Abstract
Introduction In cancer patients, current guidance suggests similar treatment for incidental and symptomatic venous thromboembolism (VTE), mainly based on retrospective data. We aimed to evaluate anticoagulant therapy in cancer patients with incidental and symptomatic VTE.
Methods The Hokusai VTE Cancer Study was a randomised controlled trial comparing edoxaban with dalteparin for cancer-associated VTE. The primary outcome was the composite of first recurrent VTE or major bleeding. Secondary outcomes included major bleeding, recurrent VTE and mortality. Outcomes in patients with incidental and symptomatic VTE were evaluated during the 12-month study period.
Results 331 patients with incidental VTE and 679 patients with symptomatic VTE were enrolled, of whom the index event was confirmed by an independent radiologist. Median durations of anticoagulant treatment were 195 and 189 days, respectively. In patients with incidental VTE, the primary outcome occurred in 12.7% of patients, major bleeding in 6.6% of patients and recurrent VTE in 7.9% of patients. Out of the 26 VTE recurrences in patients with incidental VTE, five (31%) were incidental, seven (44%) were symptomatic and four (25%) were deaths for which pulmonary embolism could not be ruled out. In patients with symptomatic VTE, the primary outcome occurred in 13.8% of patients, major bleeding in 4.9% of patients and recurrent VTE in 10.9% of patients. All-cause mortality was similar in both groups.
Conclusion Clinical adverse outcomes are substantial in both cancer patients with incidental and symptomatic VTE, supporting current guideline recommendations that suggest treating incidental VTE in the same manner as symptomatic VTE.
Abstract
In cancer patients with incidental venous thromboembolism (VTE), the risk of recurrent VTE is substantial despite initiation of anticoagulant treatment. Incidental VTE is a serious condition, which should be treated similar to symptomatic VTE. http://bit.ly/2qM4XUp
Introduction
In clinical practice ∼20% of all VTE events are related to cancer [1]. VTE can manifest as deep-vein thrombosis (DVT) and/or pulmonary embolism, and impact a cancer patient's disease journey by increasing morbidity and the risk of mortality [2, 3]. With improvements in computed tomography scanning technology, and its increased utilisation for diagnosis and monitoring of cancer, the proportion of cancer patients with incidental VTE has risen over time [4–7]. Nowadays, about half of all pulmonary embolic events in cancer patients are classified as incidentally detected on routine imaging scans for cancer staging or evaluation of cancer treatment response [8].
International guidelines suggest the same initial and long-term treatment for cancer patients with incidental VTE as for those with symptomatic VTE, mainly based on the comparable risk of recurrent VTE observed in retrospective cohort studies [9–11]. A recent prospective cohort study of 695 cancer patients with incidental pulmonary embolism reported a 1-year incidence of recurrent VTE of 6%, confirming the clinical relevance of these events [12]. Translating these results to current clinical practice is challenging because some patients received subtherapeutic doses of anticoagulants, only 5% received a direct oral anticoagulant and a control group with symptomatic VTE was not included.
The Hokusai VTE Cancer Study was the first randomised controlled trial evaluating treatment of cancer-associated VTE, which also included patients with incidental VTE [13]. Therefore, it provides a unique opportunity to compare clinical outcomes in patients with incidental and symptomatic VTE, evaluate the safety and efficacy of edoxaban versus dalteparin in these groups, and investigate risk factors for recurrent VTE in patients with incidental VTE.
Methods
Patients
The Hokusai VTE Cancer Study was an open-label trial that randomised 1050 cancer patients with symptomatic or incidentally detected VTE to either oral edoxaban or subcutaneous dalteparin once daily. Patients with cancer-associated VTE were randomised in a 1:1 ratio with the use of an interactive Web-based system with stratification of bleeding risk factors and the requirement of anticoagulant dosage reduction. The treatment duration was 6–12 months, at the discretion of the treating physician. Both VTE and recurrent VTE were considered incidental if diagnosed on radiological imaging tests performed for reasons other than a clinical suspicion of VTE. The sample size of the trial was based on the assumption of equal effectives of dalteparin and edoxaban in cancer patients and a rate of the primary outcome of 20% at 12 months of follow-up. The full list of inclusion and exclusion criteria, study protocol, and detailed description of the study design were reported previously (ClinicalTrials.gov identifier NCT02073682) [13, 14]. The CONSORT checklist and flow diagram are shown in supplementary figure S1.
All patients provided written informed consent and the study was approved by the ethical committees of participating centres.
The present post hoc analysis of this trial was restricted to patients who received at least one dose of study drug and had their index event centrally adjudicated by an independent radiologist, blinded to treatment allocation. The radiologist also evaluated the thrombus location for each DVT and pulmonary embolism, and assessed the extent of pulmonary embolism based on a predefined perfusion score (supplementary material) [15].
Outcomes
The primary outcome of the current study was the composite of first recurrent VTE or major bleeding during the 12-month study period. Secondary outcomes included recurrent VTE, major bleeding and mortality during the 12-month study period. Additionally, outcomes were evaluated during the 6-month study period, which was the minimal treatment duration with the assigned study drug, and during the on-treatment period, defined as during treatment or up to 3 days after the last dose of study drug. Recurrent VTE was defined as a new symptomatic or incidental DVT or pulmonary embolism, fatal pulmonary embolism, or unexplained death for which pulmonary embolism could not be ruled out. Major bleeding was defined according to the criteria of the International Society on Thrombosis and Haemostasis [16] and the severity of major bleeding was classified in four categories according to a previously proposed method [17]. All study outcomes in the Hokusai VTE Cancer Study were adjudicated by an independent committee unaware of treatment allocation.
Statistical analysis
Differences between patients with incidental and symptomatic VTE were assessed using standard descriptive statistics. Cumulative incidences were presented as crude proportions and time-to-event curves were constructed using the Kaplan–Meier method. Hazard ratios were computed with 95% confidence intervals using a Cox proportional hazards model. In a second model, we adjusted for age, sex, index event (pulmonary embolism versus DVT), anticoagulant treatment (edoxaban versus dalteparin), Eastern Cooperative Oncology Group (ECOG) performance status (2 versus 0 or 1), cancer type (haematological, lung, breast, gastrointestinal, urogenital, gynaecological or other), previous VTE and tumour stage (distant metastasis versus other). For patients with incidental VTE, risk factors for recurrent VTE were assessed in a univariable Cox proportional hazards model. In a sensitivity analysis, time-to-event curves of the main study outcomes were constructed with the competing risk approach proposed by Fine and Gray [18]. Likewise, the association between risk factors and recurrent VTE in patients with incidental VTE were additionally evaluated by computing subdistribution hazard ratios with the competing risk approach. A p-value <0.05 was considered statistically significant. Statistical analyses were performed in R version 3.5.1 (www.R-project.org).
Results
Patients with incidental VTE
Of the 1046 cancer patients enrolled in the Hokusai VTE Cancer Study, 340 (32%) had a reported incidental VTE, which was radiologically confirmed in 331 patients. Baseline characteristics are shown in table 1. 16 patients (4.8%) had haematological cancer and 315 patients (95.2%) had a solid tumour. The index event was incidental pulmonary embolism with or without DVT in 290 patients (87.6%) and incidental DVT alone in 41 patients (12.4%). Of the 290 patients with an incidental pulmonary embolism, the anatomical extent was considered limited in 112 patients (38.6%), intermediate in 132 patients (45.5%) and extensive in 46 patients (15.9%).
The primary outcome, which was the composite of first recurrent VTE or major bleeding during the 12-month study period, occurred in 42 patients (12.7%) (figure 1a and table 2). 26 patients (7.9%) had recurrent VTE (figure 1b), of whom 16 patients (4.8%) had recurrent pulmonary embolism with or without DVT and 10 patients (3.0%) had recurrent DVT only. Of the 16 patients with recurrent pulmonary embolism, seven (43.8%) were symptomatic, five (31.3%) were incidental and there were four (25.0%) deaths for which pulmonary embolism could not be ruled out. Of the 10 recurrent DVT events, six (60%) were symptomatic and four (40%) were incidental. On-treatment recurrent VTE occurred in 19 patients (5.7%). Major bleeding occurred in 22 patients (6.6%) (figure 1c) and on-treatment major bleeding occurred in 20 patients (6.0%), of which nine patients (45%) had a severe clinical bleeding presentation (category 3 or 4). During the 12-month study period, 123 patients (37.2%) with incidental VTE died (supplementary figure S2). The sensitivity analysis restricted to study outcomes during the first 6 months of follow-up was comparable to the primary analysis outcome and is presented in supplementary table S1. The competing risk approach showed similar results (supplementary figure S5a–c).
Patients with symptomatic VTE
Symptomatic VTE was the qualifying event in 706 patients; 679 events were radiologically confirmed centrally. Baseline characteristics of patients with symptomatic VTE are shown in table 1. The symptomatic index event was pulmonary embolism with or without DVT in 353 patients (52.0%) and DVT only in 326 patients (48.0%). Of the 353 patients with symptomatic pulmonary embolism, the anatomical extent was considered limited in 117 patients (33.1%), intermediate in 129 patients (36.5%) and extensive in 107 patients (30.3%).
The primary outcome occurred in 94 patients (13.8%) during the 12-month study period (figure 1a). Recurrent VTE occurred in 74 patients (10.9%), of whom 39 patients (5.7%) had recurrent pulmonary embolism with or without DVT and 35 patients (5.2%) had recurrent DVT alone (figure 1b). Of the 39 recurrent pulmonary embolism events, 23 (58.9%) were symptomatic, 10 (25.6%) were incidental and six (15.4%) were deaths for which pulmonary embolism could not be ruled out. Of the 35 recurrent DVT events, 32 (91.4%) were symptomatic and three (8.6%) were incidental. On-treatment recurrent VTE occurred in 53 patients (7.8%). Major bleeding occurred in 33 patients (4.9%) and on-treatment major bleeding occurred in 26 patients (3.8%) (figure 1c), of which 12 patients (46.2%) had a severe presentation (category 3 or 4). Study outcomes during the 12-month study period are presented in table 2. Supplementary table S1 shows outcomes during the 6-month study period. The competing risk analysis showed comparable results (supplementary figure S5a–c).
Outcomes in patients with incidental versus symptomatic VTE
Dalteparin was given to 171 patients (51.7%) with incidental VTE and 337 patients (49.6%) with symptomatic VTE (table 1). The median (interquartile range) anticoagulant duration was 195 (92–352) days in those with incidental VTE and 189 (76–353) days in those with symptomatic VTE (p=0.50). More patients had pulmonary embolism with or without DVT as the qualifying index event in the group with incidental VTE (87.6%) compared with the group with symptomatic VTE (52.0%) (p<0.001). While the thrombus was considered extensive in 15.9% of patients with incidental pulmonary embolism, it was considered extensive in 30.3% of patients with symptomatic pulmonary embolism (p<0.001). Fewer patients with incidental VTE had haematological cancer compared with those with symptomatic VTE (4.8% versus 13.4%) (p<0.001).
During the complete study period, there was no statistically significant difference between patients with incidental and symptomatic VTE in the risk of the primary outcome (crude HR 0.86, 95% CI 0.60–1.24; adjusted HR 0.84, 95% CI 0.56–1.25) (figure 1), recurrent VTE (crude HR 0.68, 95% CI 0.43–1.06; adjusted HR 0.68, 95% CI 0.42–1.11) and on-treatment major bleeding (crude HR 1.54, 95% CI 0.86–2.75; adjusted HR 1.22, 95% CI 0.64–2.33) (table 2). All-cause mortality after 12 months was comparable in patients with incidental and symptomatic VTE (37.2% and 38.1%, respectively; crude HR 0.92, 95% CI 0.74–1.14; adjusted HR 0.83, 95% CI 0.65–1.05).
Edoxaban versus dalteparin
In patients with incidental VTE, the primary outcome occurred in 19 out of 160 patients (11.9%) randomised to edoxaban and in 23 out of 171 patients (13.5%) who received dalteparin (HR 0.94, 95% CI 0.51–1.73) (supplementary figure S3). In patients with symptomatic VTE, the primary outcome occurred in 46 out of 355 patients (13.5%) in the edoxaban group and 48 out of 351 patients (14.2%) in the dalteparin group (HR 0.95, 95% CI 0.63–1.42) (supplementary figure S4). There was no evidence of effect modification of incidental versus symptomatic VTE in the occurrence of the primary outcome (pinteraction=0.97). There were no significant differences in any of the other study outcomes after 12 months of follow-up (supplementary table S2).
Risk factors for recurrent VTE in patients with incidental VTE
In patients with incidental VTE, age, male sex, anatomical extent of pulmonary embolism or body mass index were not associated with the risk of recurrent VTE (table 3). Patients with a decreased performance status were more likely to experience recurrent VTE during 12 months of follow-up (ECOG performance status 2 versus 0: HR 5.24, 95% CI 1.81–15.18). The subdistribution hazard ratios estimated with the competing risk approach yielded similar results (table 3).
Discussion
The present study evaluated clinical outcomes in cancer patients with incidental and symptomatic VTE using data from the Hokusai VTE Cancer Study. The main finding is that the risk of recurrent VTE in patients with incidental VTE was significant: 7.9% of patients with incidental VTE developed recurrent VTE during 12 months of follow-up despite initiating anticoagulant treatment. Approximately half of the recurrent VTE events in patients with incidental VTE were symptomatic and a quarter were deaths for which pulmonary embolism could not be ruled out. These findings indicate that incidental VTE in cancer patients is a serious condition and support the current guidelines, which recommend that incidental VTE should be treated in the same manner as symptomatic VTE.
The detailed findings from the current analysis merit closer consideration. The 12-month risk of recurrent VTE was lower in patients with incidental VTE than in those with symptomatic VTE (7.9% versus 10.9%), although this difference was not statistically significant (HR 0.68, 95% CI 0.43–1.06). A possible explanation could be differences in patient characteristics, which may impact the risk of VTE recurrence. For example, patients with symptomatic VTE more frequently had extensive pulmonary embolism (30.3% versus 15.9%; p<0.001) and poor ECOG performance status (25.8% versus 19.9%; p=0.066) (table 1), which are potential risk factors for recurrence. Also, fewer patients with incidental VTE had haematological cancer, presumably because routine imaging for monitoring of cancer progression, on which incidental findings are detected, is less frequently used in these patients compared with those with solid cancers. Nonetheless, there were no substantial changes in the hazard ratios for recurrent VTE after adjusting for these confounding factors, suggesting that symptomatic VTE may be associated with an inherently higher risk of recurrence. Conversely, the 12-month major bleeding risk was numerically higher in patients with incidental VTE (6.6%) than in those with symptomatic VTE (4.9%) (HR 1.31, 95% CI 0.76–2.24). This difference could be explained by the higher prevalence of gastrointestinal cancer in patients with incidental VTE (35.4% versus 26.1%), a cancer type found to be associated with an increased risk of major bleeding [19, 20]. There was no difference in overall survival between patients with incidental and symptomatic VTE, which is in line with the findings of several other studies [8, 21–23]. In the Hokusai VTE Cancer Study, the decision to continue or discontinue anticoagulation after 6 months of follow-up was at the discretion of the treating clinician [13]. The median duration of anticoagulant treatment during total follow-up was similar between patients with incidental VTE and symptomatic VTE (195 versus 189 days). This indicates that most of the clinicians in this study adhered to the guidelines, which suggests similar treatment durations for patients with incidental or symptomatic VTE. Within the groups of incidental and symptomatic VTE, there was no difference in the occurrence of the primary outcome between patients randomised to edoxaban and dalteparin, which confirms that oral edoxaban is an acceptable alternative to injectable dalteparin in both patient groups.
Several studies previously evaluated clinical outcomes in cancer patients with incidental VTE. In line with our findings, a prospective cohort study of 695 cancer patients with incidental pulmonary embolism reported 12-month incidences of recurrent VTE and on-treatment major bleeding of 6.0% and 5.8%, respectively [12]. Comparable results were obtained in an individual patient data meta-analysis of mostly retrospective data, comprising 11 cohort studies and 926 cancer patients, which reported a pooled 6-month risk of 5.8% for recurrent VTE and 4.7% for major bleeding [24]. However, Den Exter et al. [22] observed a much higher 12-month recurrent VTE risk of 13.3% in a small retrospective cohort study of 51 patients with incidental pulmonary embolism. Higher rates of recurrent VTE in patients with incidental VTE were also observed in studies by Font et al. [8] and Amato et al. [25]. Potential explanations for these differences include the larger proportion of patients receiving vitamin K antagonists in these older studies, a treatment known to be associated with a higher risk of recurrence, misclassification of incidental VTE or study outcomes and enrolment of cancer types with a higher risk of recurrence. In contrast, we present findings from a prospective, randomised controlled trial with strict adjudication of index event and study outcome events by specialists blinded to treatment allocation.
Few studies previously compared outcomes between patients with incidental and symptomatic VTE, reporting conflicting findings. The study by Den Exter et al. [22] demonstrated that the risk of recurrence is comparable in both groups. Recently, the pilot SELECT-D study compared rivaroxaban with dalteparin in 406 patients with cancer-associated VTE, of whom 213 patients (52.5%) had incidental pulmonary embolism or DVT at baseline [20]. The authors reported that the risk of recurrent VTE was significantly higher in patients with symptomatic VTE than in those with incidental VTE (HR 2.8, 95% CI 1.2–6.4). Differences between our results and those of the SELECT-D study may reflect the smaller sample size, the higher proportion of incidental VTE and the shorter reported follow-up duration of 6 months [20].
We have investigated potential risk factors for predicting recurrent VTE in patients with incidental VTE (table 3) and identified only a poor ECOG performance status of ≥2. Both the perfusion score and anatomical pulmonary embolism clot extent were not significantly associated with recurrent VTE in patients with incidental pulmonary embolism (extensive versus limited clot: HR 0.43, 95% CI 0.10–1.93; perfusion score: HR 3.07, 95% CI 0.15–62.2).
The strengths of the current analysis include the complete follow-up of almost all randomised patients [13]. In contrast to previous studies, anticoagulant treatment was uniform in patients with incidental and symptomatic VTE, and outcomes were adjudicated by an independent committee unaware of treatment allocation, thus minimising observer bias. Several limitations need to be taken into consideration. Although the open-label design could have influenced treatment duration in patients with symptomatic and incidental VTE, there were no differences between these groups. It should be acknowledged that incidental VTE is not necessarily asymptomatic. However, we followed the terminology and definition recommended by the International Society on Thrombosis and Haemostasis [26]. Since the Hokusai VTE Cancer Study was not primarily designed for the current research question, power to detect significant differences may have been limited. Although hazard ratios were adjusted for several potential confounders, we cannot exclude residual confounding. Since death is a competing risk for the occurrence of study events, Kaplan–Meier curves could have led to overestimating the cumulative incidence of study outcomes. However, a competing risk approach yielded similar results.
The improved quality and routine utilisation of radiological imaging in cancer patients frequently lead to the diagnosis of incidental VTE. Our findings substantiate the high risk of recurrent VTE despite anticoagulant treatment in patients with incidental events and the results strengthen current guidelines, which recommend the same treatment for incidental VTE as is used for symptomatic VTE. Future randomised trials evaluating anticoagulant treatment in cancer patients should include patients with incidental VTE by default. Additional risk factors for incidental VTE need to be identified in order to better individualise anticoagulant treatment for this important patient group.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material erj-01697-2019_Supplementary_material
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-01697-2019.Shareable
Footnotes
This article has an editorial commentary: https://doi.org/10.1183/13993003.00028-2020
This article has supplementary material available from erj.ersjournals.com
Data from the Hokusai VTE Cancer Study (ClinicalTrials.gov identifier NCT02073682) may be obtained by a third party upon reasonable request.
Author contributions: Study conception and design: F.I. Mulder, N. van Es, M. Di Nisio, C. Ay, H.R. Büller. Data acquisition: M. Di Nisio, M. Carrier, A. Segers, P. Verhamme, T. Wang, J.I. Weitz, S. Middeldorp, G. Raskob, L.F.M. Beenen, H.R. Büller, N. van Es. Statistical analysis: F.I. Mulder, N. van Es, G. Zhang. Interpretation of the data: all authors. Drafting of the manuscript: all authors. Critical revision of the manuscript for important intellectual content: all authors. Final approval of the manuscript: all authors. H.R. Büller is guarantor and takes responsibility for the integrity of the work as a whole.
Conflict of interest: F.I. Mulder has nothing to disclose.
Conflict of interest: M. Di Nisio reports personal fees from Daiichi Sankyo outside the submitted work.
Conflict of interest: C. Ay reports personal fees from Bayer, Daiichi Sankyo, Pfizer and BMS, outside the submitted work.
Conflict of interest: M. Carrier reports personal fees from Daiichi Sankyo during the conduct of the study; grants and personal fees from BMS, LEO Pharma; personal fees from Pfizer, Sanofi and Bayer outside the submitted work.
Conflict of interest: F.T.M. Bosch has nothing to disclose.
Conflict of interest: A. Segers reports grants from Daiichi Sankyo during the conduct of the study; grants from Ionis Pharmaceuticals, Daiichi Sankyo and Janssen Pharmaceuticals outside the submitted work.
Conflict of interest: N. Kraaijpoel has nothing to disclose.
Conflict of interest: M.A. Grosso reports personal fees from Daiichi Sankyo outside the submitted work.
Conflict of interest: G. Zhang reports personal fees from Daiichi Sankyo outside the submitted work.
Conflict of interest: P. Verhamme reports grants and personal fees from Daiichi Sankyo, during the conduct of the study; grants and personal fees from Bayer Healthcare, Boehringer Ingelheim and LEO Pharma; personal fees from BMS, Portola, Medscape, Medtronic and Pfizer; and grants from Sanofi outside the submitted work.
Conflict of interest: T-Z. Wang reports nonfinancial support from Daiichi Sankyo during the conduct of the study.
Conflict of interest: J.I. Weitz reports personal fees from Daiichi Sankyo, during the conduct of the study; personal fees from Bayer Healthcare, BMS, Boehringer Ingelheim, Ionis Pharmaceuticals, Janssen, Johnson & Johnson, Pfizer, Portola, Medscape and Novartis outside the submitted work.
Conflict of interest: S. Middeldorp reports grants and personal fees from GSK, BMS/Pfizer, Aspen, Daiichi Sankyo; personal fees from Bayer, Boehringer Ingelheim; and grants from Sanquin.
Conflict of interest: G. Raskob reports personal fees from Daiichi Sankyo during the conduct of the study; personal fees from Bayer Healthcare, BMS, Boehringer Ingelheim, Eli Lilly, Janssen, Johnson & Johnson, Pfizer, Portola, Merck and Medscape outside the submitted work.
Conflict of interest: L.F.M. Beenen has nothing to disclose.
Conflict of interest: H.R. Büller reports personal fees from Daiichi Sankyo, during the conduct of the study; personal fees from Bayer Healthcare, BMS/Pfizer, Boehringer Ingelheim, Portola, Medscape, Eli Lilly, Sanofi Aventis and Ionis outside the submitted work.
Conflict of interest: N. van Es reports personal fees from Daiichi Sankyo during the conduct of the study; and personal fees from Pfizer outside the submitted work.
Support statement: This study received financial support from Daiichi Sankyo. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received August 30, 2019.
- Accepted October 20, 2019.
- Copyright ©ERS 2020