In the keynote address at Issues & Answers 2018 Dr Nicholas Boon (formerly consultant cardiologist and past-President of the British Cardiovascular Society) said we are at a crossroads in cardiology with a new era of personalised care replacing wholesale reliance on the outcomes of randomised controlled trials, guidelines and meta-analyses.
He encouraged delegates to be wary of putting all their trust in evidence-based guidelines, systematic reviews and clinical trials – as this approach has ‘lumped’ patients into ever-broadening disease categories and produced a ‘one size fits all’ approach to care that may well not be ‘best practice’ for individual patients (splitting).
The terms ‘lumping’ and ‘splitting’ were first used by Charles Darwin in classifying plant and animal species, and this approach was taken up by Professor Victor McKusick of The Johns Hopkins University School of Medicine (Baltimore, MY, USA) when considering disease classification. Dr Boon suggests that it can now be applied to aspects of modern-day medicine and clinical methodology (Figure 1).
Clinical trials and guidelines: the era of ‘lumping’
The randomised controlled trial (RCT) has been a great force for good but it has been sometimes difficult to translate the findings of trials into real-world medicine for a host of reasons including:
- Patients tend to be highly selected, low-risk homogenous subjects in RCTs
- Fixed treatment regimens rather than flexible real-life scenarios
- Placebo comparators rather than current standard of care means RCT outcomes can be outmoded.
Dr Boon gave the example of the large number of recent clinical trials in novel antidiabetic agents (including SGLT2 inhibitors, GLP-1 agonists and DPP-4 antagonists) that have provided much interesting data, but no clarity on how to apply these results in clinical practice.
Despite being excellent vehicles providing treatment options for acute organic disorders such as myocardial infarction, evidence-based guidelines and protocols are not so useful in providing solutions for chronic functional illnesses (Figure 2). In addition, RCTS require large numbers of patients and events to give the statistical weight to results and Dr Boon suggested this approach ushered in the era of ‘lumping’ in the 1970s with large ‘mega-trials’ and meta-analyses.
Figure 2. Value of evidence-based guidelines and protocols.
Although RCTs have provided valuable information and been an undisputed force for good, the ‘lumping’ approach has also been used at times to distort results, resulting in inappropriate treatment in some patient groups which has exposed them to extra risk. He gave the example of the HOPE trial with ramipril (Yusuf S et al, NEJM 2000) which led some to assume that all people with heart disease should be given an ACE inhibitor, rather than only in certain sub-groups.
Clinical trials also ‘lump’ outcomes to increase the statistical power of the trial. An example is the use of MACE (major adverse cardiovascular events) which aggregates death, non-fatal MI and non-fatal stroke into one outcome. The FAME-2 trial (De Bruyne B et al. NEJM 2012) used a primary endpoint of death, non-fatal MI and urgent revascularisation but failed to emphasise that revascularisation was the only significantly reduced variable with angioplasty (percutaneous coronary intervention, PCI) compared with medical therapy in stable coronary artery disease (Table 1). Dr Boon said that what this trial really shows is that, “If you have an angioplasty now, you won’t need one in the future” rather than angioplasty having any impact on mortality or on MI incidence.
|Events||PCI (%)||Medical therapy||Hazard ratio||P|
|Death||1 (0.2)||3 (0.7)||0.33 (0.03–3.17)||0.31|
|Myocardial infarction||15 (3.4)||14 (3.2)||1.05 (0.51–2.19)||0.89|
|Urgent revascularisation||7 (1.6)||49 (11.1)||0.13 (0.06-0.30)||<0.001|
|Primary endpoint||19 (4.3)||56 (12.7)||0.32 (0.19-0.53)||<0.001|
Table 1. Clinical events in the FAME-2 trial.
‘Lumping’ trials in meta-analyses has also become very popular. The Cholesterol Treatment Triallists (CTT) Collaboration combined 14 trials of statins in 90,000 patients with trial cohorts from very diverse populations. The outcomes suggested a significant benefit of using statins – with the overall conclusion of the meta-analysis that lowering cholesterol by 1 mmol/L with a statin prevents 48 adverse vascular events per 1000 patient years of treatment. Subsequent mass media reporting of the data suggested that all patients over the age of 50 should be given statins highlighting the dangers of elevating the importance of statistical power over clinical significance.
The lumping approach has prospered because it suits the converging interests of healthcare professionals who require simple approaches to treating patients, and the pharmaceutical industry who will have wider indications for products.
Personalised medicine: a new era of ‘splitting’
Dr Boon suggested we are now entering a new era of ‘splitting’ where patients are being treated individually with a much more personalised approach with:
- Individual risk stratification
- Focused prevention
- Targeted treatment
- Highly specific new therapies
He said that it was time for cardiology to catch up with the approaches developed in other therapeutic areas such as oncology to provide personalised medicine. Recent advances in genetics, imaging and digital health now provide the tools for clinicians to develop increased personalised care in primary care cardiology.
|Precision cardiovascular medicine enablers in primary care
Genetic analysis: will improve risk stratification. This new technology is likely to provide improved information for risk scoring for primary prevention in the near future. A recent study from the British Biobank shows that genome-wide polygenic scoring can identify 8% of the population who have more than a three-fold risk of developing coronary artery disease (Khera AV et al, Nature Genetics 2018).
Improved access to advanced imaging: will also be of great value to primary care. The SCOT-HEART study (SCOT-HEART Investigators. Lancet 2015) set out to assess the value of a CT coronary angiogram in new patients presenting with possible stable angina due to coronary heart disease. The study showed that CT angiography improved patients dramatically with improvements in diagnosis, investigation and treatment, and outcomes. Recent 5-year outcomes data from the study reveal a 50% reduced chance of dying or suffering MI from the use of a CT coronary angiogram at diagnosis (SCOT-HEART Investigators. NEJM 2018), showing the dramatic impact that an investigation can have in improving outcomes. The angiogram identifies everyone with coronary artery disease regardless of symptom status and allows them access to appropriate medical therapies. This also avoids unnecessary treatment of people with chest pain and normal coronary arteries.
Improvements in digital health: with wearable devices such as smart watches will provide improved monitoring of vital functions. The Apple Watch 4 has already been licensed by the FDA for monitoring of atrial fibrillation.
In conclusion, the new era of personalised medicine offers huge benefits for patients and healthcare professionals in primary care cardiology.