Editor’s view

Alzheimer’s disease

Dolphin et al. give an overview of current thinking in Alzheimer’s disease diagnosis and management. Increasingly, Alzheimer’s disease is characterised as a clinical–biological diagnosis, rather than simply a clinical syndrome, in the light of greater use of blood and imaging biomarkers, and the emergence of potential disease modifying anti-amyloid therapies. Older people should be enabled to benefit from these advances, if such benefits exist, and will need to be included in the evidence base behind them. At the same time, new approaches to investigation and therapy raise important questions about appropriateness, investigation and treatment burden, adverse effects, prioritisation of pharmaceutical over lifestyle and support services, the reconfiguration of services necessary to deliver disease modifying therapies, and the need to consider co-morbidities and context for frail older people. Effect sizes seen in disease modifying therapy trials have been below established minimal clinical important differences (that is, the clinical effect on cognition is trivial). The understandable enthusiasm about new therapies must be tempered with realism about what benefits they might bring in practice. We must ensure that the introduction of new drug therapies does not degrade services as a whole.

Ian Scott provides a critical assessment of the recent trials of monoclonal antibody disease modifying agents in Alzheimer’s disease. Cardiologist John Hampton once said that ‘the randomised controlled trial must not merely become just another marketing tool’. Pharmaceutical companies invest heavily in drug development and evaluation, and without this investment we would have no new treatments. But it also means that there is a huge vested interest in making the most of trial findings. Scott forensically analyses multiple elements of the clinical trial framework and concludes that recent ‘positive’ disease modifying therapy trials do not provide high-quality evidence of clinically meaningful impacts at an affordable cost.

Also on the theme of dementia, Guo et al. reviewed the impact on subsequent incidence of dementia of rhythm control strategies in patients with atrial fibrillation, especially catheter ablation. Fourteen trials with nearly 200,000 participants were identified. Rhythm control strategies reduced the incidence of dementia by about 26%, including both vascular and Alzheimer’s aetiologies, a valuable consideration in decision making on therapy for atrial fibrillation.

Frailty identification and response

Orkaby et al. studied the use of routine electronic health data to derive electronic frailty indices, based on the theory of deficit accumulation, drawing on practical experiences in both the USA and UK. Implementation of three different systems was analysed using the Damschroder Consolidated Framework for Implementation Research. Health services are complex and heterogeneous, so implementation of new systems required flexibility and adaptation. The evolving evidence base around both the validity of frailty measurement and its benefits in practice is important. Classifying older people by frailty status alone is probably of little benefit, and may even be harmful, and healthcare systems require the skills, capacity and resources to interpret and respond to frailty. Financial and regulatory measures are strong incentives to the introduction of routine frailty classification for older people.

In an editorial commenting on a paper by Brack et al. on diagnostic test accuracy for frailty measures, Best et al. emphasise that the ultimate proof of frailty assessment will be whether it leads to improved outcomes, evidence for which is, for the most part, still lacking. There are many issues around measuring frailty. There is no clear ‘gold standard’ for diagnosis and the hundred-or-so different published frailty scales show notoriously poor agreement with each other. The validity of the electronic frailty index (eFI) is good, but modest compared with many traditional screening tests, in particular yielding a high false-positive rate. But Best et al. argue that this misses the point of the eFI, which is around risk identification and as an epidemiological and public health tool. Future focus should be on improvement of existing tools, and how they can be best implemented to benefit health.

Moloney et al. described a consensus process for introducing frailty screening in emergency departments. It was agreed that tools should be brief, validated and administered expeditiously. The baseline should be set 2–4 weeks prior to the current illness. Functional ability, mobility, cognition, medication use and social factors were identified as the most important variables to include, although one might argue that these are dimensions of a comprehensive assessment rather than items of frailty scale per se. Uncertainties remain about feasibility and clinical and cost-effectiveness of frailty screening in emergency departments.

Falls

Falls and falls risk factors align poorly with the usual medical-model approach based on diagnoses and physiological impairments. Dormosch et al. used an automated linguistic analysis, Natural Language Professing and Dynamic Topic Modelling statistics to analyse elements recorded in free-text electronic health records to predict those most likely to fall. Twenty-five topics from a variety of domains were associated with future falls, relating to health problems, health seeking behaviours and service use, communications and family responses. These risks change over time, and can be monitored in near real-time. To date, ‘data science’ methods in medicine have focussed on simpler tasks such as image recognition, rather than risk prediction for complex geriatric syndromes. The problem of reliably identifying falls risk, and hopefully acting on the findings, has not been solved, but this study points us towards how we might make more of routine clinical records to solve practical problems.

Huang et al. reviewed the prevalence of post prandial hypotension, reporting it in about 40% of individuals across different settings in which it had been measured. Diagnostic criteria are not standardised, and there was heterogeneity with age and comorbidity.

Cancer trials

Cancer therapy has seen remarkable development in recent years driven by a model of strong translational research from cell biology to practical therapeutics, and refinement of therapies though sequential clinical trials. The mis-match between the population with the disease and the population on whom therapies are tested is striking, however. Older, frailer and co-morbid people are at greater risk of cancer, but will miss out if they are not included in trials. Hagège et al. investigated barriers to cancer trial participation. Invitation was less likely for participants with poor performance status and adverse socio-demographic features. Interviews revealed a paternalistic concern for what clinicians perceived as burden, tolerability, social support and ability to follow trial procedures. Patients who were invited but declined participation were less educated and more socially isolated. In an editorial, O’Donovan and O’Hanlon point out that templates and guidelines exist to optimise recruitment of older people in trials, and that greater integration of oncology with professionals used to working with frail older people will likely help normalise inclusion rather than exclusion from trials. Research commissioners are rightly prioritising better diversity in trial populations, but we are also realising that trying harder alone will not improve recruitment. Rather, we need better adaptation of trial protocols and recruitment pathways, unfortunately often requiring greater effort and expense.

Syncope themed collection

Age and Ageing regularly makes themed collections of previously-published papers freely-available on the journal website. We have recently added a collection on Syncope in Older Adults. The curators of the collection, Jansen and van der Velde, summarised the content in an accompanying commentary. A wide range of diagnoses underlie syncope, often treatable, but sometimes presenting non-specifically as falls. Diagnosis may not be easy, not least because of post-syncope amnesia. Investigation approaches are discussed, along with guidance on orthostatic hypotension, vasovagal and carotid sinus syndromes and cardiac causes.