Alex Nowbar reviews the latest research from the top medical journals

The importance of TIAs

Are transient ischaemic attacks (TIAs) a warning shot for impending stroke? Lioutas and colleagues examined the incidence of stroke in people who had had a TIA compared with age and sex matched controls who had not had a TIA in a large retrospective cohort study (the Framingham Heart Study). They found that almost 30% of those who had a TIA went on to have a stroke. A fifth of the strokes happened within seven days of the TIA, and two fifths of the strokes had happened by 90 days. Half the strokes happened more than a year later. Matched control participants who had not had a TIA had much lower stroke rates. Of course, this tells us to pay attention when people have a TIA, as it is potentially an opportunity to reduce the risk of a stroke. The close temporal relation between the TIAs and strokes suggests we should consider TIAs a pre-stroke phenomenon. These data are helpful for being able to quantify risk, with the caveat that our patients may be quite different to the Framingham population.

JAMA doi:10.1001/jama.2020.25071

Still active

Older adults do quite a lot even when they have dementia, disability, or depression—according to this survey of US adults aged over 65 years living in the community and participating in the National Health and Aging Trends Study between 2011 and 2015. The researchers focused on “meaningful activity”—defined as “self-reported participation in a favorite activity that enhanced cognitive engagement (eg, reading), social connectedness (eg, socializing with others), or physical aptitude (eg, walking/jogging).” They also considered the inability to do a favourite activity, lack of a favorite activity, and passive activities such as watching television as lack of meaningful activity. In all, 84% of people without dementia, disability, or depression did meaningful activity compared with 74% of those with dementia alone, 56% of those with disability alone, and 68% of those with depression alone. Rates of meaningful activity were much lower if participants had more than one of these conditions. However, the assessment was based on self reporting, which could be biased. Also, I imagine the results would be quite different if the survey had included people in care homes. In the whole survey group the three favourite activities were walking/jogging, outdoor maintenance, and reading.

JAMA Intern Med doi:10.1001/jamainternmed.2020.7492

Aspirin to reduce pregnancy loss

We rely on the intention-to-treat analysis of randomised controlled trials because per-protocol analysis can negate the bias-resistance that randomisation provides. Yet it’s very tempting to rely on the effect seen in those who actually took the treatment. Naimi and colleagues present a per-protocol analysis of a previously published study of aspirin initiated before conception for improving pregnancy outcomes. This study did not show a positive effect, and that was attributed to non-adherence. In this post hoc analysis, taking aspirin more than four days a week improved pregnancy outcomes. The authors rightly acknowledge that “trial data for this study were analysed as observational data, thus are subject to the limitations of prospective observational studies.” The positive effects reported here are therefore food for thought but insufficient for recommending aspirin for people with one or two pregnancy losses who are attempting pregnancy.

Ann Intern Med doi:10.7326/M20-0469

Morbid but important

When life sustaining measures are withdrawn, it’s important to know when death (that is, cessation of cardiac activity) has occurred for the purpose of organ donation. Dhanani and colleagues systematically studied 631 adults in three countries, but primarily Canada, in whom withdrawal of life sustaining measures without cardiopulmonary resuscitation and imminent death was anticipated. Monitoring included blood pressure, electrocardiography, and oxygen saturations for 15 minutes after withdrawal and for 30 minutes after death. Retrospective analysis identified 67 (14%) instances of one or more cardiac cycles. Only 1% of patients were recognised to have this at the time. The latest time these occurred after pulselessness was within 4 minutes and 20 seconds. Is this a big enough cohort to trust that 5 minutes is long enough? Possibly, but there were limitations—such as large numbers of excluded patients due to incomplete data for waveform analysis, and patients who proceeded to organ recovery so the observation ended at 5 minutes. Also, the returns of cardiac activity may not have represented meaningful circulation.

N Engl J Med doi:10.1056/NEJMoa2022713

Useful coil trial

Turok and colleagues randomised women to receive either the copper intrauterine device (IUD) or a levonorgestrel IUD (which provides local progestogen hormone) for emergency contraception after unprotected sex in the preceding five days and who agreed to implantation. If only 2.5% more pregnancies occurred in the levonorgestrel IUD group, it would be considered non-inferior. This sounds pretty generous to me given the nature of the outcome being the presence of an unwanted pregnancy. Fortunately, the difference between the two groups was 0.3%, so it does seem that the levonorgestrel IUD is non-inferior to the traditional copper coil. This is good news because the former is better tolerated than the copper coil as a form of long-term contraception.

N Engl J Med doi:10.1056/NEJMoa2022141

Alex Nowbar is a clinical research fellow at Imperial College London