Richard Lehman’s journal review—6 February 2017

richard_lehmanNEJM  2 Feb 2017  Vol 376

Adding bicalutamide to RT for recurrent prostate cancer

So far, this year is proving a good one for urology studies. Here are the follow-up results of a trial which was designed in the 1990s to compare two treatments for locally recurrent prostate cancer following radical prostatectomy.

The signal to start treatment was a return of detectable PSA (0.2-4.0 ng/ml) and all the participants received local radiotherapy, while half were randomised to receive the anti-androgen bicalutamide at high dose for two years. The actuarial rate of overall survival at 12 years was 76.3% in the bicalutamide group, as compared with 71.3% in the placebo group. So what would it take to convert these data into a shared decision tool for patients? The overall number needed to treat to prevent one death in 12 years is calculated at 20. What does post-hoc subgroup analysis tell us? That the effect is greatest in younger men (<65), and that survival is best overall for those who get their radiotherapy early. What about the harms? The main one is gynaecomastia, affecting 70% in the bicalutamide group. And how has treatment moved on since the trial? Not by a lot, but you can read the detail in the editorial. So how do we present this to a real person in the urology clinic? Clearly, it’s going to take time. Clearly, it’s going to need an infographic. And clearly it’s going to result in different choices, according to how individuals value a 1 in 20 chance of survival beyond 12 years at their particular age, or the inconvenience of having gynaecomastia for 2-3 years. The main thing is to inform the choice as best you can, and respect the patient’s decision.

Intrapericardial LV assistance device

The area of shared decision making I’ve been puzzling over most is advanced heart failure. The big bucks are in left ventricular assist devices, and this week’s NEJM features two: a levitating device that I’ve covered previously and the one in this article, which is described as a small intrapericardial centrifugal-flow device. That’s rather intriguing but since I don’t have a thousand words, you’ll have to look at the picture on the website. The comparator was an axial-flow LVAD already on the market, HeartMate II. The rate of stroke in the new device group was 29.7% versus 12.1% with the older device. Case closed? Well no, the conclusion of the abstract reads: “In this trial involving patients with advanced heart failure who were ineligible for heart transplantation, a small, intrapericardial, centrifugal-flow LVAD was found to be noninferior to an axial-flow LVAD with respect to survival free from disabling stroke or device removal for malfunction or failure.” So they have conflated device removal with disabling stroke. Not the same thing in my book, nor, I suspect, in that of patients with advanced heart failure.

JAMA Intern Med  Feb 2017

What happens when you screen for lung cancer

Sorry to bang on about shared decision making this week (actually, I’m not sorry at all), but how do you share a decision about screening in a high risk group? It’s a question that arises from two papers on the JAMA IM website. The first one describes the effort required to implement a lung cancer screening system in the US Veterans Health Administration (VHA). This was a response to the National Lung Screening Trial (NLST) which found a reduction in lung cancer mortality of 3 per 1000 high-risk individuals screened. The VHA covers 6.7 million Americans, mostly male, of which about 900,000 would meet the criteria for screening. So they decided on a pilot project covering 8 sites. About half the people invited for screening actually attended, but in answer to my opening question, it isn’t clear what they were told or how they decided. This is a rather important space for qualitative research. The organisational effort required was tremendous, and the initial CTs showed abnormalities in 60% of participants, with a false positive rate of 97.5% in the initial period of just under a year. I can hear the sound of Veterans Administrators stroking their chins and pacing their offices as they contemplate rolling this out across the organisation.

Among other Americans, screening followed the usual pattern within its chaotic health system, where anyone can be screened, though not necessarily reimbursed for it, or for managing its downstream consequences. In fact a survey shows that most of those who were screened did not fall within the criteria of NLST. For those who do, the VA study suggests that for every 1000 people screened, 10 will be diagnosed with early-stage lung cancer (potentially curable), and 5 with advanced-stage lung cancer (incurable); 20 will undergo unnecessary invasive procedures (bronchoscopy and thoracotomy) directly related to the screening; and 550 will experience unnecessary alarm and repeated CT scanning (with its associated irradiation). I take these figures from an excellent editorial with the title “Important Questions About Lung Cancer Screening Programs When Incidental Findings Exceed Lung Cancer Nodules by 40 to 1”. So reflecting on this, do you think the VA should go ahead and roll out its lung cancer screening program, and if so, what should it tell those invited?

Lancet  4 Feb 2017  Vol 389

MRI for fetal brain abnormality

The lung cancer screening example shows very clearly that the offer of a diagnostic test is an intervention whose consequences cannot be known ahead of time. Every diagnostic test should be regarded in the same way: you are offering a pathway, not just an investigation, and you are only offering it because you don’t know for sure where it may lead. If you did know, you wouldn’t be doing the test. That, at any rate, is how I think we should conceptualise the use of diagnostic tests, and here’s a quite different but excellent example. Most parents gladly accept the offer of antenatal ultrasound screening, and most of the time it offers them reassurance that their baby is growing well and looks perfectly normal. A fuzzy image of the scan takes first place in the new photograph album. But what if all is not well—worst of all, what if there is something that doesn’t look right about the baby’s brain? Here’s a British prospective cohort study of pregnant women aged 16 years or older whose fetus had a brain abnormality detected by ultrasound at a gestational age of 18 weeks or more, had no contraindications to intrauterine MRI, and consented to enter a study assessing the acceptability and usefulness of this investigation.

It’s good that they included that last bit in the abstract, because it would be good to know how they explained their intentions to these anxious mothers. All went very well: IuMRI provided additional diagnostic information in 387 (49%) of 783 cases, changed prognostic information in at least 157 (20%), and led to changes in clinical management in more than one in three cases. IuMRI also had high patient acceptability, with at least 95% of women saying they would have an iuMRI study if a future pregnancy were complicated by a fetal brain abnormality. Again, I’m so glad they asked this last question.

Lots of oxygen for bronchiolitis?

It’s warm, its moist and it’s full of oxygen. What’s not to like? Well, it’s just the result really. “High-flow warm humidified oxygen did not significantly reduce time on oxygen in moderately severe bronchiolitis compared with standard therapy (cold wall oxygen 100% via infant nasal cannulae at low flow to a maximum of 2 L/min), suggesting that early use of high-flow warm humidified oxygen does not modify the underlying disease process.” A nice sensible Australian trial addressing a common clinical question.

BMJ  4 Feb 2017  Vol 356

Continuity of care and patient outcomes

I’ve always found the formal trappings of the Royal College of General Practitioners slightly embarrassing, including its motto “cum scientia caritas“. It seems a bit un-British to boast about caritas, even in Latin. The College might have done better to come up with a modern logo incorporating the words “able, amiable, and available”, which I was taught long ago were the characteristics of a good GP. Alas, they have become almost unattainable, as able has come to mean digesting endless decontextualised guidelines, amiable has meant a struggle to smile through an unbroken 12-hour day of conflicting demands, and available has come to mean whatever the government decides to impose on a diminishing workforce. It really didn’t have to be like this. General practice has always been hard work: when I started we ran a one-in-three rota which meant working weekends unbroken from Friday morning to Monday evening, during which we tried our best to keep people out of hospital. “Continuity of care” was a proclaimed basic principle of general practice and we took a certain pride in it, though it meant that most of us were less than amiable when called out for the third night in a row. A new study shows that even the remnant of continuous care that remains in general practice still plays a major role in preventing hospital admission for “ambulatory care sensitive conditions” to use the horrible current jargon. But nobody in power is making any serious attempt to nurture a workforce of GPs with sustainable patterns of work and careers that would foster and reward true continuity of care.

PROMs reveal meniscectomy outcomes

A couple of weeks ago I opined that Patient Reported Outcome Measures (PROMS) were a half-way house to patient-centred outcomes, since patients put varying values on outcomes and should be encouraged to prioritise an individual list of them, whereas PROMS reduce them to a generic list for quantitative comparison. In this Danish study, the patient questionnaire used was fairly non-specific, and the main outcome measure that resulted was the knee injury and osteoarthritis outcome score (KOOS). The point of the study was not to refine the scoring system but to determine whether patients with traumatic meniscal tears experience greater improvements in patient reported outcomes after arthroscopic partial meniscectomy than patients with degenerative tears. By the measure of KOOS, they do not.

Overseas training and patient mortality in US

Despite living in a household of devotees to American politics, the only things I know about the subject are that (a) it is based on something called checks and balances and (b) it has given rise to an interminable series called The West Wing, during which I am banned from the living room for my curmudgeonly comments. I do find the current situation quite puzzling. There don’t seem to be any checks and balances and the President now seems to be played by an orange-coloured man who is just as self-righteous as the previous guy in West Wing, only totally deranged. The effect of his ban on Muslims—sorry, people from certain Muslim countries—is going to be as catastrophic for US medicine as anything else he is doing, whether or not it remains in force. Doctors from these countries act as a reality check within the often bizarre world of American hospitals. Fewer people die there if looked after by foreign-trained residents than under the care of US-trained residents, says this timely study from Harvard.

Plant of the Week: Abeliophyllum distichum

I praise this plant every year, because it’s so easy and pleasurable and still so little grown. Usually it appears just after the snowdrops: this year it is coming out at the same time. Its flowers aren’t of as pure a white as the snowdrop’s but their fragrance is sweeter and moreover located nearer to the human nose.

It is best grown by a sunny wall to encourage flowering. Ours faces west and has always flowered abundantly. For the rest of the year it is a rather straggly object, about a metre high in poor dry soil, minding its own business while wisteria and then roses lead the eye well above it. At this time of year, it’s a very cherished plant, offering draughts of Dove soap scent to noses chilled by the late winter winds.