Alex Nowbar’s weekly research reviews—22 January 2019

Alex Nowbar reviews the latest research from the top medical journals

Annals of Internal Medicine

Young people with type 2 diabetes

People under the age of 40 with type 2 diabetes were found to be hospitalised more than those with onset after the age of 40. What are they in hospital for? Ke et al’s Hong Kong based cohort study reports that 36.8% of bed days in people with young-onset type 2 diabetes were hospitalisations for mental illness. This caught my attention as it is far higher than I would have expected. Whatever the mechanism, this study highlights the need for a multi-faceted approach to improve the lives of young people with diabetes with a particular focus on mental health even though the association is not well understood. The authors of this study performed elegant modelling to estimate bed days saved by intensified risk factor control. This is interesting, but I suspect only a holistic and empowering approach will make a difference.


Disease-modifying therapy for relapsing-remitting multiple sclerosis

Brown et al conducted a cohort study which supports the theory that disease-modifying therapy for relapsing-remitting multiple sclerosis might reduce the risk of conversion to the secondary progressive form. Perhaps it does reduce the risk, but observational studies like this should not be heralded as evidence of the benefits of a therapy. The issue is confounding. For example, patients who are more unwell are less likely to receive the therapy (because of contraindications, inability to attend for treatment, or a perception that the condition is too severe) and disease in more unwell patients is more likely to progress because they are further in the course of the disease, not necessarily because they did not receive the therapy. The therapies could easily have been given to many patients with mild disease that was never going to progress. Randomised controlled trials are the definitive way to assess the benefit of a therapy. The only good thing I can say about this paper is that the topic is important and interesting. The study itself has no meaning for the scientific community. Clinicians, patients and drug companies all like to buy into the idea that treating diseases earlier in their course is more effective, but we should only accept this with the right evidence behind it.


Medically assisted beginnings

Now here is a study that uses an appropriate study design to answer a research question. The question: to what extent is the higher risk of adverse birth outcomes with medically assisted reproduction due to the treatment versus parental characteristics? You can’t randomise to medically assisted reproduction versus natural conception because, by definition, relevant couples have had problems conceiving naturally! Goisis et al used Finnish administrative registers to compare children conceived by medically assisted reproduction with siblings conceived naturally. As expected there was an increased risk of adverse birth outcomes with medically assisted reproduction, but this relationship was less clear when adjusting for confounding using the sibling comparison approach. This suggests factors other than the medically assisted reproduction treatment are contributing to the effect. Good news for medically assisted reproduction, but not great news for the couples themselves as there is no obvious solution.

Fibre-day, the new five-a-day

Nutrition research is not one of my favourites. But this study is important so here I go. This extensive analysis of published studies by researchers in New Zealand showed a large mortality benefit in those with the highest fibre intake compared to the lowest. This is a classic observational finding, confounded by whatever unmeasured factor causes people to eat more fibre (which could be anything from where they live to personality traits) that could be contributing to the changes in mortality. But the crux of it is that I have to concede that it does look pretty likely (particularly because of the strong dose-response relationship) that there is something good in this dietary fibre. I suspect that this may well be exploited by diet supplement companies who might start making claims about the mortality benefits of their products even though it’s not their products that were studied.

JAMA Internal Medicine

Dastardly diagnostic lung procedures

Huo et al examined the complication rates and downstream medical costs of invasive diagnostic procedures for lung abnormalities (including bronchoscopy and thoracic surgery) in the community setting in the US and found a complication rate of more than 20%, with hefty costs ranging from $6320 to $56 845. Of course this is just one angle from which to view the issue, because we don’t know what happened to those who didn’t get diagnostics. Randomised controlled trials provide the best evidence for this. A previous study—The National Lung Screening Trial shows reduced mortality with low-dose CT screening compared to radiography. Huo et al’s study reports very different complication rates to those in this trial. On the face of it, this is difficult to reconcile and might suggest that screening is risky, but this was a different cohort of patients. I think one has to regard Huo et al’s results in isolation as an indication that investigation of lung abnormalities in an unselected cohort carries risks rather than use them to undermine findings of studies which are tackling a different question. It is sometimes good to be reminded that no diagnostics should be taken lightly, even the non-invasive ones as results of non-invasive tests can cascade into the more invasive.


Malaria muddle

This week the NEJM has simultaneously published two pharma-sponsored studies about tafenoquine for plasmodium vivax malaria, a condition which recurs due to dormant hypnozoites. The papers are two double-blind randomised controlled trials in which all patients received a 3 day course of chloroquine, one compared single-dose tafenoquine to 14 days of primaquine (GATHER) and the other compared tafenoquine to placebo or 14 days of primaquine (DETECTIVE). And something about glucose-6-phosphate dehydrogenase (G6PD) and asymptomatic drops in haemoglobin. Confused? So was I. But like the dormant hypnozoites that tafenoquine is supposed to clear, I resurfaced after a long interval of working out what these trials mean.

Some of you may recall that primaquine is bad in G6PD deficiency as it causes haemolysis. Most people do not have to worry about this, but in malaria endemic areas there is a high prevalence of G6PD deficiency and limited access to G6PD testing. Well tafenoquine is also bad in G6PD deficiency. Worse still, it has a much longer half-life than primaquine so the haemolysis persists. The only advantage of tafenoquine may be in adherence as it is taken as a single dose whereas primaquine is a 14 day course.

The DETECTIVE phase 3 trial of 522 participants assessed freedom from recurrence. Tafenoquine fell short compared to primaquine, but was still effective compared to placebo with no obvious adverse safety signal.

The GATHER trial of 251 participants was more concerned with haemolysis. It included female patients with moderate G6PD deficiency. There was no difference in the number of patients who had a haemoglobin decrease. The paper also reports a patient-level meta-analysis of patients from the DETECTIVE and GATHER trials, which concluded that tafenoquine was not non-inferior to primaquine.

You would have to be a detective to really gather any of this. For now, I say no to tafe-no-quine.

Platelet transfusions in neonates

What does a non-specialist do about low platelets? Worry a fair bit? After all, it is definitely a bad sign. Discuss it with colleagues? Discuss it with a haematologist? Check a guideline? Basic medical training teaches all of the above. And how many platelets to give? Are we told to restrict platelets just because of limited supplies of the product, or because too many platelets are bad for people? We all love a target threshold to help us deal with the anxiety of not being a haematology expert.

This study is laudable in attempting to answer the question of what to do with low platelets in neonates, as research in neonates is particularly challenging logistically and ethically. Curley et al randomised preterm infants to platelet transfusions with either a low threshold or a high threshold. Paradoxically the high threshold group had a higher rate of death or major bleeding. Practically this study should be very useful for neonatal clinical practice, but also it underlines the need for improved understanding of the platelet conundrum.


A “big” risk factor for kidney disease

This study from the chronic kidney disease prognosis consortium data coordinating center in Baltimore, USA establishes a convincing link between obesity and eGFR decline. The highest BMI category had the strongest relationship. The highest BMI was also linked to overall mortality.

The study design was an individual patient meta-analysis involving measurements from more than five million people in various general population and CKD cohorts over eight years. This impressive project will have required not only technical expertise, but also great diplomatic effort.

There are a few caveats to their findings. The association is confounded by risk factors for CKD that are linked to obesity such as diabetes. Adjusting for potential confounders weakened the association between obesity and eGFR decline. Also it may not be obesity itself contributing to eGFR but elements that lead to obesity such as poor diet or low levels of physical activity. Nevertheless, this is an important and meticulous analysis.

Disclosures of patient organisations contributing to NICE appraisals

I have often wondered about the impartiality of NICE health technology appraisals given their far-reaching consequences. But it didn’t occur to me to question the conflicts of patient organisations. Mandeville et al investigated the prevalence of patient organisations that had received funding from the manufacturer of the technology under appraisal or competitor products—a worrying three quarters. Of these, they found that NICE’s decision-making committees were aware of less than a quarter.

This study neatly linked this to NICE’s disclosure policy which did not require disclosure of most of these patient organisations’ financial interests. It is becomingly increasingly obvious to me that objective technology appraisals are almost impossible to achieve since relevant stakeholders and experts are so likely to be financially or otherwise conflicted. But here’s hoping NICE have a plan to tackle this.

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

Competing interests: None declared