Alex Nowbar’s research reviews—5 February 2019

Alex Nowbar reviews the latest research from the top medical journals

Annals of Internal Medicine

Messy scores

MEESSI is a score developed in Spain for prognosticating acute heart failure in the emergency department. Prognosis is something we think we are good at, even though studies show we aren’t, but we are still sure we are. This Swiss study successfully validates the score in their cohort. I am glad it was valid and believe the score probably is highly discriminatory at predicting 30 day mortality. Nice work. Whether it will be useful to apply it in practice remains to be seen. It depends what decisions would change as a result of knowing a patient’s risk of 30 day mortality. It could be used to deny some patients more aggressive therapy and that puts any score in a messy area.


Probable dementia and blood pressure

Researchers in the US performed a randomised controlled trial of patients with hypertension. They compared blood pressure control with a target of less than 120 mmHg versus a target of less than 140 mmHg and looked for a reduction in “probable dementia” cases. This was a large (9361 people) and elegant trial. It was stopped early because intensive blood pressure control was better for cardiovascular outcomes and all cause mortality. The difference in probable dementia rates between the two blood pressure control groups was not statistically significant. So, we do not know if blood pressure lowering reduces probable dementia, but we know it is good. We knew this before, but are a bit reluctant to act in practice because lower blood pressures might also be bad and it’s hard work. Also it requires more medication which, in real-world practice, isn’t necessarily good or even consumed.

On the subject of  “probable dementia”, clinically, one can put the word “probable” in front of anything and look clever, but what does it actually mean when used as an outcome in research? The protocol mandated participants to have cognitive assessments. So my best guess is that probable dementia is a way of referring to the detection of dementia by screening everyone, which isn’t how dementia is usually detected in clinical practice. However, it was a way for the researchers to standardise the definition of a case of dementia. Probably.


Hep C hopes

Heffernan et al’s study predicts the global hepatitis C epidemic by 2030 and models the impact of public health interventions on this. They estimate 640000 deaths from cirrhosis and liver cancer would be prevented by offering direct acting antivirals at the time of diagnosis and 1.5 million deaths would be prevented by offering a comprehensive package of prevention, screening, and treatment interventions. They estimate 15.1 million new infections would be prevented by the comprehensive package. This is a salient paper to increase focus on delivering public health interventions to tackle this condition globally.

Drugs for anxiety

Slee et al analysed 89 randomised placebo controlled trials of drugs for anxiety, providing comparative data for the different drugs. All the pharmacological agents appeared quite effective. The most effective, and relatively acceptable in terms of side effects, were duloxetine, pregabalin, venlafaxine, and escitalopram. Mirtazapine, sertraline, fluoxetine, buspirone, and agomelatine were also effective, but the study sizes were smaller. Benzodiazepine and quetiapine were less well tolerated, but still effective. Admirably, they included Chinese trials (that weren’t written in English). The results were not much affected by the Chinese trials, but did provide data for some drugs that weren’t assessed in any other study.


The world needed another study on statins. No really. It did. There is a prevailing belief that statins are not beneficial above a certain age. This can lead to underprescribing or deprescribing of statins. This meta-analysis compares the effects of statins at different ages. 8% of the patients in the 28 included trials were over 75. The study underlines the benefit of statins in patients of all ages to prevent vascular events and differs from previous meta-analyses by using individual patient data. Patients who participate in trials will, by their nature, rarely represent the population, but this is the best we have to go by.

JAMA Internal Medicine

Anti-antibiotics for asthma

Almost 20000 patients hospitalised with asthma exacerbations and given systemic corticosteroids were studied in this American cohort study. The authors compared patients who were prescribed antibiotics with those who weren’t. Those prescribed antibiotics had longer hospital stays and higher hospital costs. “Treatment failure” rates were no different. These observational data are interesting but provide no information about causality because of confounding by factors that contribute to antibiotic prescribing such as age and co-morbidities, which also affect length of stay and costs. It is reassuring that patients weren’t worse off without antibiotics though.


Cardiac Bypass Grafts

This trial is boss. And I don’t just mean because the authors are boss (including the late Doug Altman). I mean it’s impressive to randomise more than 3000 patients and follow them up for 10 years, especially when randomising to bilateral internal thoracic artery grafting versus single internal thoracic artery grafting. It shows what is possible in clinical trials with the right stakeholder support and infrastructure. The trial found no difference in mortality or the composite outcome of death, myocardial infarction, or stroke between the two groups.


Giving antibiotics directly into the veins has got to be better than into the gut—or so we are told. Well, it turns out oral antibiotics are just as good, as shown in the POET randomised controlled trial in patients with left sided endocarditis in Denmark. Both groups received 10 days of intravenous antibiotics and infectious diseases experts chose the antibiotic regimes. Adherence to therapy was high and may not be matched in practice, which could reduce the applicability of this study, but it is still very promising. The findings may also not be as applicable to antibiotic-resistant organisms. The patients were carefully selected to be “clinically stable,” but it would be possible to implement such a protocol in clinical practice to determine suitability for oral antibiotics.

Bone and joint infections

The OVIVA trial also looked at oral antibiotics instead of intravenous antibiotics, this time for complex orthopaedic infections. This Oxford-led trial is remarkable for its inclusion of patients with protheses; 639 of the 1054 participants had metalware-related infection. Patients were randomised to oral or intravenous antibiotics for 6 weeks and the oral option was found to be non-inferior in terms of treatment failure at one year. Interestingly the oral option was associated with shorter hospital stay and fewer complications, but in an unblinded trial, an outcome like length of stay is susceptible to bias.

E-cigarette use for stopping smoking

Hajek et al randomised 886 people who attended stop smoking services to nicotine-replacement for 3 months or a refillable e-cigarette with one bottle of nicotine e-liquid and advice to buy further e-liquid. Both groups also received behavioural support. At one year, 18% were abstinent in the e-cigarette group versus 9.9% in the nicotine-replacement group. There were higher rates of throat or mouth irritation with e-cigarettes and higher rates of nausea with nicotine-replacement. These are good data to support e-cigarette use for stopping smoking in the context of a stop smoking service with behavioural support. I would be interested to know the rates of nicotine addiction beyond one year in both groups, but it is still a practically useful study in its current form.



Sievert et al performed a meta-analysis of randomised controlled trials which compared breakfast consumption with no breakfast consumption in adults in high income countries. The aim was to work out which strategy is better for weight loss and reducing energy intake. These data debunk the myth that skipping breakfast puts on weight. This myth has always been a bit counterintuitive. All else being equal, surely an additional meal means additional calories which equates to weight gain? Weight loss is an inimitable challenge that is unlikely to be solved simplistically with or without breakfast.

By claiming that skipping breakfast isn’t as bad as we thought, unmeasured potential breakfast benefits are disregarded, such as better concentration. The study was about weight loss so this is no criticism of the study. But the media do get carried away with this sort of thing with no care for the impact.

Scribes and sundry

A randomised controlled trial of scribes to increase doctor productivity (patients per doctor per hour) was bound to strike a chord. A scribe is a person who does documentation. The role of scribe reported in this Australian study of consultants and senior trainees in the emergency department is much more than that; to do clerical tasks otherwise done by the physician including tasks such as locating nurses’ notes and faxing requests.

These researchers report that scribes increased productivity in the emergency department. But did they randomise the 88 doctors to either scribe or no scribe? No. They randomised thousands of those doctors’ shifts to scribe or no scribe. The issue here is bias, as a doctor doing a shift with a scribe might subconsciously be more productive when they have the scribe and subconsciously less so without the scribe. This could have been avoided by only including each doctor once, but this would have been impractical and the trial would have never been completed. Also, if the scribe was sick the shift was counted in the non-scribed group despite the shift having been randomised to scribe. I appreciate these design features were felt necessary to ensure completion of the trial within time and budget constraint, but they make the findings unconvincing.

Perhaps it was cost-effective to train 12 scribes with 30-40 hours of pre-clinical training, two classroom days including emergency department orientation, lectures, and patient simulation and feedback, and a clinical apprenticeship with a physician trainer until competent (range 3-16 shifts). Much as scribes may be useful, it does seem a bit topsy-turvy to employ a new staff member to deal with excess clerical tasks rather than reducing the excess clerical tasks by streamlining referral pathways and improving IT systems.

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

Competing interests: None declared