FROM THE JOURNALS Edited highlights of Richard Lehman’s ... Co-amoxiclav for otitis media in under 2s This trial firmly establishes the superiority of a 10 day course of co-amoxiclav

the bmj | 14 January 2017 61

How similar is biosimilar?

I once claimed that the easiest way to become a billionaire was to find a plausible cancer target, make an antibody to it, add the suffix uzumab, and then sell it to desperate patients with cancer at £8000 for each notional added week of life. It is amazing how all these drugs end up costing about the same. The market solution to this should be competition from other people making similar antibodies. Myuzumab should bring down the cost of Youruzumab. In your dreamuzumabs, I hear you say. And yet it might be starting to happen for some antibodies that are running out of patent. One such is trastuzumab, used alongside chemotherapy for the treatment of HER2—which we now have to call ERBB2—positive breast cancer. Somebody has now produced an as yet unnamed antibody to ERBB2 (anonymouzumab?) and compared it with trastuzumab in patients co-treated with taxane for their metastatic breast cancer. It looks like a good thing: it is certainly as effective as trastuzumab, and might be better. Let’s hope it’s 10 times cheaper.

JAMA 2017, doi:10.1001/jama.2016.18305

Not so often with zoledronic acid

In Mike Rosenblatt’s article, he refers to the long processes of dose finding that preceded the phase 3 placebo controlled trials of alendronate for osteoporosis. Zoledronic acid is a far more powerful bisphosphonate. For the treatment of postmenopausal osteoporosis, 5 mg given intravenously can put the osteoclasts out of action for a year. For metastatic cancer eating away at bone, however, the situation is less clear. Much higher doses are used, calculated according to body weight and renal function. Traditionally the drug is given every four weeks, but here’s a trial that compared this with giving it every 12 weeks to patients with bone metastases due to breast cancer, prostate cancer, or multiple myeloma. Over two years, the less frequent dosing worked equally well.

JAMA 2017, doi:10.1001/jama.2016.19425

Dementia and busy roads

Is living near a busy road associated with a higher risk of developing dementia? Yes. It says so in the Lancet. Do road traffic emissions cause dementia? Dunno. It’s plausible. However, it could be the dearth of town sparrows in high traffic environments and the absence of the protective effect of inhaling their aerosolised faeces. This is less plausible, I know, but I’d like to point out a highly suggestive study of sparrow droppings in relation to brain plaques in mice that I seem to remember reading in an ornithology journal. As I’ve pointed out before, when you go looking for associations and find one with a statistically significant effect size, that doesn’t mean that you should start hypothesising about causality. It is just as likely that you have failed to correct for some other factor that never occurred to you. The more you want to believe in causality, the more you should question yourself. I don’t want people to live in areas of high road traffic density, I don’t want motor vehicles to spew out carbon monoxide and dioxide, nitrogen oxides, or particulate matter, and I don’t want to get dementia. This paper shows that in a large population based cohort in Ontario, living close to heavy traffic was associated with a 5% higher incidence of dementia but not of Parkinson’s disease or multiple sclerosis. Avoid living near heavy traffic, feed your sparrows, don’t live in Ontario? Not sure which.

Lancet 2017, doi:10.1016/S0140-6736(16)32399-6

Co-amoxiclav for otitis media in under 2s

This trial firmly establishes the superiority of a 10 day course of co-amoxiclav over a five day course for babies and toddlers with acute otitis media. The figures are clear: with only five days’ treatment, failure of resolution at 10 days is 34%, whereas if 10 days of treatment are given, it is 16%, and the infants experience fewer symptoms. So here we have a nice simple trial to inform practice in a common and distressing clinical situation. More is sometimes more.

N Engl J Med 2016, doi:10.1056/NEJMoa1606043

research updateF RO M T H E J O U R NA L S Edited highlights of Richard Lehman’s blog on http://bmj.co/Lehman

“Incompatible” kidneys and survivalWhen I was a medical student 45 years ago, immunology was still a very young science and renal transplantation was a very new procedure. But even then we knew about human leucocyte antigen and the need to find a compatible kidney donor. Or did we? Does this need actually exist in the current era of better immunosuppression? From an outcomes study in the UK comparing human leucocyte antigen compatible and incompatible donor kidney transplants, it seems that matching may no longer be important. But this is a complex analysis in a complex area that’s way beyond my competence: if you need to know the detail, look up the study. And this whole sentence applies to pretty well everything I write about.

Lancet 2017, doi:10.1016/S0140-6736(16)31595-1

the bmj | 14 January 2017 6362 14 January 2017 | the bmj

Effectiveness of flexible sigmoidoscopy screening in men and women and different age groupsHolme Ø, Schoen RE, Senore C, et alCite this as: BMJ 2017;356:i6673Find this at: http://dx.doi.org/10.1136/bmj.i6673

Study question Does the effectiveness of flexible sigmoidoscopy screening for colorectal cancer vary between men and women, and between different age groups?

Methods Data were pooled and analysed from three large randomised trials comparing flexible sigmoidoscopy screening with usual care (the US Prostate, Lung, Ovarian and Colorectal Cancer Screening trial (PLCO), Italian Screening Colon Rectum trial (SCORE), and Norwegian Colorectal Cancer Prevention trial (NORCCAP)). Data were stratified by sex and in five year age

groups. Participants were aged between 50 and 74 years (n=287 928).

Study answer and limitations After 10.5 to 11.9 years of follow-up, incidence of colorectal cancer was significantly reduced among men in the screening group compared with the usual care group regardless of age (relative risk 0.76, 95% confidence interval 0.70 to 0.83). Screening significantly reduced incidence of colorectal cancer in younger women (age <60; 0.71, 0.59 to 0.84), but not in those aged 60 years and older (0.90, 0.80 to 1.02). Screening reduced incidence to a similar extent in the distal colon in men and women, but there was no effect of screening in the proximal colon in older women with a significant interaction between sex and age group (P=0.04). The pooled analysis included few studies of unequal size, and data from a fourth trial on flexible sigmoidoscopy screening (the UK Flexi Scope Trial) were not available.

What this study adds Flexible sigmoidoscopy screening is recommended for men and women. For women aged 60 years and older, other screening tools that more effectively detect proximal tumours should be considered.

Funding, competing interests, data sharing Funded by grants from Sørlandet Hospital Kristiansand, Norway and the Regional Health Authorities of South-Eastern Norway. The authors report no conflicts of interest. Data may be obtained from the lead authors of the individual trials.

ORIGINAL RESEARCH Pooled analysis of randomised trials

For open access to the full text of all research articles, see thebmj.com

An increasing number of countries have implemented or are about to implement programmes for colorectal cancer screening. Unlike existing screening programmes for breast and cervical cancer, colorectal cancer screening affects both men and women so we need to know whether established screening tools are equally effective in both sexes. To explore this issue, Holme and colleagues1 report a joint analysis of three large randomised trials of flexible sigmoidoscopy screening conducted in Italy,2 Norway,3 and the United States.4

Among men, colorectal cancer incidence fell significantly by 24%, irrespective of age at screening. However, an effect on incidence in women was only observed in those younger than 60 at screening. The proportion of proximal colorectal cancers, which are less preventable by flexible sigmoidoscopy than distal colorectal cancers, is higher in women than in men. This could partly explain the observed patterns but other possibilities include a potential age trend towards particularly

fast growing lesions in the proximal colon or towards proximal neoplasia without distal lesions among women.

Whatever the underlying cause, these differences have complex implications for sex specific screening recommendations. Among women over 60, for example, screening with flexible sigmoidoscopy should not be recommended, because current evidence indicates that it is not effective. Screening colonoscopy might be a reasonable alternative given the high proportion of proximal colorectal cancers in older women. However, in men, the number of proximal colorectal cancers is also relevant. When we consider the absolute risk of colorectal cancer,5 which is higher in men than in women, it becomes apparent that the 10 year risk of developing proximal colorectal cancer is similar in 60 year old men and women. Would it be acceptable to offer colonoscopy to 60 year old women but not men?

One could argue that the higher life expectancy in 60 year old women justifies more intense screening because—in theory—more life years could be gained by preventing colorectal cancer. However, using the overall gain in life years instead of disease specific outcomes leads to another debate. Is cancer screening actually intended to reduce all cause mortality, given that it will never be feasible to prove such an effect empirically?6 7 The same consideration—that is, balancing the higher absolute risk of colorectal cancer in men against the higher life expectancy in women—can also be applied to the question of whether the starting age for screening should be the same in men and women.8

Holme and colleagues’ study shows that uniform recommendations for men and women can lead to sex related inequity in benefits. Further studies should now be done on this and other screening tools to help provide a secure scientific basis for sex specific recommendations. Cite this as: BMJ 2017;356:j75

Find this at: http://dx.doi.org/10.1136/bmj.j75

Whatever the underlying cause, these differences have complex implications for sex specific screening recommendations

Flexible sigmoidoscopy screening for colorectal cancer

COMMENTARY Different outcomes in men and women are a challenge for decision makers

Ulrike Haug [email protected] thebmj.com for author details

the bmj | 14 January 2017 6362 14 January 2017 | the bmj

Cost effectiveness of a government supported policy strategy to decrease sodium intakeWebb M, Fahimi S, Singh GM, et alCite this as: BMJ 2017;356:i6699Find this at: http://dx.doi.org/10.1136/bmj.i6699

Study question How cost effective is a government “soft regulation” national policy intervention combining targeted industry agreements and public education in reducing sodium intake?

Methods In this analysis across 183 nations, the authors characterised global sodium intakes, blood pressure levels, effects of sodium level on blood pressure and of blood pressure on cardiovascular disease, and cardiovascular disease rates in 2010, each by age and sex. Country specific costs of a policy that combines government supported education and targeted industry agreements to reduce population sodium intake were estimated using the World Health Organization Noncommunicable Disease Costing Tool. Country specific impacts on mortality and disability adjusted life years (DALYs) were modelled using comparative risk assessment, based on scenarios including 10%, 30%, 0.5 g/day, and 1.5 g/day reductions in sodium intake achieved over 10 years. Cost effectiveness

was evaluated as purchasing power parity adjusted international dollars (I$, equivalent to the country specific purchasing power of US dollars) per DALY saved over 10 years.

Study answer and limitations Worldwide, a 10% reduction in sodium intake over 10 years within each country was projected to avert 5 781 000 cardiovascular disease related DALYs/year, at a mean cost of I$1.13 per capita over the 10 year intervention and with a cost effectiveness ratio of I$204/DALY. The authors did not account for possible unintended consequences of the intervention, nor model health system savings from averted cardiovascular disease events.

What this study adds A government “soft regulation” strategy combining targeted industry agreements and public education to reduce population sodium consumption by 10% over 10 years would be extremely cost effective in nearly all of 183 nations evaluated.Funding, competing interests, data sharing This research was supported by the National Institutes of Health. DM reports ad hoc honorariums or consulting fees from Boston Heart Diagnostics, Haas Avocado Board, Astra Zeneca, GOED, DSM, and Life Sciences Research Organization, none of which were related to topics of dietary sodium. The global data on sodium intake, blood pressure, and cardiovascular events are all available for download (see full paper on thebmj.com for urls).

ORIGINAL RESEARCH Global analysis across 183 nations

Five years of specialised early intervention versus two years of specialised early intervention followed by three years of standard treatment for patients with a first episode psychosisAlbert N, Melau M, Jensen H, et alCite this as: BMJ 2017;356:i6681Find this at: http://dx.doi.org/10.1136/bmj.i6681

Study question What are the effects of five years of specialised early intervention (SEI) treatment for a first episode of psychosis in schizophrenia spectrum disorder versus two years of SEI plus three years of treatment as usual?

Methods Of 400 participants with first episode schizophrenia spectrum disorder recruited in Denmark between 2009 and

2012, 197 were randomised to prolonged SEI treatment and 203 to treatment as usual. SEI treatment consists of three core elements—modified assertive community treatment, family involvement, and social skill training—with a low patient-case manager ratio. Follow-up assessments were conducted five years after the start of the SEI programme. The primary outcome was negative symptoms measured on the scale for assessment of negative symptoms (avolition-apathy, anhedonia, alogia, and affective blunting).

Study answer and limitations There was no difference between study groups on levels of negative symptoms (1.72 v 1.81 points; estimated mean difference −0.10 (95% confidence interval 0.33 to 0.13), P=0.39). Participants receiving prolonged SEI treatment were more likely to remain in contact with

specialised psychiatric services (90.4% v 55.6%, P<0.001) and had higher client satisfaction (estimated mean difference 2.57 points (95% confidence interval 1.36 to 3.79), P<0.001) than controls. The lack of effect could be due to the intensity of the treatment provided to control participants.

What this study adds This trial investigates the effect of SEI treatment for up to five years. Although few effects of the prolonged treatment were found, neither was any evidence of a deteriorating course of illness.

Funding, competing interests, data sharing Funded by the Danish Agency for Science and Technology and Innovation, Capital Region Denmark, and Central Region Denmark. The authors declare no competing interests. Full dataset will be made available at the Danish National Archives (Rigsarkivet) after the initial publication.

Study registration Clinicaltrial.gov NCT00914238.

ORIGINAL RESEARCH Randomised, superiority, parallel group trial in Denmark (OPUS II)

30 100 800Cost e�ectiveness ratio (I$/DALY)

5000 30 000

Cost effectiveness (purchasing power adjusted I$/disability adjusted life year) by country of a national policy intervention to reduce sodium consumption by 10%

the bmj | 14 January 2017 PB64 14 January 2017 | the bmj

Migraine and risk of perioperative ischaemic stroke and hospital readmissionTimm FP, Houle TT, Grabitz SD, et alCite this as: BMJ 2017;356:i6635Find this at: http://dx.doi.org/10.1136/bmj.i6635

Study question Are patients with migraine at increased risk of perioperative ischaemic stroke and, if so, does this lead to an increased hospital readmission rate?

Methods This study was a data analysis of 124 558 patients undergoing surgery between January 2007 and August 2014 at three hospitals in Boston, MA, USA. The exposure variable (history of migraine and migraine aura) and the primary outcome (perioperative ischaemic stroke occurring within 30 days after surgery) were identified using ICD-9 (international classification of diseases, ninth edition) codes. Perioperative ischaemic stroke events were validated by medical chart review. The secondary outcome was hospital readmission within 30 days of surgery.

Odds ratios of outcomes associated with migraine status were estimated using logistic regression models adjusted for various comorbidities, intraoperative parameters, and surgery related risks.

Study answer and limitations A history of migraine was associated with an increased risk of perioperative ischemic stroke (odds ratio 1.75, 95% confidence interval 1.39 to 2.21) and 30 day hospital readmission (1.31, 1.22 to 1.41). The predicted absolute risks are 2.4 (2.1 to 2.8) perioperative ischemic strokes for every 1000 surgical patients, 4.3 (3.2 to 5.3) for every 1000 patients with any migraine diagnosis, and 6.3 (3.2 to 9.5) for migraine with aura. The use of ICD-9 codes for a diagnosis of migraine may capture patients with the most severe headaches.

What this study adds Patients with migraine, and particularly migraine with aura, are at increased risk of perioperative ischaemic stroke and 30 day hospital readmission after surgery. Migraine should be incorporated in the perioperative stroke risk assessment.

Funding, competing interests, data sharing This study was funded by Jeff and Judy Buzen. TTH has received funding from Merck and the National Institutes of Health. Other authors have received honorariums for editorial services from The BMJ and Cephalalgia, travel funding from the International Headache Society, and funding from Merck.

The BMJ is an Open Access journal. We set no word limits on BMJ research articles, but they are abridged for print.

The full text of each BMJ research article is freely available on thebmj.com.

The online version is published along with peer and patient reviews for the paper, and a statement about how the authors will share data from their study. It also includes a description of whether and how patients were included in the design or reporting of the research.

The linked commentaries in this section appear on thebmj.com as editorials. Use the citation given at the end of commentaries to cite an article or find it online.

ORIGINAL RESEARCH Hospital based registry study

Incidences and multivariable adjusted odds ratios for study outcomes according to migraine status. Values are numbers (percentages) unless stated otherwise

OutcomeTotal

(n=124 558)No migraine (n=114 379)

Migraine status (n=10 179) Multivariable adjusted odds ratio (95% CI); P valueAny migraine (n=10 179)

Migraine without aura (n=8901)

Migraine with aura (n=1278) Any migraine Migraine with aura

Perioperative ischaemic stroke

771 (0.6) 682 (0.6) 89 (0.9) 71 (0.8) 18 (1.4) 1.75 (1.39 to 2.21); P<0.001 2.61 (1.59 to 4.29); P<0.001

Hospital readmission within 30 days

10 088 (8.1) 9160 (8.0) 928 (9.1) 786 (8.8) 142 (11.1) 1.31 (1.22 to 1.41); P<0.001 1.59 (1.33 to 1.91); P<0.001

STEV

E M

CCAF

EE