Alex Nowbar reviews the latest research from the top medical journals

Annals of Internal Medicine
Blood cultures before antibiotics
Researchers analysed the blood culture results of 325 patients with severe sepsis who had their blood taken before and after receiving antibiotics. Blood cultures taken before antibiotics were positive for one or more microbial pathogens in 31% of patients, whereas blood cultures taken up to two hours after antibiotics were positive in only 19% of those same patients. This study confirms the significant reduction in sensitivity of blood cultures if taken after antibiotics are started. The two things we think we know about sepsis are that outcomes are better the sooner antibiotics are started, and that blood cultures should be taken before antibiotics are started to optimise the subsequent antibiotic therapy for the identified pathogen. Unfortunately, these two things can be in conflict: the logistics of taking blood cultures might delay the administration of antibiotics. However, this study suggests that getting blood cultures first is important for pathogen identification. 

Aspirin for primary prevention
Researchers in New Zealand estimated the benefits and harms of aspirin in almost a quarter of a million people in primary care aged 30 to 79. They found that 2.5% of women and 12.1% of men would get net benefit from aspirin (i.e. avoidance of cardiovascular events, as well as major bleeding events if you take the events 1 for 1). However, if preventing one cardiovascular event was worth two major bleeding events, 21% of women and 41% of men would get net benefit from aspirin. This study is certainly food for thought but the events are based on risk scores using the PREDICT equations, which were extensively validated in a previous study, not randomised controlled trials. Secondly, the results may not be applicable as the sample were all in New Zealand and people were excluded if they were over 79 years old or if they had chronic kidney disease. While we cherish the idea of personalised medicine, i.e. aspirin only for those who will derive net benefit, we’re not there yet. 

JAMA
Linagliptin for diabetes
The CAROLINA trial randomised more than 6000 people with “early” type 2 diabetes and high cardiovascular risk to either linagliptin or glimepiride in a double blind fashion and assessed cardiovascular events over a median of six years. Rates of cardiovascular events were similar in both groups. This is hardly surprising given that all patients had intensification of glycaemic therapy as needed. Linagliptin was non-inferior but not superior for cardiovascular prevention over glimepiride, yet it was associated with fewer hypoglycaemic episodes, which will be a welcome finding. Rates of symptomatic hypoglycaemia were about a third in the glimepiride group but 10% in the linagliptin group. Rate of serious adverse events such as pancreatitis were otherwise similar in both groups. This is the first proper comprehensive evidence for preferential use of linagliptin. 

Metformin and reduced kidney function
Roumie et al compared metformin with a sulphonylurea in a retrospective cohort study of 50 000 US veterans who had diabetes and reduced kidney function (eGFR <60) but continued treatment. The primary endpoint was major adverse cardiovascular events, and metformin came out with significantly lower event rates. The time when declining kidney function triggers cessation of metformin is a bewildering time for patients and clinicians. This study is not that useful to those of us who persist with metformin until eGFR < 30. I am also surprised that renal outcomes, such as progression to end stage renal failure and rates of dialysis, weren’t reported given the purpose of this study to establish whether it is a good idea to use metformin at lower kidney functions.

Lancet
Dual diabetes therapy
The VERIFY trial looked at two types of starter therapy for type 2 diabetes over five years in a double blinded randomised controlled trial. Two thousand treatment naïve people diagnosed with type 2 diabetes in the last two years with an HbA1c between 48 and 58 were randomised to either metformin and placebo or metformin and vildagliptin and followed up for HbA1c checks. And would you believe it? Those in the dual therapy group were less likely to have HbA1c levels greater than 53 down the line. The only reason I can even bear to convey these pedestrian findings to you is that they at least demonstrated the relative safety of the dual therapy approach. This study implies that use of Novartis’ metformin/vildagliptin combination early after diagnosis of diabetes has benefits to patients, but with only an HbA1c endpoint have they really shown that? Does it really make a difference to patients and, if so, at what cost? Nevertheless, credit where credit is due, their academic buzzword chatbot is on top form: “The results reflect an enhanced understanding of the pathophysiological mechanisms underlying the progressive nature of newly diagnosed type 2 diabetes, and the expanded therapeutic armamentarium and strategy in optimisation of early diabetes management.”

Chronic rhinosinusitis with nasal polyps
Bachert et al tested the monoclonal antibody dupilumab in two international randomised controlled trials including more than 700 participants with severe chronic rhinosinusitis and nasal polyps. The drug was a resounding success, improving all aspects of the co-primary endpoint, which included patient assessed symptoms and radiographic obstruction. It is impressive that the treatment showed efficacy for both subjective and objective measures. For this population of patients, who are ridden with symptoms, this will be a welcome development with few downsides; adverse events were more common in the placebo group, but the treatment does mean an injection every two weeks. Dupilumab is not a cure, but it is an important step forward in treatment as an add-on to intranasal steroids. 

Potassium binder with spironolactone in chronic kidney disease
The AMBER trial addressed an important issue: not being able to give spironolactone to hypertensive patients with high potassium due to chronic kidney disease. Patiromer is a potassium-binder that is already licensed for treating hyperkalaemia. This international trial assessed whether it could be used to increase the ability to use spironolactone in patients with resistant hypertension and chronic kidney disease (eGFR 25 to 45) with potassium levels between 4.3 and 5.1mmol/L (so already quite a restricted population). Two hundred and ninety five patients were randomised to patiromer or placebo in a double blind fashion and all patients were started on 25mg of spironolactone. The primary endpoint was the proportion of patients on spironolactone, but you could argue that a more relevant primary endpoint would be blood pressure reduction. Anyway, at week 12, 66% of those in the placebo group remained on spironolactone compared with 86% of those in the patiromer group. While these data do support the use of patiromer, if it was okay for 66% of patients to take spironolactone without it, maybe that is what should be explored.

JAMA Internal Medicine
Survey of patients’ views on glycaemic targets
US researchers assessed the perceptions of 818 people over 65 years old about adding and removing diabetes medications from treatment plans. The risk of adverse events was rated as the most important factor when it came to deciding to add or stop medication. The authors highlight that participants thought diabetes treatment should be more aggressive in those who have more comorbidities or who have had diabetes for longer, which is in contrast to guideline recommendations. This discrepancy between patient beliefs and guidelines is fascinating, but I take surveys with a pinch of salt. In this one, people were being asked what was appropriate for a hypothetical patient, which is quite different from being asked what one would want for oneself. Individual answers would probably vary with personal experience of the condition as well as after education. What is clear is that there may be a substantial gap to bridge between patient and clinician.

NEJM
Dapagliflozin to infinity and beyond?
Confirming what cardiologists and SGLT2 makers had hoped, the DAPA-HF trial found a reduction in “heart failure worsening” and cardiovascular death with SGLT2 inhibitor dapagliflozin, regardless of whether the person had diabetes or not. This randomised controlled trial enrolled people with symptoms of heart failure and an ejection fraction of 40% or less. Heart failure worsening was defined as an unplanned hospital admission or urgent intravenous therapy for heart failure. Positive effects were even seen for all cause mortality. I am almost convinced to start prescribing dapagliflozin. My only misgiving is the lack of blinding. Is it really justifiable to be doing unblinded research with the potential to impact millions of patients? You may say, okay yes, the need for or timing of “urgent IV therapy” can be fiddled, but what about the “hard” endpoints—there are no two ways about death. But it’s not as simple as that. If patients and clinicians know the treatment allocation, they have the ability to alter their behaviour in a biased way, albeit subconsciously. For example, other medications or interventions might be more or less likely to be modified. Both randomisation and blinding are essential to provide evidence that it is the investigated treatment producing the desired effect, and not other factors.

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