Richard Lehman reviews the latest research in the top medical journals
NEJM 21 Jun 2018
Let’s hear it for sodium thiosulfate
Blooming crystals. The bloom on the crystals of sodium thiosulphate is caused by exposure to water vapour in the air (“efflorescence”). The only medical use for it that I learnt as a teenage chemist was as an antidote to cyanide poisoning. But actually there are several others, sufficient to make sodium thiosulfate (aka hyposulfite) an essential drug on the WHO list. This study describes what must be one of its rarest uses—to prevent ototoxicity in children receiving cisplatin chemotherapy for hepatoblastoma. In this trial it halved the incidence of hearing loss. It now only requires some entrepreneur to repackage the blooming crystals as Onco-Audiosave® and sell them at $350 a dose. Come on Martin, let’s do it as soon as you leave jail.
I’ll be hosting an Overdiagnosis conference this week and the journals seem to know it: there are several articles that perfectly illustrate the problem. This one is about subclinical hyperthyroidism.
The overdiagnosis community regards “subclinical” as a warning: do enough tests, and we can all be subclinically something; eventually we realize that we are all subclinically dead. Our meeting is called POSSUM, standing for Preventing Overdiagnosis through the Shared Understanding of Medicine. But we will try to avoid preaching to the converted and focus instead on changing thinking habits amongst ourselves and those we try to help. One classic bad habit is to use a blood test to rule out something, and then, if it comes back abnormal, to interpret backwards from it. Sometimes I wish TSH (thyrotropin) testing had never been invented. You order the test for someone who is feeling tired all the time. But instead of coming back high, it comes back low. You repeat it in 3 months and it’s still low, whereas the serum free T4 levels are normal. So you don’t know what to do, and the patient is left thinking there’s something not quite right with her thyroid, but there’s nothing to be done about it. Then you repeat it in a year, and so it goes on. Judging from this review, nobody actually knows what do about “subclinical hyperthyroidism”. It’s a BADASS: biochemical abnormality of doubtful actual sense or significance.
JAMA 19 Jun 2018
Stress and autoimmunity
Here’s another example of how difficult a concept “overdiagnosis” really is. It comes from a Swedish registry study linking a prior clinical coding for “stress-related illness” with a subsequent coding for auto-immune disease.
During a mean follow-up of 10 years, the incidence rate of autoimmune diseases was 9.1, 6.0, and 6.5 per 1000 person-years among the exposed, matched unexposed, and sibling cohorts, respectively. So roughly speaking, people who had a previous recorded diagnosis of a stress-related illness had a third more diagnoses of auto-immune disease than those who had no such recorded history. So can we conclude that stress increases the likelihood of auto-immune diseases? I would say not. Remember that this analysis is entirely about note entries, not systematic case finding. In retrospect, some of the “stress” diagnoses may represent a failure to consider auto-immune disease, corrected by testing later. Or, by contrast, some persisting “stress” symptoms may have caused doctors to look for auto-immune disease by sending off more blood tests. Either factor might explain a third more diagnoses, without any causal relation between stress and auto-immunity.
Fat and the city
Public health physicians remind us that our cities are obesogenic environments, with more junk food outlets in areas of highest deprivation. Two papers scan the American obesity landscape for links with ethnicity and urbanization, and how things have changed from the last decade to the current one. For both children and adults, data from the National Health and Nutrition Examination Surveys show a small increase in the overall prevalence of obesity and severe obesity. It remains particularly high amongst Black, Hispanic and under-educated groups. Urbanization is a factor too, but not as you might expect: it is inversely related to the prevalence of obesity. Cities are good.
JAMA Intern Med June 2018
Tamso-uselessin helping stones through the ureter
Tamsolusin was an early example of a “take it in case it works” drug—an immensely lucrative category of drugs that do very little for most individuals but manage to reach a p-value for something if given to enough people. Mostly it’s used for men with prostatic symptoms. Pee-value: ha-ha, yes, that was a joke. The other thing you can do with tamsulosin is give it to people writhing in the agonies of ureteric stone to “relax the ureter” and help the stone to pass. It’s a situation where “something must be done” and this was something to do (N.B. a very common reason for overtreatment, possums). Here is a report of two placebo-controlled trials in which tamsulosin was compared with placebo for people suffering from small (less than 9mm) stones in their ureters. Tamsolusin proved to be tamso-useless for helping more people pass their stones.
Ann Intern Med 19 Jun 2018
Two-test single diagnostic sampling for diabetes
Here’s a observational study which concludes: “Our results support the clinical utility of using a combination of elevated fasting glucose and HbA1c levels from a single blood sample to identify undiagnosed diabetes in the population.”
The authors call it a prospective study, but I would call it retrospective, since it looks at 25-year outcomes of people in the ARIC (Atherosclerosis Risk in Communities) study who had their fasting blood glucose and HbA1c measured on a single occasion between 1990 and 1992. “A single-sample confirmatory definition of diabetes had a high positive predictive value for subsequent diagnosis and was strongly associated with clinical end points.” That is true, especially for all-cause mortality. But you would really need a two-hour seminar to discuss this paper properly; especially whether it provides an argument for case-finding.
The Lancet 22 Jun 2018
I have seen the robots and they might be non-inferior
“In patients with bladder cancer, robotic cystectomy was non-inferior to open cystectomy for 2-year progression-free survival. Increased adoption of robotic surgery in clinical practice should lead to future randomised trials to assess the true value of this surgical approach in patients with other cancer types.” Is this not putting the cart before the robotic horse? Are robots exempt from the general principle that you should test something thoroughly in a well-designed trial before you let it loose on everybody? We should look to the Accelerated Access Collaborative. The stated purpose of this government-industry body is to ensure that “a number of the most promising products will be accelerated through the clinical development and NHS approval processes to treat conditions such as cancer, diabetes, and dementia.” Tell me, how does “accelerated” differ from “not properly tested”? Won’t the premature adoption of poorly tested devices simply waste NHS money and risk harm?
Watching and waiting
About one fifth of patients with locally advanced rectal cancer show complete disappearance of the tumour following radiotherapy and neoadjuvant chemotherapy. Traditionally all of them would still undergo major surgery, with a perioperative mortality of 1-2% and major functional effects in 60%. But might the surgery be avoidable by a policy of careful watching and waiting for recurrence? (“masterly inactivity and cat-like observation” as it used to be referred to, with a hint of irony). It would be hard to do a randomised trial of sufficient duration, so an International Watch & Wait Database (IWWD) aims to describe the outcome of the W&W strategy in a large-scale registry of pooled individual patient data. Most local recurrence happened in the first two years and distant metastasis was uncommon (8%). Five-year overall survival was 85% and 5-year disease-specific survival was 94%.
So for strictly selected patients with a clinical complete response, W&W can be a good alternative to major surgery with very little oncological risk.
e-guidance after surgery
“A personalised e-health intervention after abdominal surgery speeds up the return to normal activities compared with usual care. Implementation of this e-health programme is recommended in patients undergoing intermediate-grade abdominal, gynaecological, or general surgical procedures.” Thus saith the ending of the abstract of this study, funded by the Netherlands organisation ZonMw. The comparator was web-based general advice about post-operative care, while the intervention was a personalised interactive site. I’m a bit baffled as to why the recommendation is so specific, since the authors point out that this is the only tool of its kind in existence. It is badly needed, but if we want to use it, how do we get hold of it? I couldn’t find out from the ZonMw website whether it’s going to be available under Creative Commons or as a commercial package. But I think there’s a clue in the Declaration of Interests: two of the authors “JRA and JAFH intend to set up a spin-off company for implementation of a mobile application concerning the IKHERSTEL intervention in the Netherlands (ie, the intervention under study).”
The BMJ 22 Jun 2018
Imaging and overdiagnosis
Well possums, here’s an article to keep you busy for a while. It’s a massive umbrella review carried out by Jack O’Sullivan, a doctoral researcher in Oxford, aimed at discovering the prevalence and outcomes of incidental imaging findings. This is classic overdiagnosis territory, as readers of Steven Woloshin, Lisa Schwartz, Gill Welch, and Margaret McCartney will know. “Incidentalomas” abound in most forms of imaging done for most sites in the body, and this invaluable paper quantifies them comprehensively. The uncomfortable fact is that for all the harms of overdiagnosis, there are some lives which are actually saved. This applies particularly to renal and ovarian incidentalomas, 50% of which are malignant. But overall, the more you scan, the more you find, and most abnormalities in most sites will be unimportant. There is no simple solution, let alone any simple way to discuss this before imaging individual patients: though I very much hope to be proved wrong by cleverer possums.
Embolisation for prostate hyperplasia?
Surgery for benign prostatic hyperplasia is extremely preference sensitive, as Jack Wennberg and colleagues found when they did the first studies of shared decision making in Dartmouth College in the early 1990s. Both surgeons and patients were found to have grossly exaggerated ideas about the benefits of surgery and the risks of doing nothing. Essentially surgery is done to relieve symptoms in the long term, so the only outcomes that matter are symptoms, the long term, and harms. This trial comparing transurethral resection with prostate artery embolisation reports outcomes at 12 weeks. For me, this is an instant fail. Prostate symptoms come and go: I need to know 12-year outcomes, before I will let anyone near mine. It was conducted in a single Swiss centre, and randomised 103 patients. Sorry, another fail. Harms? How can we tell—with such numbers over such a short period? So that leaves symptoms, which were the same at three months, even though TURP led to better measurements of voiding. But the one thing I do like about this study is its frank acknowledgement of tribalism among professionals: “The discussion regarding the use of PAE to treat BPH-LUTS is highly controversial and is influenced by political interests, as a fundamental domain of urologists is challenged by interventional radiologists.” Possums, if you ever see a domain in medicine, please challenge it.
Plant of the Week: Clematis ‘Błękitny Anioł’
Of the many excellent clematis hybrids to come out of Poland over the last decades, this is one of the finest. Although it’s called “Blue Angel”, its four huge crinkly sepals encompass a palette from pale blue through faint mauve and even grey. Yet far from coming over weak and washy, they are full of character and beauty.
The man behind these hybrids was a humble monk who used them as a quiet form of political dissent during the Communist era by giving them patriotic names that the authorities were bound to dislike: General Sikorski, Cardinal Wyszinski, Polish Spirit, Warszawa Nike, Jan Pawel II, and even a Lech Wałęsa. Like Gregor Mendel in his Bohemian monastery garden, Brother Stefan Franczak spent obscure decades crossing his favourite plants, but unlike Mendel has achieved fame and recognition in his lifetime.
These varieties all contain C viticella blood, as we say in gardening circles, making them tough and vigorous. ‘Błękitny Anioł’ can easily cover a small tree. So make sure you put it next to one that gets boring in July (most do), and you will enjoy seeing it covered in an abundance of flower from now until the end of October.