Description of the condition
Venous thromboembolism (VTE) is the collective term for the clinical conditions deep vein thrombosis (DVT) and pulmonary embolism (PE). DVT is the formation of a blood clot (thrombus) in a deep vein, predominantly in the legs. Symptoms include pain, tenderness, erythema and swelling of the affected leg. PE occurs when part or all of the thrombus breaks off (embolises) and travels up to the lungs, blocking the pulmonary arteries. Symptoms of PE include breathlessness and chest pain (Blann 2006).
Guidelines published by the UK National Institute for Health and Clinical Excellence (NICE) recommend that patients with a suspected DVT should be risk stratified using the Wells score, undergo a D-dimer test and have a proximal leg vein ultrasound scan based on the risk of DVT (NICE 2012a). Similarly, patients with a suspected PE should be assessed using the Wells PE risk score, D-dimer and a computed tomography pulmonary angiogram (CTPA) depending on the PE risk. The NICE guidelines recommend that anticoagulant therapy with low molecular weight heparin (LMWH) should be administered in the interim until these diagnostic investigations are carried out if leg ultrasound is expected to take longer than four hours, or if a CTPA cannot be performed immediately (NICE 2012a). Patients with confirmed VTE should receive LMWH or fondaparinux for at least the initial five days and, except for cancer patients, all others also started on a vitamin K antagonist. The LMWH should be stopped when the international normalised ratio (INR) has been above 2 for at least 24 hours. Vitamin K antagonists should be continued for at least three months in all patients. In patients with an unprovoked VTE consideration should be given to extending anticoagulation beyond three months based on an assessment of the risk of VTE recurrence and complications of anticoagulation. However, patients with cancer-associated VTE should be treated with LMWH from initial diagnosis for six months, and considered for continuation of anticoagulation with either LMWH or a vitamin K antagonist based on the status of the underlying cancer and risks of anticoagulation (NICE 2012a). More recently, rivaroxaban has been used for the treatment of DVT and prevention of recurrent DVT and PE (NICE 2012b). Guidelines by NICE recommend that rivaroxaban should be administered at a dose of 15 mg twice daily for the first 21 days followed by a continual dose of 20 mg once daily to prevent recurrence. Evidence from the EINSTEIN-DVT study showed that, in a subgroup of patients with active cancer, rivaroxaban was less effective than LMWH at preventing VTE recurrence (hazard ratio 1.32, 95% confidence interval (CI) 0.06 to 32.3) but was associated with fewer major bleeding events (Einstein Investigators).
The difference in management of patients with a cancer-associated VTE is due to their significantly higher risk of VTE recurrence, which is estimated to be three time higher than in patients with no cancer-associated VTE (Levitan 1999). The underlying pathophysiology of cancer and thrombosis involves a complex process of several mechanisms including a "release of inflammatory cytokines, activation of the clotting system, expression of haemostatic proteins on tumour cells, inhibition of natural anticoagulants and impaired fibrinolysis" as described by Rodrigues 2010. Activation of the clotting system has also been implicated in cancer-associated angiogenesis, tumour metastasis and aggressiveness. Indeed patients with cancer and an associated VTE have a poorer overall prognosis compared to patients without a VTE, with a 12% one-year survival from the diagnosis of VTE (Kakkar 1969; Ruiz 2003).
A proportion of patients with VTE have no underlying or immediately apparent cause and the VTE is referred to as idiopathic or unprovoked. Unprovoked VTE can often be the first clinical manifestation of underlying malignancies such as cancer of the blood, kidney, ovary, pancreas, stomach and lung (Bick 1978; Kakkar 2003; Lee 2003a; Prandoni 1997; White 2005). Results from a Swedish prospective cohort study of almost 62,000 patients determined that the standardised incidence ratio of a cancer diagnosis within the first two years of an unprovoked VTE was 4.4 (Baron 1998), and there was an overall absolute incidence of cancer of 11% (NICE 2012a). A UK study of 339 patients with a first episode of an unprovoked VTE measured that the relative risk (RR) of cancer-related mortality at two years was 0.52 (95% CI 0.10 to 2.75) while the RR for early-stage cancer detection was 3.21 (95% CI 0.88 to 11.79) (Piccioli 2004).
Patients who present with an apparent unprovoked VTE, therefore, have a significant underlying risk of malignancy or cancer-associated VTE, with significant implications for the management of the VTE itself (three months vitamin K antagonist versus six months LMWH), the prognosis related to risk of VTE recurrence and the precipitating cancer. This has raised the question of whether patients with an unprovoked VTE should be investigated for an underlying cancer. Some authors refer to this as 'screening for cancer' although this is somewhat misleading, as screening refers to the investigation of asymptomatic patients. Instead, patients with VTE are better regarded as presenting with symptoms suggestive of an underlying cancer and the aim of investigations is to better refine the diagnosis of VTE based on the underlying cause, so that the patient may receive a more accurate diagnosis and appropriate treatment for their VTE. In this context, VTE represents a symptom of a disease rather than a diagnosis per se. So to what extent should patients with an unprovoked VTE be investigated for a potential underlying cancer?
Description of the intervention
The NICE guidelines on venous thromboembolic disease (NICE 2012a) recommend that all patients diagnosed with a first episode of unprovoked VTE (DVT or PE) should undergo a history and physical examination directed to detecting an underlying malignancy, and further tests guided by the history and examination including blood tests (complete blood count, serum calcium and liver function tests), urinalysis and chest X-ray. If none of these initial investigations confirm signs and symptoms of cancer then further tests including an abdomino-pelvic computerised tomography (CT) scan (and a mammogram for women) are recommended in patients over 40 years (NICE 2012a).
Complete blood count is a test that measures the number of red cells, white cells and platelets in the blood. This test can diagnose certain conditions including polycythaemia, anaemia, leukocytosis, leukopenia, thrombocytosis and thrombocytopenia, all indicators for various cancers (Tefferi 2005).
Serum calcium tests indicate hypercalcaemia, an elevated calcium level in the blood greater than 2.6 mmol/L. Hypercalcaemia affects between 10% to 40% of cancer patients at some time during their disease (Mundy 1997). The type of malignancies associated with hypercalcaemia include solid tumours such as breast and squamous cell lung cancer and haematologic malignancies such as multiple myelomas (Muggia 1990).
The liver is a common site of metastases. Liver function tests are a group of tests used to determine evidence of liver damage. Liver function tests look for levels of enzymes, proteins and compounds made by the liver including alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), gamma-glutamyl transferase (gamma GT), ammonia, bilirubin, albumin, prothrombin and fibrinogen (Limdi 2003).
Urinalysis is a simple test used to check urine for blood, protein and other substances. Blood in the urine (haematuria) is found in 80% to 90% of patients with bladder cancer (Fritsche) while high levels of protein in the urine (proteinuria) is found in 50% to 60% of patients with renal cell carcinoma (Motzer 1996).
Tumour markers refer to substances found in the blood, urine or body tissues which, when elevated, are indicative of cancers including breast, colorectal, gastrointestinal, ovarian, lung and pancreatic cancers as well as glioma, lymphoma and leukaemia. The most commonly used tumour marker is the prostate-specific antigen (PSA) blood test used to detect prostate cancer in men (Sturgeon 2009).
Sputum cytology refers to the analysis of sputum to detect the presence of a pulmonary malignancy and is most commonly used to detect lung cancer cells.
Polymerase chain reaction (PCR) based tests are relatively new tests used to identify and measure circulating tumour DNA and RNA markers extracted from serum, plasma, saliva and urine. Initial evidence suggests that using this combination of DNA and RNA markers may diagnose over 80% of lung cancer cases (Bremnes 2005).
Chest X-rays can show evidence of metastases, effusions or the primary tumour and are often the first diagnostic evaluation for patients with suspected lung cancer.
A CCT scan of the abdomen and pelvis is used for diagnosis and in combination with chest CT to determine the stage of the cancer. The CT scan takes pictures of the body from different angles and gives a series of cross sections to build up a detailed picture of inside the body showing the tumour size and location as well as its proximity to major organs.
A mammogram is an X-ray of the chest which is used to detect calcification, very early non-invasive cancers such as ductal carcinoma in situ (DCIS) and breast cancer. In the UK mammograms are routinely offered as an NHS breast screening programme to all women aged 50 to 70 years.
How the intervention might work
The interventions for detecting an underlying cancer will enable a diagnosis of cancer-associated VTE to be made or excluded. This will enable the patient to receive appropriate anticoagulation with LMWH or vitamin K antagonist, for three or six months, and for any underlying cancer to be treated without the need for additional symptoms to emerge before it is diagnosed. It has been shown that the combination of tests recommended by NICE detects cancer in approximately 10% of patients with a first episode of unprovoked VTE with no prior cancer diagnosis (Piccioli 2004).
Why it is important to do this review
The pharmacological management of VTE in patients with and without cancer is considerably different, both in terms of choice and duration of anticoagulation. Therefore, an appropriate cancer diagnosis would ensure that patients received the optimal form and duration of anticoagulation which, in turn, could reduce the overall population VTE recurrence rate and associated morbidity. Establishing whether a patient with an apparently unprovoked VTE has an underlying cancer is important since this may lead to cancer diagnosis at an earlier, potentially curative stage, avoiding the risk of cancer progression whilst waiting for additional symptoms. This may in turn lead to improvements in cancer-related mortality and morbidity. To date, no systematic review has been conducted to measure the effectiveness of testing for cancer in patients with an unprovoked VTE. This review will provide evidence as to whether such tests for underlying cancer, followed by appropriate alteration in the management or treatment of VTE, or both, are effective in reducing morbidity (VTE recurrence) and mortality (VTE and cancer-associated).
- To determine whether testing for undiagnosed cancer in patients with a first episode of unprovoked VTE (DVT or PE) is effective in reducing cancer mortality and morbidity.
- To determine whether testing for undiagnosed cancer in patients with a first episode of unprovoked VTE (DVT or PE) is effective in reducing VTE-related mortality and morbidity.
- To determine which tests for cancer are most useful.
Criteria for considering studies for this review
Types of studies
Randomised and quasi-randomised trials in which patients with an unprovoked VTE were allocated to receive different tests for cancer or no further tests. We will primarily look at randomisation within three months of a VTE, as used in the SOMIT trial (Piccioli 2004). However, we will also include trials where randomisation occurred at different time points. We will include this as a subgroup analysis. We will include published studies and studies in progress if preliminary results are available. Non-English studies will also be eligible for inclusion in the review.
Types of participants
Patients with a first episode of unprovoked VTE (DVT or PE) with no pre-existing or clinically apparent cancer diagnosis.
Types of interventions
Tests for cancer (e.g. complete blood count, serum calcium, liver function test, urinalysis, chest X-ray, all forms of computed tomography imaging, mammogram, tumour markers, sputum cytology, ultrasonography, positron emission tomography scan and colonoscopy) versus no tests for cancer or alternative tests, followed by appropriate treatment for cancer or change in VTE treatment regime, or both. We will not include studies where these tests are routinely used in all arms of the studies. However, we will include any study that focuses on some other aspect of care than cancer only if the test for cancer is the subject of randomisation.
Types of outcome measures
- Non-cancer-related mortality (death due to some cause other than cancer or cancer-related treatment).
- Cancer-related mortality (defined as death due to a malignant disease itself, or death due to complications of treatments or procedures to diagnose or treat the cancer).
- VTE-related mortality (fatal PE). PE diagnosed "on the basis of a lung scan indicating a high probability of its presence, as indicated by the presence of new or enlarged areas of segmental perfusion defects with ventilation–perfusion mismatch; an abnormal perfusion scan with documentation of new or recurrent deep-vein thrombosis; the presence of non-enhancing filling defects in the central pulmonary vasculature on helical computed tomography; a finding of intraluminal filling defects on pulmonary angiography; or evidence of fresh pulmonary embolism at autopsy" (Lee 2003b). Fatal PE including probable fatal PE and unexplained sudden death will be used if reported as defined by individual studies.
- VTE-related morbidity (e.g. frequency of recurrent VTE). Recurrent DVT diagnosed if a previously compressible proximal venous segment or segments could no longer be compressed on ultrasonography or if there were constant intraluminal filling defects in two or more projections on venography. Unequivocal extension of the thrombus required for the diagnosis of recurrence if the results were abnormal on previous testing (Lee 2003b).
- Complications of anticoagulation (e.g. warfarin versus LMWH-associated bleeding). We will report on major bleeding and minor bleeding if reported in the included studies. Major bleeding includes bleeding associated with death, bleeding at a critical site (intracranial, intraspinal, intraocular, retroperitoneal or pericardial area), bleeding resulting in a need for a transfusion of at least two units of blood, or bleeding leading to a drop in haemoglobin of at least 2.0 g per decilitre (Lee 2003b). Minor bleeding includes any other bleeding.
- Side effects of cancer tests (e.g. radiation exposure, bleeding, as defined in included studies).
- Characteristics of diagnosed cancer (e.g. primary tumour, stage, localised (curable) versus advanced (palliative) as defined in included studies).
- Time to cancer diagnosis, as defined in included studies.
- Frequency of an underlying cancer diagnosis (that is the number of times cancer has been diagnosed through screening following an unprovoked VTE as defined in included studies).
- Patient satisfaction (if assessed in individual studies, we will report results descriptively using the definition provided by the trialists).
Search methods for identification of studies
There will be no restriction on date or language of publication.
The Cochrane Peripheral Vascular Diseases Group Trials Search Co-ordinator (TSC) will search the Specialised Register and the Cochrane Central Register of Controlled Trials (CENTRAL), part of The Cochrane Library, (www.thecochranelibrary.com). See Appendix 1 for details of the search strategy which will be used to search CENTRAL. The Specialised Register is maintained by the TSC and is constructed from weekly electronic searches of MEDLINE, EMBASE, CINAHL and AMED, and through handsearching relevant journals. The full list of the databases, journals and conference proceedings which have been searched, as well as the search strategies used, are described in the Specialised Register section of the Cochrane Peripheral Vascular Diseases Group module in The Cochrane Library (www.thecochranelibrary.com).
The TSC will search the following trial databases for details of ongoing and unpublished studies:
- World Health Organization International Clinical Trials Registry (http://apps.who.int/trialsearch/);
- ClinicalTrials.gov (http://clinicaltrials.gov/);
- Current Controlled Trials (http://www.controlled-trials.com/);
- Nederlands Trials Register (http://www.trialregister.nl/trialreg/admin/rctsearch.asp).
Searching other resources
We will search the reference lists of relevant articles retrieved by electronic searches for additional citations. Furthermore, we will search conference proceedings of the following societies to find abstracts: the International Society for Thrombosis and Haemostasis (ISTH) (from 2003 to present) and the American Society for Haematology (ASH) (2004 to present).
Data collection and analysis
Selection of studies
One review author (LR) will use the selection criteria to identify trials for inclusion. The second review author (SEY) will independently confirm this selection and any disagreements will be resolved by discussion.
Data extraction and management
Two review authors (LR, SEY) will independently extract the data. We will record information about the trial design, VTE definition and investigations to confirm diagnosis, baseline characteristics of patients and tests for cancer. We will record non cancer-related mortality, cancer-related mortality and VTE-related mortality data as the primary outcome measures. We will collect information on VTE-related morbidity (e.g. frequency of recurrent VTE), complications of anticoagulation (e.g. warfarin versus LMWH-associated bleeding), side effects of cancer tests (e.g. radiation exposure, bleeding), characteristics of diagnosed cancer (e.g. primary tumour, stage, localised (curable) versus advanced (palliative)), time to cancer diagnosis, frequency of an underlying cancer diagnosis and patient satisfaction in accordance with the secondary outcome measures. We will contact authors of included studies for further information if clarification is required. We will resolve any disagreements in data extraction and management by discussion.
Assessment of risk of bias in included studies
Two review authors (LR, SEY) will independently use The Cochrane Collaboration's tool (Higgins 2011) for assessing risk of bias for each of the included studies. The tool provides a protocol for judgements on sequence generation, allocation methods, blinding, incomplete outcome data, selective outcome reporting and any other relevant biases. We will judge each of these domains as either high, low or unclear risk of bias according to Higgins 2011 and provide support for each judgement. We will present the conclusions in a 'Risk of bias' table. Any disagreements will be resolved by discussion.
Measures of treatment effect
We will base analysis on intention-to-treat data from the individual clinical trials. The majority of outcomes are binary measures (mortality, morbidity, complications, side effects, characteristics of diagnosed cancer, frequency of an underlying cancer diagnosis). For these outcomes we will compute odds ratios (ORs) using a random-effects model and calculate the 95% confidence intervals (CI) of the effect sizes. For time to cancer diagnosis we will compute hazard ratios (HR) while for patient satisfaction we will report results descriptively (Deeks 2011).
Unit of analysis issues
The unit of analysis within each trial will be the individual participant.
Dealing with missing data
We will seek information about dropouts, withdrawals and other missing data and, if not reported, we will contact the study authors.
Assessment of heterogeneity
We will assess heterogeneity between the trials by visual examination of the forest plot to check for overlapping CIs, using the Chi
Assessment of reporting biases
We will assess reporting biases such as publication bias using funnel plots. There are many reasons for funnel plot asymmetry and we will consult the Cochrane Handbook for Systematic Reviews of Interventions to aid the interpretation of the results (Sterne 2011).
The review authors will independently extract the data. One review author (LR) will input the data into RevMan (RevMan 2012). A second review author (SEY) will cross-check data entry and we will resolve any discrepancies by consulting the source publication.
We will use a random-effects model to meta-analyse the data. We will stratify analyses according to the individual test being assessed. If data are available in individual studies, we will stratify further analysis similarly to the combination of tests used in the SOMIT trial (Piccioli 2004).
If the I
Subgroup analysis and investigation of heterogeneity
- DVT or PE
- Cancer site
- Treatment post-investigation with vitamin K antagonist or LMWH
- Duration of anticoagulation (e.g. three or six months)
- Age and gender of patients (comparing those in age/gender groups for national screening programmes to those not in these age/gender groups)
- Time of randomisation after VTE diagnosis (within three months compared with after three months)
We will perform sensitivity analysis by excluding studies that are judged to be at high risk of bias.
Appendix 1. CENTRAL search strategy
Contributions of authors
LR: drafted the protocol, will select studies for inclusion, assess the quality of studies, carry out data extraction, perform data analysis and write the review.
RA: contributed to the protocol and will provide clinical input to the review.
SEY: will select studies for inclusion, assess the quality of the studies and carry out data extraction.
Declarations of interest
Sources of support
- No sources of support supplied
- National Institute for Health Research (NIHR), UK.LR is supported by a programme grant from the NIHR.
- Chief Scientist Office, Scottish Government Health Directorates, The Scottish Government, UK.The PVD Group editorial base is supported by the Chief Scientist Office.
- National Institute for Health Research (NIHR), UK.The PVD Group editorial base is supported by a programme grant from the NIHR.