Alex Nowbar reviews the latest research from the top medical journals
Selepressin for septic shock
SEPSIS-ACT was a double-blind randomised trial seeking an optimal dose of vasopressin V1a agonist, selepressin, to improve outcomes in patients with sepsis requiring noradrenaline. Selepressin did not live up to expectations, with no reduction in deaths or days on ventilation or vasopressors. Previous trials had suggested it would help with blood pressure and fluid balance without the adverse effects of catecholamine-based vasopressors, but that wasn’t the case. One interesting explanation posed by the authors was that the drug was started when patients had already been on noradrenaline for a median of 8 hours beforehand. This could have limited the ability to observe the benefits of selepressin as an alternative vasopressor. I would look to this trial for its smart features in trial design and statistical analysis such as response-adaptive randomisation and Bayesian inference model to guide interim decision-making.
Vitamin C infusions for sepsis
Intravenous vitamin C to reduce organ failure in patients with sepsis and acute respiratory failure? Whatever next? In this trial 167 people were randomised to vitamin C or placebo infusion every six hours for a period of 96 hours. There was no evidence of benefit based on the change in modified sequential organ failure assessment (SOFA) score, nor in the inflammatory biomarkers (C reactive protein and thrombomodulin). Surprisingly, given the lack of benefit in almost all the secondary outcome measures, death rates were lower in the vitamin C group. This difference was striking, but could have been due to chance. It is difficult to know what to make of this. It is mysterious that death rates were reduced without improving markers of organ dysfunction.
Annals of Internal Medicine
The meat trials
Not quite as catchy as “The Hunger Games,” but just as intriguing, a recent analysis of meat trials has turned previous recommendations on their head. Zeraatkar et al’s review identified 12 trials that randomised participants to varying quantities of unprocessed red meat or processed meat and assessed health outcomes. Most were small studies lasting a year or less and at high risk of bias. The one study that had the capacity to usefully answer the question had over 40 000 participants and 12 years of follow-up, but found no mortality benefit with the lower red/processed meat diet. Common sense still dictates that burgers aren’t the healthiest choice, but this will put many minds at rest while baffling others. Diets are so complex that we may always have to take dietary research with a pinch of salt.
It’s not really news that continuous skin-to-skin contact and exclusive breastfeeding (“kangaroo care”) is good for neonatal health, but Mazumder et al take this concept further in a large randomised controlled trial of infants with low birth weight set in India. The study was designed to compare death rates in the first 28 days and in the first 180 days. The babies randomised to kangaroo care were visited at home intensively (9 times) in that month to support kangaroo care. There was a significant reduction in death rates compared with routine care. There were also better growth rates, breastfeeding rates, and dramatic reductions in infection rates with kangaroo care. The researchers should be commended for their tenacity in conducting this trial to demonstrate the benefits of a key lifesaving intervention in neonatal care in a resource-limited setting.
JAMA Internal Medicine
Intervention for dementia dyads
Possin et al performed a randomised controlled trial of a collaborative care delivery system called the Care Ecosystem compared to usual care for people with dementia. They randomised 780 dyads, by which they mean the units of one person with dementia and one caregiver. The Care Ecosystem involved telephone-based education, support and care coordination (advanced nurse, social worker and pharmacist). Usual care involved contact information for Alzheimer’s support groups and quarterly newsletters. The Care Ecosystem model performed well in improving the person with dementia’s quality of life, reducing emergency department visits and improving caregiver depression and burden. Dyads were just as likely to get admitted to hospital though, suggesting that merely the “unnecessary visits” got avoided rather than actually changing underlying disease processes. The study follow-up was 12 months which is probably long enough to demonstrate a worthwhile benefit, but it would be useful to know whether the effects persist and whether the model would need to be delivered indefinitely.
Which temperature after arrest
French researchers provide some answers to a longstanding question in people resuscitated after cardiac arrest. Their trial randomised 584 people who had been resuscitated from cardiac arrest with a non-shockable rhythm to either therapeutic hypothermia of 33° or targeted normothermia of 37°. The hypothermia arm involved active cooling. More patients had a favourable neurological outcome (that is, a score of 1 or 2 on the 1-5 cerebral performance category scale, assessed by phone call at 3 months) in the hypothermia group compared with the normothermia group (10.2% v 5.7%). The findings are limited by lack of blinding. Arguably, maintenance of blinding when the treatment is administered over a 24 hour period would have been challenging. While this is reasonable evidence that hypothermia is protective for the brain, knowledge of treatment arm could have had an impact on the result. Could we build better infrastructure in hospitals (electronically and environmentally) to foster blinding in research studies so these questions can be answered with more certainty?
Nul points for ticagrelor
The THEMIS trial was a double-blind randomised controlled trial of aspirin versus aspirin and ticagrelor for patients with stable coronary artery disease and diabetes. The participants had to have no previous history of a stroke or myocardial infarction. The fact that this 19 thousand patient study “met the primary efficacy endpoint” isn’t as exciting or game-changing as one might hope though. Who is interested in reducing their risk of a cardiac event from 8.5% to 7.7% with the burden of not only extra tablets, but also an increased risk of bleeding? It doesn’t matter how small the increased bleeding risk is (up from 1% to 2.2%), I cannot imagine many patients being impressed by being told to take more tablets. Discontinuation rates were high in both groups, but more so in the aspirin and ticagrelor group. I am impressed though with the quality of the design, conduct and reporting of this enormous international study.
Stents or bypass for left main stem coronary disease
Cardiologists and cardiothoracic surgeons are both known for their devil-may-care attitude to the volume of procedures that come their way… not. A bias-resistant trial is therefore more important than ever to compare outcomes of stents or bypass. The EXCEL trial was a carefully designed randomised controlled trial of 1905 people now reporting the rate of death, stroke, or myocardial infarction at five years. The EXCEL investigators had already shown non-inferiority of stenting at three years. This seemed to also be true at five years, that is if you’re not bothered by the following features of the stent group compared to bypass: (1) a 3% absolute increase in death rate, (2) a 3% increase in myocardial infarction (excluding periprocedural myocardial infarction because that is of dubious significance), (3) a 7% increase in rate of needing another revascularisation procedure and (4) unreported bleeding rates on dual antiplatelet therapy with stents (whereas the bypass group mostly received only aspirin). I would interpret these data as confirming bypass as the preferred option. In practice, these data will be very useful for patients to make informed decisions and undoubtedly there is a role for stents. As with many high-stakes studies, the paper presents the results as a success for the product manufactured by the study sponsor (stents) when actually in several ways it was highly disappointing.
An addition to the antibody armamentarium
My work isn’t done until I’ve mentioned a monoclonal antibody. This time it’s for the rare disease, chronic spontaneous urticaria, and the mab is ligelizumab which is described as “a next-generation high-affinity humanized monoclonal anti-IgE antibody.” Wowzer, that’s got to be good, right? And it was, at least at 12 weeks in this double blind, dose finding, randomised controlled trial. Of the people in the 72 mg ligelizumab dose group, 51% had complete control of hives (weekly hives-severity score of 0), compared with only 26% in the active comparator (anti-IgE mab omalizumab) group and 0% in the placebo group. Quite impressive. The German researchers found higher rates of adverse events in the ligelizumab groups, particularly injection site reactions. Overall, not too bad, especially since the injections are only once every four weeks.
Alex Nowbar is a clinical research fellow at Imperial College London