Alex Nowbar’s weekly review—29 November 2019

Alex Nowbar reviews the latest research from the top medical journals.

Annals of Internal Medicine

Bedside test for raised intracranial pressure

Just lie back and close your eyes. Here’s some gel on your closed eye lid so I can measure your optic nerve sheath diameter by ultrasound. Is this any good as a diagnostic test for raised intracranial pressure? Certainly from a practical perspective it’s more appealing than a CT, MRI or lumbar puncture. Koziarz et al. meta-analysed 71 studies involving over 4000 patients to determine the accuracy of this non-invasive technique in traumatic and non-traumatic brain injury settings. They found that optic nerve ultrasonography was a useful test for ruling out raised intracranial pressure when normal (nerve sheath <5mm) and when raised (>5mm) may indicate raised intracranial pressure and indicating a need for a confirmatory test.


Ubrogepant for migraine

The ACHIEVE II trial is a double blinded randomised controlled trial of two doses of a calcitonin gene related peptide antagonist called ubrogepant as an acute treatment for migraine. The participants were adults experiencing between two and eight episodes of migraine with or without aura each month. The primary endpoint was a combination of freedom from pain 2 hours after the treatment and absence of the patient’s self-selected other most bothersome symptom—a novel but laudable addition to the endpoint. While ubrogepant was more effective, still only 20% of people were pain free, which isn’t many given that 14% were pain free in the placebo group. And for the absence of the most bothersome symptom, it was only really the higher ubrogepant dose that was effective. The main missing ingredient from this study was comparison with other acute migraine treatments and in the future we will also need more evidence of long term safety in a broad population.

Umbilical cord management for the premature brain

Randomised evidence in perinatal care is limited so Katheria et al. should be congratulated on their study comparing umbilical cord milking with delayed umbilical cord clamping. They set out to assess rates of death or severe intraventricular haemorrhage in preterm infants. For some study centres, randomised was permitted with deferred consent if delivery was imminent and then parents were approached for consent afterwards. That approach doesn’t sit well with me ethically plus it introduced selection bias as parents could withdraw consent post-randomisation. However the lack of randomised evidence in this area does not sit well either so I will have to hold my tongue on this matter. The trial was stopped early because at the first interim analysis, umbilical cord milking was associated with much higher rates of severe intraventricular haemorrhage (although without a signal of higher death rates). Unfortunately, as noted by the authors, definitive conclusions from this study are limited because it was stopped early. Umbilical cord milking is pushing placental blood down the cord to the newborn before clamping the umbilical cord. Perhaps this is too much for the underdeveloped newborn. For example, the authors note that that premature infants lack adequate cerebral autoregulation. They explain that improved blood flow in a fragile brain may do more harm than good.


Anti-CD38 for multiple myeloma

Ah yes, cluster of differentiation 38. As we all know, CD38 has an important role in white blood cell function. Isatuximab is a monoclonal antibody that binds to CD38 thus targeting tumour cells for programmed cell death in myeloma. The ICARIA-MM study tested its impact on top of pomalidomide and low-dose dexamethasone in an open-label randomised controlled trial of 307 people with refractory multiple myeloma. In the control group median progression-free survival was 6.5 months, but with the addition of isatuximab, progression-free survival was 11.5 months. In this study, progression-free survival was adjudicated by an independent committee, but there was a high risk of bias. There were tests for myeloma protein, imaging assessments, and sometime bone marrow aspiration. The latter two of these seemed to be at the discretion of unblinded investigators. So their beliefs about the treatment under investigation could have affected their assessment. The trial would have provided more useful evidence had it been designed and powered for overall survival. Alternatively it could have been fully blinded. As it stands, it is really unclear whether or not it truly benefits patients.

JAMA Internal Medicine

Increasing HIV tests with more HIV tests

Enhanced HIV awareness is key to early diagnosis, treatment and transmission prevention. There have been a number of campaigns around this but few are backed by randomised controlled trials. MacGowan et al. randomised 2665 American men who have sex with men to a group who were sent 4 HIV self-tests, and a group who were not. They then measured HIV testing frequency and the number of newly identified HIV infections in participants and participants’ social network members. 77% of those who were sent the self-tests reported testing themselves 3 or more times in a year compared with only 22% of those in the control group. The incidence of new infections was twice as high in the group who were sent the self-tests. This is in the context that all study participants were given access to online resources and telephone counselling and that participants were recruited via the internet using online banner advertisements. This strategy appears to have potential to prevent HIV transmission in men who have sex with men. 

Inpatient cardiac care in the homeless

It is obvious that homeless people get worse healthcare than people who are not homeless, but unless you quantify a problem, people rarely take any notice. Wadhera et al. compared intensity of care and mortality for cardiac conditions like myocardial infarction and heart failure in thousands of homeless and nonhomeless people in New York, Massachusetts and Florida. What is intensity of care though? Is it a good measure of quality of care? Probably not, but it is still worth understanding why there was a disparity in it between the homeless and nonhomeless. Rates of angiography, stenting, and coronary artery bypass grafting were dramatically lower in the homeless even after adjusting for confounders like comorbidities and insurance payer. Why should that be? One explanation is discrimination against the homeless by clinicians which would be quite concerning, but another is the lack of support for the challenges in treating homeless people e.g. lack of primary care facilities. What’s worse is that mortality rates were significantly higher in the homeless. For example in STEMI, where it makes sense that less intervention leads to worse outcomes, mortality rates were 8.3% in the homeless and 6.2% in the nonhomeless (p<0.04).


Ticagrelor after stenting

For a cardiologist, the most feared complication of a stent is stent thrombosis. Aside from the technical aspects of performing the procedure, dual antiplatelet therapy for 1 year has been the main approach to preventing this. In recent years, there has been a lot of work into whether dual antiplatelets can be switched to a single antiplatelet agent sooner. This isn’t just a pill numbers game. This is about reducing bleeding risk. Mehran et al. randomised 7119 people who had had a stent 3 months ago to either ticagrelor and aspirin or ticagrelor only in a double blinded fashion. The primary endpoint (bleeding) was much less common in the ticagrelor only group with no difference in rates of death, myocardial infarction of stroke. I have no criticisms of this trial other than that it tested the wrong single antiplatelet agent. Why not use aspirin or if it has to be P2Y12 inhibitor then why not cheap clopidogrel? 

Cholesterol targets after stroke

The Treat to Target Stroke trial is one of those studies where the answer seems so obvious (surely the lower the better?), people might have said why do we need a trial. Amarenco et al. compared two different LDL-cholesterol targets (less than 1.8 versus 2.3 to 2.8) in a randomised controlled trial of 2860 people who had had a stroke or TIA. But we did need this study. There has been a fear that statins would increase the rate of diabetes and intracranial haemorrhage. Also there is a question about whether to blanket treat with high dose statin or whether to aim for a particular target. This study found a lower rate of death, myocardial infarction and stroke in the lower cholesterol target group with no concern about diabetes or intracranial haemorrhage. However, what this trial does not do is tell us whether cholesterol reduction is what mediates the benefit or is it other effects of the statin and other drugs used. 

Alex Nowbar is a clinical research fellow at Imperial College London