Richard Lehman’s journal review—12 January 2015

A vaccine that works really well is the best kind of medical intervention. But a vaccine that gives partial protection is a headache.

NEJM 8 Jan 2015 Vol 372

113 A vaccine that works really well is the best kind of medical intervention. But a vaccine that gives partial protection is a headache. Sanofi Pasteur has developed a tetravalent vaccine which is 60.8% protective against symptomatic dengue in children in Latin American countries where dengue is endemic. It is least effective in serotypes 1 and 2 and most effective in serotypes 3 and 4, and it is a distinct improvement on previous experimental vaccines against this increasingly common flavivirus. There is a huge market to be tapped, but is this product ready for prime time? The accompanying editorial is very informative on the science but non-committal on the billion dollar question.

134 When I see a cancer trial in one of the top journals, I generally scan the abstract to see if there is anything of general interest or importance. I drew a complete blank with this one. “Among women with predominantly stage I HER2-positive breast cancer, treatment with adjuvant paclitaxel plus trastuzumab was associated with a risk of early recurrence of about 2%; 6% of patients withdrew from the study because of protocol-specified adverse events.” Um, compared with what? Well, with what might have happened otherwise, elsewhere, possibly. “We recognize that a prospective, randomized trial would have been the best option. However, we did not believe that such a design would have been feasible given the accumulating evidence from retrospective studies. Patients and their providers may have been unlikely to enroll in a trial that included a group in which patients would not receive trastuzumab. Some clinicians and investigators might have argued for a trial of trastuzumab alone versus trastuzumab plus chemotherapy, but there are limited data indicating that trastuzumab alone is an effective approach.” Aha, but isn’t that exactly why people do RCTs? Why does this trial—which is just a large case series—get space in the New England Journal? Any time your well conducted, randomized, double blind trial gets rejected by the NEJM, you might like to ask them.

JAMA 6 Jan 2015 Vol 313

37 A study of life expectancy in type 1 diabetes kicks off with the statement “Major advances in treatment of type 1 diabetes have occurred in the past 3 decades.” An acquaintance recently met the Oxford consultant who taught me about diabetes 40 years ago and asked him what he thought the advances had been. After a pause for thought, he replied “Insulin needles are now a lot finer and less painful.” We may be on the verge of really transformative treatments, but this Scottish paper—a mixture of observation and modelling—shows that on the basis of data from 2008-2010, people with type 1 diabetes at the age of 20 have an estimated loss of life expectancy of approximately 11 years for men and 13 years for women compared with the general population. A bit better than 30 years ago, but not better enough.

45 The best data we have about the importance of good blood sugar control in type 1 diabetes come from the Diabetes Control and Complications Trial (DCCT), in which a total of 1441 healthy volunteers with diabetes mellitus aged 13 to 39 were randomized to 7 years of tight control or usual care. After a mean of 27 years’ follow-up, those in the tight control group were more likely to be alive (hazard ratio for death = 0.67 [95% CI, 0.46-0.99]—note how close to the significance boundary this gets). There is a slightly downbeat editorial about these T1DM studies (“Two Steps Forward, One Step Backward”) which comments that “Despite renal protective therapies, end-stage renal disease from diabetes has continued to increase” which makes you wonder how renoprotective these therapies really are.

62 For people who are obese and have type 2 diabetes, however, there has been an enormous advance—bariatric surgery. Comparison between 2500 Veterans Administration patients—with or without diabetes—who underwent such surgery with matched controls shows a halving of mortality in the first five years. This difference is maintained in the next five years too, and was equal in diabetic and non-diabetic subjects.

Ann Intern Med 6 Jan 2015 Vol 162

27 Having an unprovoked deep vein thrombosis is a powerful risk factor for having a second DVT. The risk declines over time but never quite goes away. Researchers in Canada sought to establish whether it would be useful to measure D-dimer before stopping anticoagulation, and then, if the D-dimer was negative, stop the anticoagulants, measure it again after a month and restart the drugs if it had become positive. Or leave them off if both tests were negative. It didn’t quite work in men, who showed a 9.7% annual recurrence rate even after two negative tests. In women the rate was 5.4%, which might just be acceptable.

OL Are you the kind of clinician who actually treats blood pressure? If you’re a GP, that means you’re probably responsible for about 400 people taking drugs for let’s say an average of 15 years = 52 million drug-hours. So do click on the link and take half an hour of your own time to mull over this free systematic review of the effects of BP reduction in mild hypertension. Consider what you are trying to do. Reduce cardiovascular risk, right? If you have come across John Yudkin’s Ten Commandments, you will remember the one that says “Thou shalt treat according to level of risk and not to level of risk factor.” So for each “patient” on BP-lowering medication, you have calculated a risk score such as QRISK? And discussed each element of it with each individual and what non-drug and drug treatments might help to reduce it? Giving every person an individualized number-needed-to-treat and number needed-to-harm for each intervention? If you have answered yes to all of these questions, you must be lying, because the information to support this detail of shared decision making simply isn’t there. This review, which lumps together all sorts of individuals—with and without diabetes, some with previous treatment and some not—ends up concluding that treating people with “grade 1 hypertension” is probably going to reduce cardiovascular events but that the confidence intervals are huge. So which risk-reducing intervention is it going to be? More physical activity? A statin? A BP lowering agent? A more “Mediterranean” diet? Metformin? All of the above, or none of them? After all, it is the symptomless individual who has to decide on the basis of the information you give them. This is the mess we are in and the mess we need to get out of if shared decision making about risk reduction is going to become a reality. Because “mild hypertension” is not a thing in itself: it is just a single risk factor. And “all-cause mortality”—the thing that gets us all in the end—does not seem to be postponed by any of the mild BP treatments described in this paper.

JAMA Intern Med Jan 2015

Perhaps the grim reaper can be held off a while by eating more whole grain products. This is one possible conclusion to be drawn from dietary information fed into those two familiar American cohort studies, 74 341 women from the Nurses’ Health Study (1984–2010) and 43 744 men from the Health Professionals Follow-Up Study (1986–2010). ” These data indicate that higher whole grain consumption is associated with lower total and CVD mortality in US men and women, independent of other dietary and lifestyle factors.” So the genes that make people like to eat grainy things may be associated with the genes that make people die less from cardiovascular disease. Or maybe the grainy things themselves have that effect. Either way, I shall eat what I like, avoiding buckwheat and couscous and similar vile things.

Lancet 10 Jan 2015 Vol 385

Most of this week’s print Lancet is taken up with the huge Global Burden of Disease Study which I alluded to before Christmas. Just to clear up some misinformation I was guilty of, this Gates-funded survey is not behind a paywall, but it is not open access either: you have to subscribe (for free) to get to it. It is well worth the minimal effort.

OL The Cholesterol Treatment Trialists’ Collaboration offer a paper on “Efficacy and safety of LDL-lowering therapy among men and women: meta-analysis of individual data from 174 000 participants in 27 randomised trials.” But although there is now some weak evidence LDL-C lowering with ezetimibe may improve cardiovascular outcomes, all the evidence in this review is about statins. These drugs certainly lower LDL-C and “These results indicate that, for each 1 mmol/L reduction in LDL cholesterol, statin therapy reduced major vascular events by about a fifth, major coronary events by a quarter, coronary revascularisations by a quarter, and ischaemic stroke by just under a fifth, and that these proportional reductions were similar in men and women, even though on average women had somewhat lower absolute cardiovascular risk in these trials.” But it’s not just pedantic obstinacy that makes me chary about putting all this down to LDL-C lowering. I’m worried that people will therefore carry on treating cholesterol as some kind of target independent from total cardiovascular risk, whereas we need always to obey the commandment “Treat to level of risk and not to level of risk factor.” And scientifically I still can’t understand how statins can produce marked improvements in acute events before they have had time to reduce LDL-C.

BMJ 10 Jan 2015 Vol 350

Tiny babies frighten me at the best of times and I really admire people who can look after extremely premature infants. Their whole future can depend on how much oxygen reaches their brains in the first days of life: they can cope with brief periods of anoxia but not with prolonged hypoxia. A method of measuring actual oxygen delivery to cerebral tissue sounds like a wonderful thing, and this study shows that it probably is. Cerebral near infrared spectroscopy (NIRS) oximetry can measure the top 2cm of brain oxygenation and use this monitoring helped to keep it at optimal levels in this exploratory European trial. It was not powered to establish an effect on clinical outcomes but the effect it did show was positive.

Last week The BMJ published a Korean meta-analysis claiming to show that blood sugar elevation was an important risk factor for pancreatic cancer. This week it prints a Greek “umbrella review of meta-analyses of observational studies” linking type 2 diabetes and cancer. Pancreatic cancer does not make it into the significant association group, and the authors—overseen by that master of rigour, John Ioannidis—conclude that “Though type 2 diabetes has been extensively studied in relation to risk of developing cancer and cancer mortality and strong claims of significance exist for most of the studied associations, only a minority of these associations have robust supporting evidence without hints of bias… only the associations between type 2 diabetes and risk of developing breast cancer, intrahepatic cholangiocarcinoma, colorectal cancer, and endometrial cancer remained nominally significant. This result does not necessarily mean that only these four associations are genuine. It does, however, suggest that there is still substantial uncertainty about the other nominally significant associations.” Looking back at last week’s paper, I’m not sure it would have passed the Ioannidis test. Perhaps The BMJ should have asked him.

Literature and Medicine

My first degree was in English Literature, taken as a medical student to escape the physiology degree course which, 45 years later, is still routine in Oxford. I’ve often dreamt of setting up an undergraduate Medical Humanities option there, but I’ve never proceeded further than occasionally helping teach clinical students in a “Medicine and Literature” special study module.

To me the main point is to teach the students that medicine provides just a partial and distorted insight in the totality of life. Any texts involving doctors are out. For me, the course should be mainly about teaching young aspiring doctors—increasingly from privileged and professional families—how unimportant they are. There are so many other people telling them the opposite. These weeks should be spent impressing on them what a tiny realm of discourse medicine is when regarded within the unbounded totality of human thought and expression.

No wonder I don’t get invited often.

There is a nice little editorial about Literature and Medicine in this week’s Lancet to accompany an excellent piece by Gabriel Weston, “Developing judgement, not being judgmental.”

Here is Józef Teodor Konrad Korzeniowski as he struggles to turn himself into the English author Joseph Conrad. Contrasting the artist’s role with that of a scientist or thinker, he declares:

“His appeal is less loud, more profound, less distinct, more stirring—and sooner forgotten. Yet its effect endures for ever. The changing wisdom of successive generations discards ideas, questions, facts, demolishes theories. But the artist appeals to that part of our being which is not dependent on wisdom; to that which is a gift and not an acquisition—and, therefore, more permanently enduring. He speaks to our capacity for delight and wonder, to the sense of mystery surrounding our lives; to our sense of pity, and beauty, and pain; to the latent feeling of fellowship with all creation—and to the subtle but invincible solidarity that knits together the loneliness of innumerable hearts, to the solidarity in dreams, in joy, in sorrow, in aspirations, in illusions, in hope, in fear, which binds men to each other, which binds together all humanity—the dead to the living and the living to the unborn.”

Preface to The Nigger of the “Narcissus”, 1897.