]. The relative risk of VTE in patients with Sjögren's syndrome is 2.05 (95% CI: 1.86–2.27) [41,42]. Patients with psoriasis are in a 2.02-fold higher VTE risk compared to age and sex-matched healthy individuals (95% CI: 1.42–2.88). The risk of VTE is significantly higher in the first year of follow-up after diagnosis (OR = 3.30, [95% CI: 1.45–7.55]) than after one year (OR = 1.68, 95% CI: 1.13–2.49) [43,44].

Patients with vasculitis (polyarteritis nodosa, granulomatosis with polyangiitis and giant cell arteritis) are at increased risk of VTE (pooled OR = 3.94; [95% CI: 1.11–14.01]) [45]. Giant cell arteritis (GCA) is also an independent risk factor for VTE. The VTE risk increases shortly before GCA diagnosis, peaks at the time of diagnosis, and progressively declines thereafter [46].

Various forms of lupus erythematous (LE) ranging from cutaneous localized disease (CLE) to systemic involvement (SLE), are independently associated with the risk of VTE. A recent retrospective cohort study including 3234 patients with CLE and 3627 patients with SLE compared to 5,590,070 individuals in the reference population showed that the incidence rates of VTE were higher in patients with LE, (OR: 1.20, 3.06, and 5.24 for the reference population, CLE, and SLE, respectively). The relative VTE risk compared to the reference population was 1.39 (95% CI: 1.10–1.78) in patients with CLE and 3.32 (95% CI: 2.73–4.03) in patients with SLE [47].

Inflammatory bowel disease - Crohn's disease and ulcerative colitis – is a well-recognized risk factor of VTE. Recent evidence shows that patients with inflammatory bowel disease who have a Clostridium Difficile infection are at higher risk of VTE compared to those who are negative for CDI (OR = 1.7; [95% CI = 1.4–2.2], p < 0.001) [48].

Patients with celiac disease (CD) are at an increased risk of VTE (OR = 1.25; [95% CI: 1.02–1.53]) due to chronic inflammation and vitamin deficiency.

Patients with chronic immune thrombocytopenic purpura (ITP) are also at increased VTE risk as compared to background population (OR = 2.95; [95% CI: 2.18–4.00). The risk of VTE is higher within one year after ITP diagnosis [49]. Noteworthy, VTE may occur in patients with chronic ITP even in the presence of low platelet count and in the absence of detectable antiphospholipid antibodies whereas this risk increases after splenectomy [50].

Idiopathic inflammatory myopathy - polymyositis (PM) and/or dermatomyositis (DM) - is an independent risk factor of VTE. A meta-analysis of five studies, including 8858 patients with PM/DM revealed a significant association between PM/DM and VTE risk (OR = 4.364, [95% CI: 2.128–8.949], p < 0.005) [51,52].

It is well established that chronic HIV infection is associated with increased VTE risk [53]. Chronic HIV infection in patients with tuberculosis amplifies the risk of VTE (OR = 8.2; [95% CI: 2.9–22.7]) [54]. In addition, other chronic viral infections are also associated with increased VTE risk; the pooled relative risk of VTE in hepatitis C virus (HCV)-infected patients versus subjects without HCV infection is 1.38 ([95% CI: 1.08–1.77], I2 = 40%) [55].

The risk of VTE substantially increases with greater comorbidity in patients hospitalized for acute infection, i.e. acute pancreatitis (OR = 1.47), acute kidney injury (OR = 1.08), acute respiratory failure (OR = 1.40), pseudo cyst (OR = 1.41), total parenteral nutrition (OR = 1.28) and central venous catheter placement (OR = 3.01) [56,57].

Oral contraception with estrogen/progestin agents and estrogen based hormone replacement treatment in women is an established risk factor for VTE (reviewed in [58]). In some countries, testosterone administration is an emerging treatment in men. This treatment is a new risk factor for VTE which should not be neglected [59].

Several treatments used in oncology are independent risk factors for VTE. Among them, chemotherapeutic agents, immunomodulatory drugs (such as thalidomide and lenalidomide) as well as erythropoiesis-stimulating agents figure among the most important independent risk factors for VTE [[58], [59], [60], [61], [62]].

Antipsychotic drugs used in ambulatory patients, represent an important risk factor for non-hospital related VTE risk. Analysis of the data from a continuous multicenter drug surveillance program in 264,422 psychiatric inpatients monitored from 1993 to 2011 in 99 psychiatric hospitals showed that antipsychotic drug exposure should be considered as a risk factor for VTE for patients older than 65 years. Additionally, the diagnosis of an affective disorder seems to increase the risk for VTE. Interestingly, VTE risk varies according to the class of antipsychotic agents. Butyrophenones (48/100,000) followed by atypical antipsychotics (36/100,000) showed the highest occurrence rate for VTE compared to thioxanthenes (23/100,000), which were less associated with VTE. If imputed alone, pipamperone (61/100,000) and risperidone (55/100,000) were most frequently associated with VTE. In general, there was no difference in occurrence rate of VTE between high- and low-potency antipsychotics [63]. The risk of VTE is higher in men over 65 years of age with mood disorders as well as in postmenopausal women who are on antipsychotic treatment (OR = 1.90-fold; 95% CI = 1.64–2.19) [64].

The relative risk of VTE related with the most recently described risk factors is summarized in Table 2.

Several RAMs have been derived from cohorts of acutely ill hospitalized medical patients and examined in independent validation cohorts [reviewed in [65], [66], [67]]. Among them, the Kucher Model, the Padua Prediction Score (PPS), the International Medical Prevention Registry on Venous Thromboembolism (IMPROVE-RAM), and the Geneva Risk Score underwent extensive external validation. The Kucher Model is an electronic alert tool with 8 differently weighted predictors. It functions via the link between the computer program and the database of hospitalized patients and classifies patients in high, intermediate and low risk for VTE [68].

The PPS uses 11 predictors scored with 1, 2 or 3 points and stratifies patients in a binary fashion: high risk for VTE are designated with a score ≥4 warranting pharmacologic prophylaxis, whereas patients at low risk for VTE are designated with a score <4.

The IMPROVE RAM derived from a registry of 15,125 medically-ill patients and includes 11 predictors scored with 1, 2 or 3 points. Patients at high risk for VTE are designated with a score ≥4 [69,70]. These patients had an observed VTE risk of 5.7% versus <1% in the low-risk group.

The three RAMs (Kucher, PPS and IMPROVE) share a common structure. They share 6 common VTE predictors which however have a different weight in each RAM. The PPS includes more risk factors related to comorbidities, underlying pathologic conditions and co-medications (in regard to hormone treatment), whereas IMPROVE RAM emphasizes on factors related to immobilization and disease severity.

The Geneva VTE RAM follows a different structure since it is based on the assessment of 19 predictors which represent the classical VTE risk factors. The weight of each predictor is based on a dichotomous analysis (1 or 2).

Table 3 summarizes and compares the predictors of VTE included in each RAM.

Section snippets

Performance of the RAMs for VTE in hospitalized acutely ill medical patients

The implementation of the electronic alert system proposed by Kucher increased the rate of application of mechanical prophylaxis from 1.5% to 10.0% (p < 0.001) and that of pharmacological thromboprophylaxis from 13.0% to 23.6% (p < 0.001) reducing the hospitalization-related VTE risk at 90 days by 41% (hazard ratio: 0.59; [95% CI: 0.43–0.81]; p < 0.001) [68].

The development of the Automated Padua Prediction Score (APPS) to auto-calculate a VTE risk score using electronic health record data had

RAM and e-@lert systems

The introduction of an electronic alert system for VTE scoring into hospital's t network is associated with improved management of VTE risk [66]. Several newer studies confirmed the beneficial effect of electronic alert systems on the implementation of VTE prevention strategies. This effect is mirrored both on the optimization of antithrombotic agents' prescription and on the decrease of in-hospital VTE [[84], [85], [86]]. However, the over-exposure of physicians to electronic alert procedures

Conclusions

Improvement of VTE prevention in hospitalized acutely ill medical patients stems from a global strategy that includes (a) implementation of preventive protocols and hospital guidelines (b) availability of a functional risk assessment tool and education of the medical staff on its use (c) continuous education of physicians and nurses on personalized management of VTE (d) computerized systems for data retrieval and continuous auditing and feedback to the medical staff and (e) multidisciplinary

Author contributions

All authors contributed to drafting of the paper.

Conflicts of interest

Grigoris Gerotziafas and Ismail Elalamy have served as consultants or on advisory committees for Sanofi, Leo-Pharma, Boehringer Ingelheim, Rovi, Bayer Healthcare.

References (86)

  • A. Falanga et al.

    Anticancer treatment and thrombosis

    Thromb. Res.

    (2012)
  • S. Barbar et al.

    A risk assessment model for the identification of hospitalized medical patients at risk for venous thromboembolism: the Padua Prediction Score

    J. Thromb. Haemost.

    (2010)
  • A.C. Spyropoulos et al.

    Predictive and associative models to identify hospitalized medical patients at risk for VTE

    Chest

    (2011)
  • H. Decousus et al.

    Factors at admission associated with bleeding risk in medical patients: findings from the IMPROVE investigators

    Chest

    (2011)
  • D. Spirk et al.

    Electronic alert system for improving appropriate thromboprophylaxis in hospitalized medical patients: a randomized controlled trial

    J. Thromb. Haemost.

    (2017)
  • A.T. Cohen et al.

    D-dimer as a predictor of venous thromboembolism in acutely ill, hospitalized patients: a subanalysis of the randomized controlled MAGELLAN trial

    J. Thromb. Haemost.

    (2014)
  • M.M. Samama

    An epidemiologic study of risk factors for deep vein thrombosis in medical outpatients: the Sirius study

    Arch. Intern. Med.

    (2000)
  • S.M. Pinto et al.

    Venous thromboembolism as predictor of acute care hospital transfer and inpatient rehabilitation length of stay

    Am. J. Phys. Med. Rehabil.

    (2017)
  • M.M. Samama et al.

    A comparison of enoxaparin with placebo for the prevention of venous thromboembolism in acutely ill medical patients. Prophylaxis in Medical Patients with Enoxaparin Study Group

    N. Engl. J. Med.

    (1999)
  • A. Leizorovicz et al.

    PREVENT Medical Thromboprophylaxis Study Group. Randomized, placebo-controlled trial of dalteparin for the prevention of venous thromboembolism in acutely ill medical patients

    Circulation

    (2004)
  • A.T. Cohen et al.

    Efficacy and safety of fondaparinux for the prevention of venous thromboembolism in older acute medical patients: randomised placebo controlled trial

    BMJ

    (2006)
  • R.D. Hull et al.

    EXCLAIM (Extended Prophylaxis for Venous ThromboEmbolism in Acutely Ill Medical Patients With Prolonged Immobilization) study. Extended-duration venous thromboembolism prophylaxis in acutely ill medical patients with recently reduced mobility: a randomized trial

    Ann. Intern. Med.

    (2010)
  • A.N. Nicolaides et al.

    Prevention and treatment of venous thromboembolism—International Consensus Statement

    Int. Angiol.

    (2013)
  • J.F. Bergmann et al.

    Venous thromboembolism risk and prophylaxis in hospitalised medically ill patients. The ENDORSE Global Survey

    Thromb. Haemost.

    (2010)
  • A. Amin et al.

    Are hospitals delivering appropriate VTE prevention? The venous thromboembolism study to assess the rate of thromboprophylaxis (VTE start)

    J. Thromb. Thrombolysis

    (2010)
  • F.M. Al-Hameed et al.

    Thromboprophylaxis and mortality among patients who developed venous thromboembolism in seven major hospitals in Saudi Arabia

    Ann. Thorac. Med.

    (2017)
  • F. Violi et al.

    Should all acutely ill medical patients be treated with antithrombotic drugs? A review of the interventional trials

    Thromb. Haemost.

    (2013)
  • A.M. Schleyer et al.

    Preventing hospital-acquired venous thromboembolism: improving patient safety with interdisciplinary teamwork, quality improvement analytics, and data transparency

    J. Hosp. Med.

    (2016)
  • M.M. Samama et al.

    Quantification of risk factors for venous thromboembolism: a preliminary study for the development of a risk assessment tool

    Haematologica

    (2003)
  • J.A. Heit et al.

    Relative impact of risk factors for deep vein thrombosis and pulmonary embolism

    Arch. Intern. Med.

    (2002)
  • K. Kröger et al.

    Venous thromboembolism in Germany: results of the GermAn VTE registry (GATE-registry)

    Int. J. Clin. Pract.

    (2014)
  • A.A. Khalafallah et al.

    Venous thromboembolism in medical patients during hospitalisation and 3 months after hospitalisation: a prospective observational study

    BMJ Open

    (2016)
  • C. Minet et al.

    Venous thromboembolism in the ICU: main characteristics, diagnosis and thromboprophylaxis

    Crit. Care

    (2015)
  • G.V. Fontaine et al.

    Venous thromboembolism in critically ill medical patients receiving chemoprophylaxis: a focus on obesity and other risk factors

    Clin. Appl. Thromb. Hemost.

    (2016)
  • M. Boddi et al.

    Thrombosis in intensive care

    Adv. Exp. Med. Biol.

    (2017)
  • F.Y. Lee et al.

    Increased risk of deep vein thrombosis and pulmonary thromboembolism in patients with aortic aneurysms: a nationwide cohort study

    PLoS One

    (2017)
  • G. Gyamlani et al.

    Association of serum albumin level and venous thromboembolic events in a large cohort of patients with nephrotic syndrome

    Nephrol. Dial. Transplant.

    (2017)
  • T. Wun et al.

    Sickle cell disease: an inherited thrombophilia

    Hematology Am. Soc. Hematol. Educ. Program

    (2016)
  • D. Kelley et al.

    Evaluating the safety and effectiveness of venous thromboembolism prophylaxis in patients with sickle cell disease

    J. Thromb. Thrombolysis

    (2017)
  • I. Little et al.

    Venous thromboembolism in adults screened for sickle cell trait: a population-based cohort study with nested case-control analysis

    BMJ Open

    (2017)
  • S.Z. Goldhaber et al.

    Prevention of venous thromboembolism among hospitalized medical patients

    Circulation

    (2005)
  • R. Alikhan et al.

    Risk factors for venous thromboembolism in hospitalized patients with acute medical illness: analysis of the MEDENOX Study

    Arch. Intern. Med.

    (2004)
  • P. Prandoni et al.

    An association between atherosclerosis and venous thrombosis

    N. Engl. J. Med.

    (2003)
  • Cited by (22)

    • Thrombosis and cachexia in cancer: Two partners in crime?

      2023, Critical Reviews in Oncology/Hematology
    • Clinical practice in COVID-19: The most frequently asked questions to infectious diseases specialists

      2021, Brazilian Journal of Infectious Diseases
      Citation Excerpt :

      Hospitalized acutely ill patients, including those with other infections such as pneumonia, have an increased risk of venous thromboembolism (VTE).37 Patient-specific VTE risk factors such as advanced age, a prior history of VTE, a history of or active cancer, immobility, and thrombophilia, had been incorporated before the COVID-19 era to assess overall VTE risk using standardized VTE risk assessment scores such as Padua VTE or IMPROVE VTE risk scores.38 The overall estimated VTE prevalence in COVID-19 was 14.1% in non-ICU patients, 22,7% in ICU patients, reaching up to 40.3% with ultrasound screening.3940

    • AGA Technical Review on Coagulation in Cirrhosis

      2021, Gastroenterology
      Citation Excerpt :

      Current guidelines for medical patients admitted to the hospital recommend VTE prophylaxis with anticoagulation for patients at high risk to develop VTE.80 Risk stratification should be performed with risk assessment models, such as the Padua Prediction Score, to determine which patients are at highest risk and would benefit from prophylaxis.87,88 Risk assessment models have been applied successfully to cohorts with cirrhosis in smaller retrospective observational studies.18,89

    • Standardized Practices to Address Maternal Venous Thromboembolism: AWHONN Practice Brief Number 7

      2020, Nursing for Women's Health
      Citation Excerpt :

      However, the development of easily implementable treatment guidelines to address maternal VTE is challenging because the well-being of two patients must be considered: the woman and the fetus (Bates et al., 2018). Although strategies to prevent VTE have been well studied in the general population (Gerotziafas, Papageorgiou, Nikolopoulou, & Elalamy, 2018; Schünemann et al., 2018; Zhou et al., 2018), clear, evidence-based guidelines for VTE prophylaxis and treatment during and after pregnancy are limited by the lack of randomized controlled trials in pregnant women (Eubanks, Deering, & Thiel, 2019). The published recommendations of many professional nursing and medical organizations related to VTE overlap, conflict, and/or are nonspecific to women in the perinatal period (American College of Obstetricians and Gynecologists, 2018; Bates et al., 2012; Bates et al., 2018; Leffert et al., 2018).

    View all citing articles on Scopus
    View full text