EBM in critically unwell patients
Evidence based medicine (EBM) was undoubtedly one of the major medical advances of the last century. EBM is at its best when it changes daily clinical practice and challenges well-established norms, but this only really happens every now and then. Two recent New England Journal papers about management of critically unwell patients have done just that.
Since medical school, doctors learn that the optimal treatment of patients with severe infection (sepsis) and low blood pressure (shock) is to pour in intravenous fluids. Not something that is up for debate, you would think. There are much more pressing things like drug-eluting stents to take to clinical trials. Well, think again. A paper by Maitland and colleagues reports a trial in Africa where over 3000 children with severe sepsis and shock were randomised to receive either boluses of fluid (albumin and saline) or no bolus in the early stages of treatment. Children with malnutrition or gastroenteritis were excluded. Amazingly, any bolus treatment led to an increase of 45% in the risk of mortality at 48 hours (relative risk 1.45; 95% CI, 1.13 to 1.86; P=0.003). The results were consistent across all subgroups of patients, and now researchers and clinicians all over the world are scratching their heads to understand whether what they have been doing since they were medical students is actually wrong and too much fluid is a bad thing in septic patients.
Intensive care, like surgery, is often cited as a difficult area of medicine for EBM to infiltrate. The traditional mantra is to give ITU patients nutritional support as soon as possible. In a trial of nearly 5000 patients, the two arms were either early initiation of intravenous (also known as parenteral) nutrition on days 1 and 2 of ITU admission, or late initiation of parenteral nutrition after day 7. Late initiation was associated with reduced complications and faster recovery. So EBM is possible in the ITU and it does change practice. What other areas of treatment of the critically unwell do we need to test?
Should we be relaxed about glucose control in diabetes?
The underlying problem in diabetes is that the blood glucose levels are too high, either due to lack of insulin production or resistance to the effect of insulin. While I was at medical school, two studies changed practice in diabetes: DIGAMI and UKPDS.
The UK Prospective Diabetes Study (UKPDS) found that intensive blood-glucose control by either oral agents or insulin reduced the risk of the so-called “microvascular” complications (i.e. kidney disease, retinal disease and neuropathy), but not macrovascular disease (heart attacks and stroke) in patients with type 2 diabetes. The study led to a focus on the microvascular complications of diabetes.
The DIGAMI (Diabetes and Insulin-Glucose Infusion in Acute Myocardial Infarction) study suggested that after heart attacks, all diabetic patients should receive intensive glucose control with an insulin/dextrose infusion, even if they were not usually on insulin. This study meant that all diabetic patients with MI were put on insulin infusions for 48 hours after their MI. The DIGAMI-2 trial and other later studies have not been so clear- cut in their results and so this practice has largely stopped.
Although a recent meta-analysis has shown that diabetes leads to a doubling in the risk of vascular disease (including MI), independently from other conventional risk factors, fasting blood glucose levels are, AT BEST, modestly associated with this risk. The authors of that meta-analysis concluded that “In people without a history of diabetes, information about fasting blood glucose concentration or impaired fasting glucose status did not significantly improve metrics of vascular disease prediction when added to information about several conventional risk factors”.
Another meta-analysis of 5 trials of intensive versus standard glucose-lowering therapies found no difference in effect on stroke or death, but a 17% reduction in non-fatal heart attacks.
In diabetic patients, the benefits of intensive glucose control do not seem to be as great as initially thought, whether in primary (before a heart attack) or secondary (after a heart attack) prevention. If you had to treat anything intensively, I would go for Intensive blood pressure control which is a more effective treatment in reducing vascular disease in these patients.
Knowledge is power but should it be free?
I spent last night eating and drinking with two old friends who happen to be web programmers. It is fair to say that we have a lot to learn from the way they share knowledge, and our evening’s conversation made me feel very uncomfortable about the whole medical research machine.
The origins of the internet lie in sharing of information among like-minded individuals, initially scientists. Much of the protocols, programming, infrastructure and software of the internet rely heavily on web programmers not to patent their ideas and to make them available to the broader public. Ideas and innovations are often introduced for free and programmers and experts evaluate, give feedback and improve ideas. Although Bill Gates and many others have hugely profited from various aspects of mushrooming of IT in the past few decades, imagine a world where we had to pay per e-mail, or pay for the right to write this very blog which you are currently reading! Open knowledge and an ethos “for the greater good” ensure this culture.
In medicine, the two major types of knowledge we share are innovations (most commonly of drugs or devices) or research studies (from basic science to epidemiology, from animal studies to patient studies). The way in which these new types of ideas are shared or disseminated is generally through publication in peer-reviewed journals, which are either subscription only, or open-access. Subscription-only journals involve a cost to readers and open-access publication involves a cost to the author or their institution. Subscription-only journals have started releasing certain articles as open-access. The cost is justified by the cost of peer review and the cost of publication.
The balance of medical research is very much tipped in favour of intellectual property rights rather than open information. Sharing of data is relatively uncommon, either because of fears of “data theft” or due to conflicts of interest, particularly in industry-funded studies. The conclusions of the research are heavily influenced by the way in which the research is funded and there is good evidence that researchers and clinicians can be influenced. Open-access datasets are a relatively new and evolving concept.
Research careers and progression are still very much judged by the publication model: “publish or perish” and often only indirectly by ability of knowledge to change practice and change norms. Although clinicians and researcher are driven by the Hippocratic oath and wanting to do good for patients, the current research model is often too investigator-driven or funding body-driven and only relatively recently have the end users (the patients) been involved in designing and prioritising research. The current global interest in the Polypill is an example of how different research institutions around the world are competing to gain the accolade of being the first to conduct the first trial of the Polypill. All the Polypills being suggested use generic drugs, but they will all be patented and promoted as new entities. Imagine if the different research groups worked together with “open Polypill” technology and the Polypills remained generic.
I was asked for examples of medical research where researchers put open access to knowledge before their own interests. Three examples came to my mind. Firstly, defibrillation as a process has never been patented. Secondly, oral rehydration therapy was originally trialled to treat cholera in Bangladesh and has been used to treat hundreds of millions of people with diarrhoea and dehydration worldwide without patent of process or the product. Thirdly, the results of the Human Genome Project are being made gradually openly available. Let me know of the other examples that must exist so that I can defend medical researchers better next time....