Alex Nowbar reviews the latest research from the top medical journals

JAMA

Delirium in ICU
This Brazilian study compared flexible family visits (up to 12 hours a day) to an intensive care unit (ICU) with restricted visits (<4.5 hours) in a cluster randomised controlled trial for prevention of delirium. The flexible family visits intervention was coupled with an education programme for family members. Unfortunately there was no difference between the two groups in incidence of delirium, even after adjusting for duration of visits, nor any of the secondary outcomes, which included infections and the impact on staff and family members. The prevailing belief had been that if only ICUs were able to accommodate family members, then there would be lower rates of delirium. Well, now we know ICUs can accommodate family members, but it made no discernible difference. 

Post-traumatic stress disorder after ICU stay
A diary of events filled by family members and clinicians is an attractive strategy for improving psychological distress when people come out of intensive care. The idea is that patients use the diaries afterwards to help fill memory gaps. This is based on the premise that symptoms of post-traumatic stress disorder are related to memory gaps. However, in a rigorous randomised controlled trial of patients receiving mechanical ventilation in 35 French units, diaries did not reduce post-traumatic stress disorder. Perhaps digital diaries might have been better, because almost a third of the randomly sampled diaries were unreadable. Also, use of the diaries after ICU discharge was not mandatory, and presumably the impact of the diary is dependent on how much it gets looked at afterwards. The other factor limiting these data is that only half of those randomised completed follow-up at three months.

NEJM

The value of systolic versus diastolic blood pressure
Two numbers, one patient. What do you pay attention to and at what threshold? This analysis advocates tight control (that is, a 130/80 threshold) and keeping an eye on each number. Flint et al assessed the relation between stroke and myocardial infarction and systolic and diastolic blood pressures in over a million patients over a period of eight years. They report that systolic and diastolic blood pressure each independently predicted adverse outcomes. Perhaps this is statistically beyond me, but systolic and diastolic blood pressure are not independent of one another. However, there is a clear relationship between the highest blood pressures and outcomes. I am surprised death wasn’t in the combined outcome though. The second point the authors make is that this relation with outcomes was present regardless of blood pressure threshold (140/90 or 130/80). These data touch on the conflict between overtreating and undertreating hypertension in this range. Try not to be convinced either way based on these observational data as all they do is identify risk levels. Observational data can’t tell us what treatment strategy to pursue.  

HIV prevention in Africa
Usually I lament the lack of truly effective treatments. HIV is an exception. Anti-retroviral therapy is effective, and not only for the individual, but for preventing transmission to others. In areas without good healthcare provision though, the challenge is how to effectively deliver the treatment to the population. This week in NEJM there were three randomised controlled trials comparing universal HIV testing and treatment (covering Uganda, Kenya, South Africa, Zambia and Botswana) to control. For prevention of HIV they were concordantly negative so let’s just consider one of the trials, the Ya Tsie Trial.

It was conducted from 2013 to 2018. The method of randomisation was interesting. They took 30 communities in Botswana and paired them by matching geographical location, population size and other factors. Then there was randomisation within each pair. The intervention had several components including testing, links to clinical care, and early anti-retroviral therapy. While there was a lower incidence of HIV in the intervention group compared to control, the difference was not statistically significant. However, viral suppression rates in HIV-positive individuals went up more in the intervention group compared to control. The impact of the intervention compared to control may have been limited by the introduction of universal anti-retroviral therapy (as opposed to based on CD4 count) in 2016 which was therefore available in both groups in this study. Despite this community-based strategy, many people will not have been reached, and non-citizens weren’t eligible for free treatment so could have contributed to persisting incidence of HIV.

JAMA Internal Medicine

E-cigarettes to quit smoking
Smoking is one of the strongest risk factors for heart and lung disease and cancer, and yet our profession doesn’t have a clear message for patients who want to try e-cigarettes to help give up. This is a tricky area. We can say with some certainty that there are more harmful toxins per cigarette compared with an “equivalent” amount of e-cigarette inhalation, but that’s not really enough information for a behaviour with complex psychosocial context. The real question is whether e-cigarettes help overall smoking cessation or if they encourage more nicotine consumption in the population. This large French cohort study found that smokers using e-cigarettes regularly were more likely to reduce cigarette intake or stop smoking compared with those not using e-cigarettes. However, former smokers using e-cigarettes were more likely to start smoking again compared with those not using e-cigarettes. These findings are susceptible to confounding by features of people who use e-cigarettes that are also associated with cigarette use such as socioeconomic status and amount of cigarette use. Any differences are not simply explained by the use of e-cigarettes. 

Undertreated and so what?
Alore et al make some interesting claims in their analysis of primary hyperparathyroidism, that it is “underdiagnosed” and “undertreated” with surgical parathyroidectomy. They leave no holds barred in their introductory statements, making it clear that more patients should be having surgery for hyperparathyroidism even if asymptomatic for fracture prevention and cognitive benefits (although I’m not sure there is randomised controlled trial evidence to support these ambitions). They found that less than a quarter of people with chronic high calcium levels had had a parathyroid hormone level checked. This certainly begs the question about the cause of the high calcium in the rest. However, the burden of testing and subsequent interventions in this group would be high in order to bring unknown benefits to an unknown proportion of people. Just because a condition has a treatment, isn’t a good enough reason to aggressively seek it out. If symptomatic patients weren’t being tested, that would be different. But this study doesn’t capture symptoms.

There must be a reason that surgery rates have decreased over time despite rising incidence of primary hyperparathyroidism. Perhaps this is because the benefit of surgery in mild asymptomatic cases is not clear. Alore et al found that 12.8% of patients with high calcium and high parathyroid hormone levels (diagnostic of primary hyperparathyroidism) received surgery. The authors and current guidelines find this inadequate as it’s a missed opportunity for surgeries. From these data though, lack of testing and operations isn’t demonstrably wrong. These data don’t tell us if the supposed “underdiagnosis” and “undertreatment” had adverse consequences.

Annals of Internal Medicine

Anticoagulation in chronic kidney disease
Cardiac and renal physicians often see the same patients. Both groups are naturally very conscious of the risk of bleeding with anticoagulation because, if there is a risk of something, it tends to happen in a patient with chronic kidney disease (CKD). We are fortunate to have many trials of anticoagulation, but one of the limitations has been the exclusion of patients with CKD, especially at the more severe end. Ha et al’s meta-analysis combines results from 45 randomised controlled trials of anticoagulation for any indication in patients with CKD. Eight of the trials included patients with dialysis-dependent end-stage kidney disease. The main finding was a lower risk of bleeding with novel oral anticoagulants compared with vitamin K antagonists, but this was not statistically significant. 

Colorectal polypectomy and anticoagulation
A cold snare strategy for polypectomy with continuous oral anticoagulants had a 4.7% rate of major bleeding in a Japanese randomised controlled trial of 184 participants. A hot snare strategy (i.e. involving use of electrocautery) with heparin bridging had a 12% rate. Polyps could not be larger than a centimetre. The trial design was based on noninferiority and thus the trial found that the cold snare strategy was noninferior. The cold snare strategy was also associated with shorter procedure time and hospital stay. This trial is unusual in that it compared both polypectomy procedure types and anticoagulation types simultaneously. This makes it difficult to know what factor made a difference to the outcome, but it is thought that electrocautery damages the submucosal vessels leading to rebleeding after initial immediate haemostasis. The risk of thromboembolism means that clinicians are reluctant to stop anticoagulation so this result is very welcome.

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