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

Annals of Internal Medicine

Adverse effects of methotrexate

Adverse events happen to patients on medication all the time, and it is easy for some drugs to get an erroneous reputation for causing adverse effects. Methotrexate is an important case in point because it is used to treat diseases which themselves have a myriad of multisystem manifestations. So the only way to know which events are attributable to methotrexate is from a double-blind randomised controlled trial. Solomon et al analysed 4786 participants without rheumatic disease who were randomised to methotrexate or placebo. Active treatment was associated with moderate increases in rates of gastrointestinal, pulmonary, and haematologic events with hazard ratios of 1.91, 1.52 and 1.15 respectively. The participants were quite a specific population of people with cardiovascular disease (in particularly 81% were male), but these data still provide a valuable insight into the true rates of adverse effects of methotrexate.

JAMA

Lifting the polygenic cloud

Polygenic risk scores for coronary artery disease have turned out to have no incremental value beyond cohort equations (that is, existing risk calculators). Mosley et al assessed the predictive accuracy of a polygenic predictor including millions of common mutations in two US (white, middle aged) cohorts and found it didn’t add much. Elliott et al did essentially the same using the UK Biobank (middle aged and mostly European ancestry) and found a “modest improvement” in predictive accuracy that was statistically significant but which practically did not change most participants’ predicted probabilities of coronary artery disease. Taken together, these studies suggest that genetic information does not add much in risk prediction and thus polygenic risk scores aren’t clinically relevant in coronary artery disease risk prediction. This is hardly surprising since some of the biggest drivers of cardiovascular disease are hypertension, diabetes, and hyperlipidaemia, and there is a degree of heritability for all of these so the additional genetic impact was always going to be minimal.

Lancet

Making headway in a cure for HIV?

The RIVER trial, set at six sites in the UK, is a landmark randomised controlled trial in HIV infection. It is the first to test the effect of a “kick and kill” therapy on markers of the HIV reservoir. The existence of a reservoir of latent HIV is what has so far made it impossible to cure. You want to “kick” the latent HIV virus into expressing proteins, which can then be targets for the immune system, and provide antiretroviral drugs to “kill.” The HIV reservoir was measured by number of HIV DNA copies from peripheral blood CD4+ T cells at 16 and 18 weeks after randomisation. Unfortunately the kick and kill therapy did not demonstrate any effect of the HIV reservoir markers in this 60 patient study. This is despite evidence that the kick had an effect and the kill had an effect. They just didn’t result in the overall desired benefit. Maybe the dosing was wrong. Maybe some patients were resistant to this type of treatment. Maybe the stage of the disease in these patients was wrong—that is, these were patients with a recent diagnosis rather than chronic HIV infection, which may have quite different susceptibility to this therapy. This well designed and well executed trial may have had a disappointing outcome, but it has generated many interesting avenues to explore.

Add-on to thrombectomy for ischaemic stroke

Endovascular thrombectomy is a life-saving treatment for acute stroke due to large vessel occlusion but suddenly restoring blood flow to an ischaemic area can be problematic. This is referred to as ischaemia-reperfusion. Nerenitide is an intravenous peptide from the senior author’s Canadian biotech company, NoNO, intended to improve functional outcomes in these patients by reducing the neurotoxicity of ischaemia-reperfusion. In the ESCAPE-NA1 trial, over a thousand patients having thrombectomy were randomised to nerenitide or placebo in a double-blind fashion. Some also received the thrombolysis drug, alteplase, “as indicated” which is a bit vague. The primary endpoint was function at 90 days assessed on the modified Rankin scale which is a scale from 0 to 6 reflecting the severity of disability. There was no difference in rates of people achieving a good functional outcome between the two groups. The authors cling to hope in the fact that a positive effect of nerenitide was seen in people who did not receive alteplase. But that was not a planned analysis. And as people were not randomised to alteplase or not we cannot know if it’s alteplase that modified the treatment effect of nerenitide as opposed to confounders like time or patient characteristics. I’m pleased that the Lancet didn’t let the authors get too bowled over by this differential effect because it is not sufficient evidence of benefit.

JAMA Internal Medicine

Pre-hydrating for contrast CT in chronic kidney disease

This Dutch randomised controlled trial compared prehydration with sodium bicarbonate with no prehydration in over 500 people with stage 3 chronic kidney disease who were undergoing contrast-enhanced computed tomography. The main outcome was the change in serum creatinine at two to five days after administration of the contrast medium compared with baseline. There was no difference between the two groups, allowing Timal et al to conclude that it is safe not to prehydrate. I am not sure this is completely definitive because the prehydration fluid comparator was 250 mL of 1.4% sodium bicarbonate, and other prehydration strategies that were not assessed in this study might be effective at reducing post-contrast acute kidney injury. However, evidence that no prehydration is as good as or no worse than sodium bicarbonate prehydration, which has been a commonly used therapy, is useful. Timal et al even determined that there was no increase in healthcare expenses in the two months after randomisation.

NEJM

Value of plasma exchange in severe ANCA vasculitis

The PEXIVAS trial tested plasma exchange versus no plasma exchange and also separately compared two oral steroid regimens (standard dose versus reduced dose). It was not blinded so patients could have been differentially treated in other ways alongside their randomised treatment allocation. The authors comment that no therapies are known to reduce the risk of death or end stage kidney disease (the main outcomes assessed in this study) and that crossover rates were low, but that isn’t really enough of a defence. I might have accepted the defence that blinding is logistically challenging and thus expensive, but is that even a defence? That aside, it was large well-conducted elegant study. It found that plasma exchange did not reduce the risk of death or end stage kidney disease. And it found that a reduced dose oral steroid regime was non-inferior to the standard dose for risk of death and end stage kidney disease, but led to fewer infections which are all fascinating and clinically useful results.

A no for no sedation in mechanical ventilation

Olsen et al compared a plan of no sedation against a plan of light sedation with daily interruption in 710 mechanically ventilated patients in intensive care units (ICUs). In the non-sedation group, mortality at 90 days was 42.4% compared with 37% in the lightly sedated group. There was also no difference in the number of ICU-free days or ventilator-free days. Perhaps the lack of difference could be attributed to the lack of difference in the amount of sedation between the non-sedated and lightly sedated groups (especially since the non-sedated group often ended up crossing over to sedation (27% in the first 24 hours)). Another important limitation of this study is that the authors couldn’t recruit just any patient on mechanical ventilation. Patients had to be excluded if they were anticipated to need sedation for oxygenation or for the patient to remain in a prone position. This puts us in the quandry almost always faced in randomised trials: what if your population isn’t representative? Nevertheless, I think the take-home message is a “no” for no sedation.

Alex Nowbar is a clinical research fellow at Imperial College London

Competing interests: None declared