Alex Nowbar’s weekly review—8 January 2020

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

Annals of Internal Medicine

Another arthritis flop

So much has been invested in therapeutic avenues to modify the osteoarthritis disease process, but to no avail. The cathepsin enzyme is involved in bone and cartilage breakdown, and MIV-711 is selective cathepsin K inhibitor. Conaghan et al randomised 244 patients with knee osteoarthritis in six European countries to MIV-711 or placebo in a 26 week, double-blind study. The primary outcome was change in average weekly pain severity. There was no difference between the treatment and placebo groups for this. For the secondary outcomes of bone and cartilage progression, there was a hint of promise, however. I would be interested in seeing this tested over a longer period, because I suspect the most we can hope for in this space is slowing progression, not reversing the process. As osteoarthritis progresses over years, it follows that it would take years for a difference between groups to emerge. This makes for a very challenging research pipeline especially when the long-term safety is unknown. It is hard to get excited about a moderately effective treatment that needs to be taken for a long time, but I wouldn’t dismiss this approach completely. It may be all we have. Supposed quick fixes and magic bullets, for example, as promised by stem cell therapy, don’t work.

Homage to home care?

Levine et al randomised 91 patients with selected conditions including heart failure and infections living within five miles of the hospital to either home hospital care or traditional hospital care. They used risk algorithms to exclude patients at risk of deterioration. Off the bat, that doesn’t sound like nearly enough to show a benefit for an important outcome, or to convince healthcare providers that it is safe and effective. But this trial managed to do both by using cost as the primary outcome (which is highly susceptible to bias in a unblinded study) and by presenting positive interim outcomes to hospital leadership causing the trial to be stopped early “in light of local operational needs to quickly increase home hospital capacity.” This is flawed for so many reasons. However, the real story of this study is how difficult it is to perform randomised controlled trials of a service change. Hospital home care evaluation would have been better suited to a cluster randomised trial where individual patient consent is not required and therefore everyone’s data is captured.


Evolving epidemiology of nontypeable Haemophilus influenzae

You may remember H influenzae as a bug that causes respiratory infections. There is a vaccine for subtype b but not for nontypeable H influenzae thus manipulating the epidemiological landscape. In 2018 in Atlanta, Georgia there was a sharp rise in rates of invasive nontypeable H influenzae infections in HIV-infected men who have sex with men (even in patients with well-controlled HIV). Collins et al were interested in whether strains were different between those infected with HIV and not. They assessed 553 adults with an invasive nontypeable H influenzae infection, of whom 60 had HIV and discovered two genetically distinct clonal groups of nontypeable H influenzae. These were particularly associated with septic arthritis among black men who have sex with men in geographic proximity. Genetic clusters of H influenzae are not the norm. And neither is H influenzae septic arthritis. And nor would it be expected to affect those with well-controlled HIV. Transmission of this strain of bug within a social network is the most likely explanation for these findings. These findings may only be applicable in the context of the HIV epidemic in Georgia.


PCI or bypass for left main stem coronary disease

For the unindoctrinated, the more proximal a coronary narrowing, the greater the territory of myocardium that would be affected should it get blocked. The left main stem is about as proximal and important as it gets. The NOBLE trial is an open label, non-inferiority randomised trial of percutaneous coronary intervention (PCI) versus coronary artery bypass grafting (CABG) for left main stem disease, looking at rates of death, stroke, or myocardial infarction. At five years, mortality was the same in both groups, but the PCI group had higher rates of myocardial infarction. PCI was therefore inferior to CABG for this indication. Even if you hark back to 2019 when the controversial 5-year EXCEL data indicated that PCI was non-inferior to CABG, it’s clear that CABG is better. We will never know exactly what it is about CABG that mediates the benefit but it is a more extensive revascularisation technique that maintains blood supply even when other coronary regions are diseased. Compared to PCI it is a big operation that could psychologically trigger substantial lifestyle changes such as stopping smoking but that’s pure speculation.

JAMA Internal Medicine

Opioid deaths and plant closure in the US

Venkataramani et al compared opioid mortality rates between US counties that experienced an automotive assembly plant closure and counties in which they remained open from 1999 to 2016. This innovative study used commuting zones related to the county in question on the basis that individuals may not reside in the county in which they work. These geographical units are a little crude for assessing the impact of an event because they rely on a number of assumptions, but the authors performed a number of sensible sensitivity analyses which did not change the findings. Plant closure was robustly associated with increased county-level opioid overdose mortality. Opioid overdose mortality before plant closure was similar in both counties in which plants closed and those in which plants stayed open. Five years after plant closure there was a dramatic rise in opioid overdose deaths in the relevant counties. Overdose deaths from other drugs also increased. These data highlight the role of economic downturns in opioid overdose deaths.

Device recall

Sengupta et al’s single-centre retrospective case series is a crucial piece of work. It follows the outcomes of patients with a Medtronic cardiac resynchronization therapy pacemaker at the Minneapolis Heart Institute. The device was recalled in 2015 by the FDA due to high internal battery impedence decreasing flow from the battery which could result in abrupt cessation of pacing. Any recall can have huge implications for patients, but replacing a pacemaker is not a minor undertaking. The manufacture recommendation was that “physicians should weigh the risks and benefits of device replacement on an individual patient basis” and per-patient mortality risk was estimated to be very low. These single-centre outcomes tell a very different story so I commend Sengupta et al’s paper in holding the manufacturer and the FDA to account. Serious adverse events related to the device included death, cardiac arrest, and syncope. Cessation of pacing also occurred during device telemetry. In that particular case the failure was found to be lifted bond wires on the electronic circuit—a failure that the manufacturer was aware of long before the recall. There is strong evidence that the manufacturer and the FDA may have failed in their duties with severe clinical consequences. I recommend a read of the discussion for the full details.


Alcohol abstinence for atrial fibrillation

Have you heard of holiday heart? It’s when someone has much too much fun on holiday—fun being a euphemism for excess alcohol use—and comes back in atrial fibrillation (AF). While the association is well established, this idea has not really been properly exploited therapeutically until now. In this multicentre Australian study, 140 people with paroxysmal or persistent AF who consume 15 or more units of alcohol per week were randomised to abstain or continue drinking their usual amount of alcohol. The researchers assessed freedom from AF after a two week blanking period and proportion of time in AF over six months of follow-up. Both outcomes were overwhelmingly positive: that is, in the abstinence group AF was less likely to recur, the time before recurrence was longer, and the six month AF burden was lower. This will give clinicians the impetus to recommend alcohol abstinence with far more vigour from now on.

Vaping associated lung injury update

If you’re a medical geek like me, you will have been fascinated by the US epidemic of electronic-cigarette product use-associated lung injury (EVALI). With Blount et al’s study the plot thickens. The authors collected broncho-alveolar lavage fluid from 51 US patients with the condition and analysed the contents with mass spectrometry to identify toxins. The comparator samples were from 99 healthy non-smokers (although 18 were e-cigarette users) who were part of a separate study about smoking. The key finding was the presence of vitamin E acetate in 94% of the EVALI patients and none of the healthy volunteers. These data point towards a role for this substance in the lung injury, but the absence of the substance in six of the 51 patients detracts from this conclusion. The authors suggest that perhaps their condition could have been misattributed to EVALI. While this is true, this line of thinking is not that constructive. Once you’ve selected the EVALI patient samples to analyse you are committed to analysing them as cases—you can’t change your mind about the diagnosis just because the results don’t fit a hypothesis. That being said, this condition has only been recognised recently so definitive diagnosis is still elusive.

Alex Nowbar is a clinical research fellow at Imperial College London