Access - The BMJ logo

Alex Nowbar’s journal reviews—14 May 2019

Alex Nowbar reviews the latest research from the top medical journals


Progesterone to prevent miscarriage

Underlying pregnancy there is a complex medley of hormones, so the theory that a hormone treatment would prevent miscarriage is appealing. Bleeding in early pregnancy suggests higher risk of loss of the pregnancy so it makes sense to target these women. A group of UK researchers randomised 4153 women with bleeding in early pregnancy to twice daily progesterone vaginal suppositories or placebo taken until week 16 of pregnancy. They found no difference in the live birth rate between groups. There was a suggestion of benefit in the subgroup of women with three or more previous miscarriages. But for everyone else, I think it’s case closed for this therapy—it doesn’t work.

The House of Alteplase

Here’s a title that leaves little to the imagination: Thrombolysis Guided by Perfusion Imaging up to 9 Hours after Onset of Stroke—also known as the EXTEND trial. It’s what House Alteplase have been waiting for, a way of extending the time window to be able to offer their treatment to more patients, especially those who wake up with symptoms. And why not? After all, the patients were assessed as having less neurological deficit (based on The National Institutes of Health Stroke Scale [NIHSS]) with alteplase compared to placebo in this randomised trial. And their use of imaging to identify people with brain tissue that might benefit sounds useful.

What does seem to be unclear is whether it was a double-blind trial as indicated in the original protocol, as blinding is not mentioned in the manuscript. And I was left with a few other questions. Firstly, the benefit is presented as an adjusted risk ratio for NIHSS. If a trial is randomised, why pre-specify (and only in later versions of the protocol) that the primary analysis is adjusted for baseline characteristics? The difference became non-significant when non-adjusted and for the secondary outcome about function, there was no difference between groups. Secondly, the rate of symptomatic intracranial haemorrhage was quite high compared to placebo. Thirdly, the trial was terminated early due to the publication of the WAKE-UP trial (which itself was terminated early for the unusual reason of cessation of funding). WAKE-UP studied patients with an unknown time of onset of stroke, much milder baseline symptoms and had different imaging criteria, particularly aiming for the imaging to suggest that the stroke occurred within 4.5 hours. WAKE-UP might have seen some benefit of alteplase but with the ever-present adverse safety signals. The EXTEND authors conclude that more trials will be needed for the longer time window which is surprising as they were in the middle of such a trial and yet stopped due to supposed “loss of equipoise.”

Annals of Internal Medicine

Ischaemic stroke and pulmonary embolism

A large French prospective cohort study looked for the presence of patent foramen ovale (PFO) on transthoracic echo and for presence of strokes on magnetic resonance imaging (silent or symptomatic) in 361 people who had acute pulmonary embolism. Of those with a PFO, 21.4% had recent ischaemic stroke. Of those with no PFO, 5.5% had recent ischaemic stroke. This suggests PFO are part of the mechanism of increased risk of stroke in people with pulmonary embolism.

SGLT2 inhibitors adverse event reports

Between 2013 and 2019, 55 cases of Fournier Gangrene in patients taking SGLT2 inhibitors were reported to the FDA. Fournier Gangrene is a serious life-threatening condition that can occur in people with diabetes. While it may be worth knowing that these cases happened, these data are far from a systematic analysis of people taking these drugs as reporting of adverse events to the FDA is incomplete and there is no comparator group. Even if there was an association between these drugs and Fournier Gangrene, the relationship is confounded by the indication for SGLT2 inhibitors which, until recently, was mostly for people with diabetes that was not well controlled with other medications. This was already a population who might be at higher risk of a recognised complication of diabetes. However, given that these drugs are now likely to be used in millions more people in the coming years, real-world safety data will be key.


Colorectal cancer screening and aspirin

Screening for colorectal cancer involves detecting blood in the faeces. Brenner et al randomised more than 2000 people to either aspirin or placebo to discover whether aspirin increased the ability to detect blood in the faeces, ie, to make the test more sensitive, leading to fewer missed cancer cases. The trial design is fascinating. To find enough cases of cancer to be able to detect a difference between groups, but without having to include an impractical number of people, they recruited people who were planned for a colonoscopy, although the indication for this could have been unrelated to rectal bleeding. This isn’t the same as the population who receive colorectal cancer screening, but still seems valid for evaluating the hypothesis. Of the study population, 10.5% were found to have cancer. The sensitivity of the faecal occult blood test was no higher in the aspirin group than in the placebo group. However, the aspirin given was a one-off dose of 300 mg two days before the test. This may be quite different to the long term daily aspirin which many people undergoing screening will be taking. Aspirin might increase the sensitivity of the faecal occult blood test, but if it does, the effect size was too small for this study to detect, or the dose of aspirin used in this study was insufficient. Either way, there is no evidence to support the use of aspirin for this reason at the moment.

Sunscreen screening

Matta et al applied sunscreen to 24 people (they were randomised to one of four types of sunscreen) and then tested their blood 30 times over seven days to look for the active ingredients, particularly avobenzone. For all four sunscreens, the participants’ blood contained more than 0.5 ng/mL of avobenzone at day 1. This is still tiny but the significance is unknown. Given the widespread use of sunscreen in people of all age groups, one would hope the screening process for its safety is vigorous. And it probably is. These results simply highlight that additional testing is needed determine the significance of the ingredients’ absorption into the body as it is above the threshold for the FDA waiving the need for nonclinical toxicology studies.


Alcohol forecasts

We all like a good forecasting, and not just the royal baby name variety, but forecasts of health outcomes. Manthey et al looked at alcohol intake from 1990 to 2017 and model global alcohol exposure up to 2030. Half of all people are expected to be current drinkers by 2030, up from 43% in 1990 and from 47% in 2017. Some 23% are expected to be heavy episodic drinkers by 2030, up from 18.5% in 1990 and 20% in 2017. This shines a spotlight on the growing issue of alcohol exposure and the associated harms it brings.

JAMA Internal Medicine

Phosphate binders in end stage renal disease

A study result that suggests a cheap drug might not be worse than an expensive one is not something we see every day. In people requiring dialysis for end stage renal disease there has been concern about adverse cardiovascular events of calcium based phosphate binders so the more expensive drug sevelamer (which is calcium free) has become increasingly popular.

Spoendlin et al’s cohort study compared thousands of people in the US from 2017-2018 taking calcium based phosphate binders with those taking sevelamer. They adjusted for many possible confounders that might have impacted what drug people were given and also might have impacted their cardiovascular outcome. They found no difference in cardiovascular outcomes between the two groups. Of course randomised controlled trials are needed, but they need careful thought to address the limitations of previous open-label randomised controlled trials in this area.

Prostate cancer and 5 alpha reductase inhibitors

Little known fact: 5 alpha reductase inhibitors like finasteride reduce PSA levels. The implication of this is that while treating benign prostatic hyperplasia with these drugs, prostate cancer might be missed. This US population-based cohort study was designed to assess whether these drugs were associated with worse prostate cancer outcomes. In men with prostate cancer who had been PSA-screened, those who had taken these drugs prior to their diagnosis of prostate cancer had more advanced disease at diagnosis (higher Gleason score and higher chance of having nodal or metastatic disease at presentation), worse prostate cancer related mortality and worse all-cause mortality.

Perhaps men prescribed 5 alpha reductase inhibitors were more likely to already be experiencing early symptoms of prostate cancer and it is their symptoms that drove the prescription in the first place. The authors reduced this possibility by only considering men as taking the drugs if they had started the drug at least a year before the cancer diagnosis. This study suggests that an adjustment needs to be made for PSA levels measured in those on 5 alpha reductase inhibitors. It also highlights the ongoing population-level challenge of differentiating the benign from the significant when it comes to prostates.


Apgar scores

At birth an Apgar score is used to indicate impending neonatal problems. This Swedish population-based cohort study confirms the good predictiveness of the score. Razaz et al found much higher neonatal morbidity and mortality with lower scores (less than 7 out of 10). A decreased score recorded at 10 minutes when the 5 minute score was 10 was particularly associated with an increase in neonatal morbidity. An early accurate indicator of prognosis is good to have as it might attract early interventions that can improve outcomes.

Alex Nowbar is a clinical research fellow at Imperial College London.