Introduction
Cancer-associated thrombosis remains an important complication of the cancer journey and continues to increase in incidence. The current evidence only informs the management of CAT for the first 6 months of diagnosis, with no strong trial data to inform what should be done for patients with ongoing active cancer reaching 6 months’ anticoagulation treatment. Despite the need to address such a significant knowledge gap, this feasibility study strongly suggested that a Phase III RCT to answer this question was unlikely to recruit the necessary number of patients within a reasonable time frame that made funding such a study worthwhile. The lessons learned from the design, set-up and conduct of the ALICAT study have been highly informative and offer useful insights pertaining to the design of future similar studies. The challenges faced and lessons learned are discussed below. The challenges are categorised according to whether they are logistical or attitudinal.
Challenges: procedural
On reflection, the design of the ALICAT study was an ambitious one. Although trials units have considerable experience co-ordinating multicentre studies, undertaking a RCT in three clinical environments and across two nations brought additional challenges, some of which could be avoided if all stakeholders involved in clinical research streamlined their processes, minimised repetition of processes and addressed bottlenecks arising from arguably unnecessary bureaucracy. Although the best research questions are the simple ones and simple trial designs are more likely to complete recruitment, one should not always pick the low-hanging fruit. By only focusing on the simple common questions, one will neglect delivering the answers to more complex or rare problems.
One site’s opening was delayed by 2 months while awaiting approval from the trust research and development (R&D) radiation lead to sign off the study, to say there were no additional radiation risks posed by the trial. What took 1 day for approval in one organisation took 56 days in another, with no excuse offered save consultant workload. Such examples reinforce the responsibility that all those who commit to involvement in research have. One cancelled risk review committee here, one R&D lead who is off sick and not backfilled there, can lead to a 6-month delay in a trial set-up before any legitimate concerns are raised.
The slightly different approvals processes between the NISCHR in Wales and NIHR in England led to a significant delay in setting up sites outside of Wales. This suggests a degree of disconnect between nations, which should be remediable.
Another challenge identified through the pathways mapping section of the study was the fact that CAT is managed differently across health organisations with no clear consensus nor consistency for ownership. Although the management of thrombosis is usually a haematology issue, which is then taken on by primary care, this is not the case for CAT. Haematology services seem best suited to manage CAT; it is frequently taken on by the patient’s oncologist who often does not have the time or equivalent expertise to manage the more complex cases that may arise. From the perspective of the feasibility study, the mapping exercise highlighted the breadth of environments and clinical pathways across which CAT is managed, thereby outlining the challenge of recruiting to such a RCT.
Challenges: attitudinal
Even if it were possible to establish a robust recruitment strategy and environment, it appears that we would still be unlikely to recruit sufficient patients to the study. From the RCT component of the ALICAT trial, only 5 of the 32 eligible patients identified consented to randomisation, suggesting that patients do not wish to participate in such trials. Furthermore, the scoping exercises suggest that clinicians likewise may be reticent to recruit to such a study since they already hold fixed views regarding the management of VTE beyond 6 months. These issues are discussed further in Influence of prior experience and Belief in equipoise.
Influence of prior experience
From the embedded qualitative study, it was possible to ascertain reasons for patients declining randomisation into the RCT and these seemed equally divided between a fear of randomisation to continue LMWH and a fear of randomisation for stopping it. It is clear that the patient experience of the index VTE event had a strong influence on their attitudes and beliefs regarding anticoagulation; in particular, there appears to be a divide between those diagnosed with symptomatic and incidental VTE.
For those patients with symptomatic VTE, there remains a strong association with a distressing symptomatic event and a subsequent improvement of symptoms with commencing LMWH treatment. Such patients lived with a concern that by stopping the LMWH, the VTE would come back and were therefore reticent to risk participation into a study, which may discontinue the drug. Conversely, those with incidental and, hence, largely asymptomatic VTE had few associations with distressing symptoms. Rather, they perceived the requirement to self-inject for 6 months an inconvenience, which they could not wait to desist from. Participation in a study, which may risk further anticoagulation, did not appeal.
Belief in equipoise
The responses from patients declining randomisation, suggest they were influenced largely by prior experience of VTE and injecting LMWH. While they acknowledged an evidence deficit in the management of VTE they made their decisions based on their fixed views on whether or not they wished to continue LMWH, and not whether or not they needed to. To some extent, they used the evidence deficit to justify their decision and continuing or not continuing their LMWH in the absence of strong data did not appear to cause distress.
From the clinician interviews and focus groups there was acknowledgement of the lack of evidence around the management of VTE beyond 6 months. However, anticoagulation beyond 6 months off-license had become such custom and practice that most clinicians held fixed beliefs that this was the best form of treatment. These beliefs were further reinforced by satisfactory, if not formally evaluated, clinical outcomes. Interestingly, the haematology views were based far more on knowledge of the existing data and understanding the complexities of thrombosis and haemostasis, and as a group gave the haematologist a cogent case for why they feared the study lacked equipoise. Of those clinicians willing to randomise patients to the study, there was a clear sense that they would ‘cherry pick’ the cases in which they perceived there was less certainty about VTE recurrence risk. As such, those perceived to be more thrombogenic would not be approached, thereby leading to a highly biased sampling strategy.
Although there were established views that patients should continue with LMWH, there was acknowledgement that the required ongoing dose needed clarification.
It was interesting to observe that clinicians would be more willing to recruit to a study which randomised patients to weight-adjusted (full-dose) LMWH versus a primary thromboprophylaxis dose. This appears to be a more acceptable study design for clinicians and is likely to overcome any challenges of clinician recruitment. It is not clear whether or not this would lead to increased consent from patients; it is unlikely to change the views of patients with incidental VTE, as their reason for declining randomisation was to ensure they stopped the LMWH. However, the remaining majority of potential patients may find this study design more acceptable.
Priorities for future research
The management of CAT will remain a clinical challenge as it is a complication of cancer, which is increasing in parallel with the use of chemotherapy in patients with metastatic and curable disease.
Currently research strategies have focused on CAT as a single entity, which behaves consistently across the different tumour types. In reality, is seems counterintuitive to research the properties of CAT in breast cancer patients alongside those with advanced pancreatic cancer since the diseases have different levels of tumour-related thrombogenicity, extrinsic thrombotic risk factors and overall prognoses. An urgent consideration for the CAT research world is to focus research strategies on specific tumour types. One such option would be to focus on a tumour which is known to be less thrombogenic overall but still carries uncertainty of best management at the 6-month anticoagulation stage. Such a cancer would be breast cancer, which is not only common, but also common as a metastatic disease with a favourable overall prognosis.
Another area of research that is much needed is an adequately powered non-inferiority trial between one of the new oral anticoagulants and LMWH. To date, the data support the use of new oral anticoagulants in the non-cancer population, with non-inferiority to warfarin. However, it is yet to be evaluated head to head in the cancer population against LMWH. If an oral agent were to demonstrate non-inferiority to the current injectable gold standard, this would negate the reasons for non-recruitment based on disliking injections.
Conclusion
The ALICAT study has established that at the current time it is not feasible to recruit sufficient patients with advanced cancer and VTE to a RCT comparing 6 months’ anticoagulation with indefinite anticoagulation. The reasons for this are multifactorial, and include practical and procedural challenges, patients’ attitudes and those of the recruiting clinicians. Current practice to continue long-term anticoagulation, although not evidence based, is supported strongly by consensus and clinician experience. A study to confirm or refute that current practice is the most appropriate treatment regime is unlikely to be supported by clinicians in particular, until current practice is demonstrated to be financially unsustainable or clinically unsafe.
