Blocking veins to heal ulcersBritish surgeons are doing rather well in the New England Journal of Medicine. This week’s print copy features the ground breaking PRECISION trial of prostate magnetic resonance imaging, which I commented on a couple of weeks ago. And queuing on the journal’s website we have the EVRA trial of early versus delayed vein surgery for varicose ulcers. Pooling and retrograde flow of blood in the legs is thought to be the mechanism that maintains skin breakdown in people with varicose ulcers. That’s the logic of ablating superficial leg veins in the hope that the blood then diverts to the deeper veins and allows the ulcer to heal faster. The hypothesis that early ablation would result in faster healing than deferred ablation was tested in 20 hospitals around the UK. And it did: median healing time 56 days versus 82 days.

N Engl J Med doi:10.1056/NEJMoa1801214

Arteries make the best coronary artery bypass graft Gosh: a systematic review has sneaked its way into the New England Journal. Or should we say “snuck,” as this is America. Conveniently for surgeons who wish to bypass blocked coronary arteries, the human extremities contain two spare vessels that can be used for grafting: the radial artery and the saphenous vein. There have been small randomised controlled trials comparing outcomes from the two, but none of sufficient size to be definitive. But now the NEJM has discovered this thing called individual patient data meta-analysis, which can solve the problem by, well, combining individual data from trials. There was no mortality difference between 534 patients with radial artery grafts and 502 patients with saphenous vein grafts, but the use of radial artery grafts for coronary artery bypass resulted in a lower rate of adverse cardiac events and a higher rate of patency at five years of follow-up.

N Engl J Med doi:10.1056/NEJMoa1716026

Fluids in abdominal surgery And now from Australia and New Zealand, a third surgical paper from the NEJM website, imparting useful knowledge for practice. Fashions in fluid replacement are always changing, and current guidelines for patients undergoing major surgery recommend a restrictive intravenous fluid strategy to promote recovery. But why, when there is no evidence one way or the other? Too often guidelines are about what eminent people think: Claude Bernard’s reply

would have been “Why think, when you can do the experiment?” So the Ozzies and the Kiwis did the experiment and lo! “Among patients at increased risk for complications during major abdominal surgery, a restrictive fluid regimen was not associated with a higher rate of disability free survival than a liberal fluid regimen and was associated with a higher rate of acute kidney injury.”

N Engl J Med doi:10.1056/NEJMoa1801601

Grouped care for grouped pain Pain is a solitary experience. Group effects may tell you which treatment is most likely to work in general, but not what treatment will actually take away the pain of the individual you are trying to help. Expect analgesic failure; pursue analgesic success is a great essay setting out these principles, though as it was written in 2013 its authors sometimes slide back into discussing analgesia in terms of conventional meta-analytics. Here’s a systematic review of psychological interventions for pain which comes up with few insights, except that group therapy tends to be more effective than individual therapy. This suggests the loneliness of pain becomes part of the pain for many.

JAMA Intern Med doi:10.1001/jamainternmed.2018.0756

Bare metal stents for saphenous grafts I’ve tried to apply my fading powers to coronary stents for 20 years, but I still haven’t worked out which is the winner. Yes, I know drug eluting stents were declared better than bare metal stents about 15 years ago, but it’s the why that escapes me. But here is a clear trial that compares bare metal with drug-eluting stents in a post-coronary artery bypass population. “In patients undergoing stenting of de-novo saphenous-vein-grafted lesions, no significant differences in outcomes between those receiving DES and BMS during 12 months of follow-up were found.”

Lancet doi:10.1016/S0140-6736(18)30801-8

the bmj | 19 May 2018 241

education

educationeducationFROM THE JOURNALS Edited highlights of Richard Lehman’s blog on http://bmj.co/Lehman

Water cure for kidney disease? In 1945, the US Board for Food and Nutrition declared that the ideal fluid intake of a human being is 2.5 L a day. The concept that humans can’t rely on thirst to tell them how much to drink has taken on a strange life of its own, and the survival of Western humanity has come to depend on the constant consumption of water from plastic bottles. Moreover, in recent years, a large percentage of people have been found to have an estimated glomerular filtration rate of between 30-60 mL/min per 1.73 m2, conferring on them the status of “stage 3 kidney disease.” Should they be coached/coaxed into increasing their water intake even further? In a Canadian trial, a coaxed group produced more urine, but their kidney function was the same after a year as an uncoaxed group. Time for a trial among the inhabitants of the Kalahari desert.

JAMA doi:10.1001/jama.2018.4930

242 19 May 2018 | the bmj

PRACTICE POINTER

Orofacial painAdonye Banigo,1 David Watson,2 Bhaskar Ram,1 Kim Ah-See1

1Aberdeen Royal Infirmary, Aberdeen AB25 2ZN, UK2Hamilton Medical Group, AberdeenCorrespondence to: A Banigo a.banigo@nhs.net

Orofacial pain or facial pain described as an ache in the front part of the head (including the oral cavity) is a common presentation in primary care. Nearly a quarter of patients in a British primary care study (2504 adult patients) reported orofacial pain.1 The pain may be musculoskeletal, dental, neural, or sinogenic in origin.2

In our clinical experience orofacial pain is often incorrectly attributed to rhinosinusitis, commonly referred to as sinusitis. Some patients refer to the pain as a sinus headache. In a case series of 973 patients with presumed rhinosinusitis, only 1 in 10 patients were confirmed to have paranasal sinus disease on endoscopy and computed tomography.3 As such, the correct diagnosis may be missed or delayed and result in inappropriate treatment and prolonged symptoms.

This article aims to enable readers to achieve a more accurate diagnosis of orofacial pain focusing on the characteristics and associated features of facial pain and presents an initial approach to managing these patients in primary care. We focus on chronic orofacial pain—that is, pain lasting more than 12 weeks—as usually by then a trial of treatment might have been explored in primary care, prompting reconsideration of the diagnosis if the symptoms persist.

What are the common causes

While there is no robust evidence on the prevalence of different conditions, migraine and mid-segment facial pain are among the commonest causes of chronic orofacial pain.4 However, these are often misdiagnosed as rhinosinusitis because of accompanying nasal symptoms. We describe below these conditions and their presenting features.

RhinosinusitisRhinosinusitis is inflammation of the lining of the nasal cavity and the sinuses, with or without the formation of nasal polyps. Symptoms can be acute (<12 weeks) or chronic (>12 weeks). It is estimated to affect 2-10% of the population.4 A population based study from Canada (73 000 participants) reported a prevalence of 3.4% in men and 5.7% in women.5

Orofacial pain is not a hallmark feature of rhinosinusitis. The European Position Paper on Rhinosinusitis and Nasal Polyps (EPOS)4 defines four symptoms of rhinosinusitis, of which the first two must be present to support a diagnosis (see box 1). As such, patients who have orofacial pain but do not have nasal congestion, nasal blockage, or anterior or posterior nasal discharge are unlikely to have rhinosinusitis.

Data from cohort studies show that orofacial pain is a poor predictor of chronic rhinosinusitis, and patients are more likely to be misdiagnosed if pain is used as a criterion to screen patients for rhinosinusitis.6-9 A prospective observational study involving 108 patients with confirmed chronic rhinosinusitis found that less than a third of patients had orofacial pain.7 This figure drops to 1 in 10 patients when nasal polyps are present,8 which suggests that orofacial pain is even more uncommon in patients with chronic rhinosinusitis with nasal polyps.

READING

0.5 HOURS

READING

0.5 HOURS

WHAT YOU NEED TO KNOW

•  Orofacial pain is not a hallmark feature of rhinosinusitis and affects only 10% of patients with rhinosinusitis

•  Consider alternative diagnoses such as migraine, midfacial segment pain, and cluster headaches, which can present with facial pain and nasal symptoms such as rhinorrhoea and nasal congestion

•  Offer a trial of treatment for 4 weeks for the most likely diagnosis and ask the patient to return if symptoms do not improve

READING

0. 5 H O U RS

LEARNINGMODULE

READING

0.5 HOURS

LEARNINGMODULE

READING

0.5 HOURS

LEARNINGMODULE

READING

LEARNINGMODULE

0.5 HOURS

See http://learning.bmj.com for linked learning module

HOW PATIENTS WERE INVOLVED IN THE CREATION OF THIS ARTICLEWe discussed the patient journey with several migraine patients in our ENT practice who had been misdiagnosed with chronic rhinosinusitis before attending their appointment. Despite their symptoms failing to improve with medical and surgical therapy, these patients were still convinced of the latter diagnosis. Patients shared the need for detailed communication explaining the diagnosis. We have emphasised this in the section on management.

P Box 1 | Symptoms of rhinosinusitis• Nasal blockage, congestion, obstruction—Essential for

diagnosis• Nasal discharge (anterior rhinorrhoea or post-nasal drip)—

Essential for diagnosis• Facial pain• Reduced sense of smell• Additional educational resources• Fokkens WJ, Lund VJ, Mullol J, et al. European Position Paper

on Rhinosinusitis and Nasal Polyps 2012. www.ep3os.org/EPOS2012.pdf.

• Pocket Guide. European Position Paper on Rhinosinusitis and Nasal Polyps. 2012. www.southernstatesrhinology.org/files/2013_SpeakerTalks/EPOSpocketguide2012.pdf.

• National Institute for Health and Care Excellence. Headaches in over 12s: diagnosis and management (clinical guideline CG150). 2015. www.nice.org.uk/guidance/cg150.

• National Institute for Health and Care Excellence. Temporomandibular disorders (TMDS). 2016. https://cks.nice.org.uk/temporomandibular-disorders-tmds.

the bmj | 19 May 2018 243

Disclaimer: This infographic is not a validated clinical decision aid. This information is provided without any representations, conditions, or warranties that it is accurate or up to date. BMJ and its licensors assume no responsibility for any aspect of treatment administered with the aid of this information. Any reliance placed on this information is strictly at the user's own risk. For the full disclaimer wording see BMJ's terms and conditions: http://www.bmj.com/company/legal-information/

Orofacial painVisual summary

Chronic orofacial pain (lasting more than 12 weeks) can be debilitating for patients. After this time, primary care treatments have often been exhausted, and referral is an obvious next step. However, many cases are incorrectly attributed to rhinosinusitis, which can lead to inappropriate referrals and delay for patients. A more accurate diagnosis may be performed by focusing on the characteristics and associated features of facial pain, as described below.

Identification and initial management Person withorofacial pain

Do they have nasal blockage or congestion, with or without nasal discharge?Yes

No

Do they have symptoms of skin or soft tissue tenderness or swelling?

Yes No

MigrainesRhinosinusitisMidfacial segment pain

Clusterheadache

Temporomandibulardisorder

Tension-typeheadache

NauseaPurulent nasaldischarge

Bilateral painUnilateral pain Clicky jawsNon pulsatile pain

No nausea

No vomiting

Bruxism

Jaw locking

Temporomandibularjoint laxity

Severe

Lacrimation

Conjunctivalinjection

Constant Photophobia

Phonophobia

Improves with rest

Patients with migraine commonly report pain in one half of the head, but may also report a facial location of pain

Inflammation of the lining of the nasal

cavity and the sinuses. Can be acute or

chronic, although chronic is more

commonly seen in clinical practice

Has similar characteristics to

tension-type headache. The

aetiology is unknown and it is often

considered as a diagnosis of exclusion

Can be initiated in primary care

if confident with diagnosis

Minimum treatment duration 6 months.

Wean off if no longer required

A diffuse headache described as a band

around the head. Usually a featureless

headache

Pain is short lived, lasting about 1 hour

and can occur in “clusters” of up to 8

times a day for weeks

Acute therapy

May also require

Affects the temporomandibular

joint and the masticatory muscles

Management is multifaceted

includingRefer immediately

to neurology

Consider referral if symptoms are refractory to community management

Address triggers by managing: Sleep Exercise Diet Caffeine Stress

Acute therapy Acute therapy

Preventive therapy

Preventive therapy

Aspirin

Propranolol

Topiramate

Ibuprofen

A triptan+/- paracetamol

+/- NSAID

ParacetamolIbuprofen

Candesartan

Tricyclicantidepressant

Flunarizine*

Acupuncture

Tricyclicantidepressant

Amitriptyline

Neurology; Ear nose and throat

Ear nose and throat

Neurology Primary care Oral and maxillofacial

Lifestyle changes

Relaxation techniques

Simple analgesia

Neuropathic agents

Specialist management, such as joint injections,

arthroscopies, arthroplasties, and joint replacement

Low dosebenzodiazepines

High flow oxygen

Subcutaneous triptan

Neuroimaging

Simpleanalgesics

Acute management with paracetamol,

aspirin and NSAIDs are not recommended

Low doseamitriptyline

Patients should be re-evaluated after 4

weeks, if their symptoms have improved then

treatment can be continued long term

Nasal steroids

Oral steroids

Nasal douching

A 10-day course can be initiated if nasal

polyps are clearly seen

* Fl

unar

izin

e is

not

lice

nsed

in th

e U

K

244 19 May 2018 | the bmj

The International Headache Society states: “chronic rhinosinusitis is not validated as a cause of headache or facial pain unless relapsing into an acute stage.”10 Patients with acute rhinosinusitis are more likely to present with orofacial pain caused by the irritation of sensory nerves by inflammatory mediators, pressure changes, and a blocked non-draining sinus.11 The pain is severe and usually unilateral. It is associated with fever and nasal obstruction, and, in the case of acute maxillary rhinosinusitis, can present with dental pain.12 True recurrent acute rhinosinusitis is rare, and patients will usually have a vascular aetiology for their pain such as migraines or cluster headaches.4

As some patients with chronic orofacial pain may report associated nasal symptoms, it can be difficult to assess if the pain is related to sinusitis, especially if the patient clearly points to the anatomical sites of the sinuses like their cheeks or forehead as the site of their pain (see infographic).13 Table 1 lists certain predictors that support or diminish the diagnosis of rhinosinusitis.4-11

MigraineChronic disabling headaches are more likely to be related to migraines than rhinosinusitis.14 EPOS states: “the majority of sinus headaches can actually be classified as migraines.”4 For this reason, patients who present in the community with “sinus headaches” should have a migraine diagnosis explored before considering rhinosinusitis.

A recent systematic review of community-based studies (over 6 million participants) reported a prevalence of 1 in 10.15 While patients with migraine commonly report pain in one half of the head, a subset of patients can have a facial location of pain,10 16 with otherwise characteristic symptoms of migraine such as nausea, vomiting, sensitivity to light and sound.

In a study of 517 patients with migraine, nearly 9% of patients had pain involving their head and lower half of the face,17 and these patients were also more likely to have associated trigemino-autonomic symptoms like rhinorrhoea and nasal blockage. Migraine is often misdiagnosed as sinus headache or rhinosinusitis due to associated symptoms such as nasal congestion, rhinorrhoea, or inflamed eyes or cheek with the migraine attacks.14-19 In a prospective cohort study of 2991 patients with a history of self described or physician diagnosed sinus headache, 88% of patients (2396) were diagnosed with migraines using the International Headache Society diagnostic criteria. Of these, over 80% of patients reported sinus pain or pressure, 63% reported nasal congestion, and 40% reported rhinorrhoea.19

Tension-type headachePatients commonly report symmetrical frontal or temporal headache4 described as a band around the head (see infographic). It has a lifetime prevalence of up to 78%, which suggests that most people would have experienced it at some point in their life.10 Like migraine, if it occurs more than 15 days in a month it is considered chronic. Unlike migraine, tension-type headache is usually a featureless headache. Some patients have pericranial tenderness in the frontal or temporal region, and in the masseter, pterygoid, sternocleidomastoid, splenius, and/or trapezius muscles.

Midfacial segment painThis condition is not well known in clinical practice despite affecting a third of patients presenting with facial pain to the ENT clinic.4 Midfacial segment pain has similar characteristics to those of tension-type headache and can be seen as a category of tension-type headache. The main distinguishing feature is the midface location of the pain in the forehead, peri-orbital region, retro-orbital region, cheeks, or nose (see infographic).20 The aetiology is unknown, and it is often a diagnosis of exclusion. Nasal blockage can be present, which may be associated with the condition or an incidental coexisting rhinitis.4 Symptoms are episodic and gradually become persistent. Tenderness and swelling over the cheeks and/or forehead and hyperaesthesia are common findings.

Cluster headacheCluster headache is characterised by attacks of severe unilateral pain in the orbital, supraorbital, and/or temporal region (see infographic). The hallmark feature is the severity of the pain. Patients are usually restless or agitated and unable to lie still. Pain is short lived, lasting about an hour (range 15-180 minutes) and can occur in “clusters” of up to eight times a day for weeks. It is accompanied by prominent cranial autonomic symptoms such as nasal congestion, eye watering, ptosis, and sweating.

National Institute for Health and Care Excellence (NICE) guidelines reiterate that these patients should not be given simple analgesics such as paracetamol, aspirin, or non-steroidal anti-inflammatory drugs (NSAIDs) during an acute attack.21 The International Headache Society recommends that patients should be managed by a neurologist who can prescribe high flow oxygen and subcutaneous triptan for acute therapy and request neuroimaging if required.10

Table 1 | Good and poor predictors of sinogenic facial painGood predictive value Poor predictive valueIncrease in severity from sitting to lying supine, or on flying or skiing

Increase in severity on bending forward

Reduced sense of smell Normal sense of smell

Improvement with antibiotic or corticosteroid treatment

No improvement with antibiotic or steroid treatment

Purulent, offensive nasal discharge Severe facial pain affecting activities of daily living

Worse with upper respiratory tract infection Tenderness or swelling in facial skin

EDUCATION INTO PRACTICE• How often do I diagnose a patient with rhinosinusitis based on orofacial pain alone?• Think of a patient with orofacial pain or sinus headache whom you have seen recently

in your practice. Based on reading this article, are there additional features you will look for on history and examination to consider alternative diagnoses?

• How will you explain to your patient about the possible causes of their pain and the need for a trial of treatment?

the bmj | 19 May 2018 245

Temporomandibular disordersTemporomandibular disorders are a group of musculoskeletal conditions affecting the temporomandibular joint and the masticatory muscles and are a common cause of chronic orofacial pain.21 Pain typically affects the pre-auricular region and can radiate around the ear to the cheek, temple, teeth, or jaw angle. An American study of 196 patients showed that referred pain in the cheek, forehead, and ear on palpating specific trigger points in the head and neck region is a common sign.22

Dental painDental pathology is an important cause of orofacial pain. In our experience, odontogenic pain usually does not cause diagnostic uncertainty as the patient accurately reports dental pain, has a history of poor dentition or previous dental work, and any current dental disease is evident on examination.

The characteristic features and site of pain can help distinguish these conditions on history and examination (see infographic).

Severe pain that interferes with a patient’s activities of daily living is more likely to be migraine or cluster headaches, unlike pain from temporomandibular disorder, midfacial segment pain, and tension-type headache. Associated nasal symptoms generally support rhinosinusitis but be aware of trigemino-autonomic symptoms in migraines and the possibility of a coincidental rhinitis given its high prevalence in the adult population. Ask for a family history or previous history of migraine.

In examination, palpate the skin overlying the sinuses. Tenderness over the sinuses should raise suspicion of midfacial segment pain. Palpate over and around the temporomandibular joint for tenderness going as far superiorly as the temple, inferiorly at the jaw line, posteriorly over the mastoid, and anteriorly over the cheek, as trigger points for temporomandibular disorders lie within this area. While palpating the temporomandibular joint, ask the patient to open and close their mouth and feel for jaw clicking and laxity. Laxity refers to increased joint mobility and protrusion of the joint laterally on mouth opening. Laxity and jaw clicking are signs of abnormal temporomandibular joint. Anterior rhinoscopy can be performed with an otoscope to look for mucopus in the nasal cavities (which supports a diagnosis of rhinosinusitis) or enlarged inferior turbinates (which suggest rhinitis). We recommend a full cranial nerve examination in all patients with chronic orofacial pain to exclude rare intracranial tumours that can present with orofacial pain.

Investigations are not usually required to confirm the diagnosis.

How is it managed?

Because of the overlapping symptoms, often two or more diagnoses are possible. Offer the patient a trial of treatment for the most likely diagnosis for four weeks. If there is no improvement, discontinue it and consider an alternative diagnosis and treatment. It is important that patients are not dismissed when they express doubts in the diagnosis. Explain to your patient the complex nature of orofacial pain that can cause diagnostic difficulties. Patients may find it reassuring to know that this is not uncommon. Emphasise the need to follow up if symptoms do not improve.

The symptoms mentioned in box 1 are sufficient to support a diagnosis of rhinosinusitis and to start treatment in primary care. The management for chronic rhinosinusitis depends on the severity of symptoms and the presence of polyps. The EPOS guidelines recommend offering a nasal steroid spray and nasal douching.23 Re-evaluate the patient at four weeks for relief of symptoms. If symptoms have improved then treatment can be continued long term.

We recommend following relevant guidelines for the management of migraine and tension-type headache. Advise patients to follow a daily routine with regular sleep, exercise, diet, and periods of relaxation and to reduce caffeine intake. Patients with tension-type headache can be offered paracetamol 1 g or ibuprofen 400 mg for an acute episode.

If mid-facial segment pain is the likely diagnosis, low dose amitriptyline (10 mg at night for at least six weeks) is recommended.20 This should be continued for a minimum of six months, and the patient then weaned off if it is no longer required.20

The management of temporomandibular disorder is multifaceted, including lifestyle changes, relaxation techniques, simple analgesia such as paracetamol or NSAID, low dose benzodiazepines, and neuropathic agents.24 Specialist management involves injections into the temporomandibular joint, arthroscopies, arthroplasties, and joint replacement surgery.

In most cases a referral is only required when the patient’s symptoms are refractory to medical therapy started in the community or there are associated red flag symptoms. Cluster headache is the only condition that warrants immediate referral to neurology because of the potential need for neuroimaging and specialist treatment.

If the diagnosis is in doubt, or in the presence of red flag signs—such as cranial neuropathies, dysphagia, dysphonia, or neck lump—arrange an urgent review with an ear, nose, and throat (ENT) specialist in secondary care. The diagnosis of rhinosinusitis is confirmed with the evidence of polyps, discharge, or oedema on anterior rhinoscopy or endoscopy, or sinus features of mucosal changes on computed tomography. Sinus computed tomograms (CTs) are usually performed in patients where sinus surgery is being considered when medical therapy has been unsuccessful.Cite this as: BMJ 2018;361:k1517Find the full version with references at http://dx.doi.org: 10.1136/bmj.k1517

SOURCES AND SELECTION CRITERIAWe searched PubMed using the keywords “facial pain,” “atypical facial pain,” “facial neuralgia.” All articles published in English that discussed the differential diagnoses and management of facial pain were reviewed. Articles on specialist treatment of one type of facial pain were excluded as we agreed to present an approach to orofacial pain for non-specialists and primary care doctors.

When to refer

How to make a diagnosis

246 19 May 2018 | the bmj

UNCERTAINTIES

Should patients with ductal carcinoma in situ be treated with adjuvant whole breast radiotherapy after breast conservation surgery?Jessamy Bagenal,1 Nicola Roche,1 Gill Ross,1 Anna Kirby,1 David Dodwell2

1The Royal Marsden Hospital, London, UK2Nuffield Department of Population Health, University of Oxford, Oxford, UKCorrespondence to: J Bagenal jessamy.bagenal@googlemail.com

Ductal carcinoma in situ (DCIS) affects around 8000 women a year in the UK.1 Since the introduction of mammographic screening, the incidence of DCIS has increased and it now represents around 20% of all new screen detected breast cancers.2

DCIS is categorised into low, intermediate, and high grade based on histological features. Most cases of DCIS are treated with breast conserving surgery (BCS), often followed by whole breast radiotherapy (WBRT). An individual patient level meta-analysis (four randomised controlled trials, 3729 women) found that WBRT approximately halved the rate of ipsilateral DCIS or invasive recurrence at 10 years compared with no radiotherapy following BCS.3 However, WBRT can cause side effects such as impaired cosmesis, skin changes, and late cardiac toxicity4 Patients might also find WBRT inconvenient and expensive.

National Institute for Health and Care Excellence (NICE) guidelines recommend offering WBRT to all patients with DCIS treated by BCS.5 The European Society of Medical Oncology guidelines suggest that WBRT might be omitted in some low risk patients.6

However, observational studies done in the UK, US, and Europe note wide variations in the use of adjuvant radiotherapy in these patients.7 This variation possibly reflects uncertainty as to whether the benefits of WBRT are large enough to warrant the blanket use of adjuvant WBRT or whether WBRT can be safely omitted in a subset of lower risk patients.

In this article we discuss the evidence surrounding radiotherapy use in DCIS.

READING

0.5 HOURS

READING

0.5 HOURS

WHAT YOU NEED TO KNOW

•  Women with ductal carcinoma in situ are usually offered breast conserving surgery (BCS), often followed by whole breast radiotherapy (WBRT).

•  WBRT reduces ipsilateral breast events but there is insufficient evidence that it improves breast cancer specific or overall mortality.

•  Patients will have different perceptions of the risks and benefits of WBRT and should be assisted in their decision making by clear presentation of the information.

HOW PATIENTS WERE INVOLVED IN THE CREATION OF THIS ARTICLEA draft manuscript was reviewed by representatives of ICPV (http://www.independentcancerpatientsvoice.org.uk/). They were happy with the overall manuscript but had specific statements about what a patient can expect from their doctor and these were addressed in the “what patients need to know” box. A patient with DCIS kindly reviewed this paper. She endorsed the uncertainty around appropriate treatment for DCIS and that medical professionals must recognise that patients have different personal risk tolerances. We have emphasised this and suggest that doctors discuss the risk and benefits of WBRT and assist patients in making a shared decision.

P

Light micrograph of tissue from an

affected breast

the bmj | 19 May 2018 247

What is the evidence of uncertainty?What is the risk of an ipsilateral breast event in a DCIS patient?Evidence from three prospective studies in patients with DCIS who have undergone BCS suggests that the risk of an ipsilateral breast event continues to rise with time.8-11 These studies failed to identify a sufficiently low risk group of patients that gain no appreciable benefit in terms of a reduced risk of recurrence from WBRT. In a trial with 636 women with low or intermediate grade DCIS, at a median follow-up of seven years, the ipsilateral event rate was 0.9% in the WBRT arm versus 6.7% in the observation arm (hazard ratio, 0.11; 95% confidence interval, 0.03 to 0.47; P<.001).9 A single-arm trial in 158 women with low and intermediate grade DCIS closed early as the ipsilateral breast event rate of 2.4% per patient-year met the predetermined stopping rules.10 In another prospective non-randomised study, the risk of an ipsilateral breast event at 12 years was 14.4% for the low-intermediate grade group (561 women) and 24.6% for the high grade group (104 women).11 Patients might have differing opinions on what constitutes an acceptable risk of an ipsilateral breast event.

Scoring systems12-14 have been developed to predict the risk of local recurrence in an attempt to identify patients at low risk of recurrence who may be able to avoid WBRT and thereby guide adjuvant therapy recommendations (table 1, see bmj.com). However, these scores are used variably in practice as they do not provide sufficiently precise estimates.

Table 3 | New trialsTrial name Location Design Population Intervention Comparator Outcome Expected resultsLow riskDCIS LORD

International, multicentre

Prospective open label randomised phase III non-inferiority trial

1240 women >45 with screen detected asymptomatic pure low grade DCIS on vacuum biopsy

Active surveillance Standard treatment for DCIS

10 year ipsilateral breast cancer free percentage

2029

Low riskDCIS TrialLORIS

Multi centre UK based

Prospective randomised controlled phase III non-inferiority trial

932 >46 screen detected asymptomatic pure non-high grade DCIS

Active surveillance Surgery 10 year ipsilateral invasive breast cancer-free survival time

2024

Table 4 | Clinical pathological features found to be associated with recurrenceFeature EvidenceYoung age High quality: analysis from one randomised trial and large well conducted

observational studies19-21

Symptomatic presentation

High quality: analysis from two randomised trials1 22

Family history Low quality: small retrospective cohort study23

Multifocality High/moderate quality: large retrospective cohort studies24 and central pathology review of patients entered into randomised trial25

Size High/moderate quality: large retrospective cohort studies26 and central pathology review of patients entered into randomised trial25

Margin status High quality: analysis from three randomised trials and retrospective study of pathological samples1 27 28

Volume of disease at closest margin

Moderate quality–one retrospective cohort study29

Nuclear grade High quality: analysis from randomised trial1 and case cohort study within a randomised trial30

Presence of comedo necrosis

High quality: analysis from randomised trial31 case cohort study within a randomised trial30 and central pathology review of patients entered into randomised trial9

Architectural pattern

High quality: analysis from randomised trial1 and central pathology review of patients entered into randomised trial25

WHAT PATIENTS NEED TO KNOW

•  Ductal carcinoma in situ very rarely leads to death.

•  Most DCIS is treated with breast conserving surgery (BCS), often followed by whole breast radiotherapy (WBRT).

•  WBRT reduces the risk of recurrence, but WBRT does not reduce the risk of dying from breast cancer or improve overall survival.

•  You should expect your doctor to advise you on the likely benefits and risks of radiotherapy, taking into consideration the severity of your disease and any other health problems, and accepting that there is some uncertainty about the benefits of treatment.

•  Because of this uncertainty, you might wish to explore what type of treatment is best for you based on your personal risk tolerance and life expectancy.

STEV

E G

SCH

MEI

SSN

ER/S

PL

248 19 May 2018 | the bmj

Does WBRT reduce overall or breast cancer specific mortality?Although WBRT has been shown to decrease the risk of within-breast recurrence by around 50%, it has not been shown to improve breast cancer mortality.15 16 Five randomised controlled trials in women with DCIS showed that WBRT did not influence breast cancer mortality or overall survival (table 2, see bmj.com). The EBCTCG patient-level meta-analysis reported a 10 year breast cancer mortality of around 4%3 and therefore a very large study would be required to show a modest reduction in breast cancer mortality from WBRT. Similarly, recent large population based longitudinal studies find insufficient benefit of WBRT in reducing breast cancer specific mortality at 10 and 20 years to warrant its use in all patients.16 17

A recent study analysed 10 and 20 year breast cancer specific mortality using the Surveillance Epidemiology and End Results (SEER) data (18 registries, 108 196 women with DCIS). Invasive recurrence increased the risk of dying from breast cancer (hazard ratio 18.1, 95% confidence interval 14.0-23.6; P<0.001), but the prevention of recurrence by WBRT did not diminish breast cancer specific mortality at 10 years.16 It is not possible to identify and account for the potential confounding factors in a study of this type, but these results suggest that the increased mortality risk associated with invasive recurrence is unlikely to be great enough to warrant WBRT.

Another large population based longitudinal study also used SEER data17 to study breast cancer specific mortality after BCS alone and in patients who received BCS and WBRT. The study used a patient prognostic scoring model comprising clinical and pathological features for risk stratification and propensity scoring to address possible confounding. In this cohort of 32 144 women, breast cancer mortality rate was 0.9%. The 10 year breast cancer mortality rate was 1.8% in the WBRT group and 2.1% in the non-WBRT group (absolute difference, 0.3%; log-rank test, P=0.003; hazard ratio, 0.73; 95% confidence interval, 0.62 to 0.88). However, the hazard ratios depicting the apparent effect of WBRT for each

of the defined prognostic groups had wide confidence intervals, and a causal relationship between WBRT and reduced breast cancer mortality cannot be confirmed.

Any small benefit derived from WBRT in breast cancer specific mortality, if this exists, must be appreciated in the context that patients with screen detected DCIS are much more likely to die of other causes than those related to breast cancer.17

Is ongoing research likely to provide relevant evidence?We searched clinical trial registries and found three large studies that might shed light on some of the uncertainty around the use of radiotherapy in patients with DCIS. The two largest trials aim to identify low risk DCIS patients based on clinical and pathological factors and compare active surveillance with standard therapy (breast conserving surgery and WBRT) using 10 year rate of invasive local recurrence as an endpoint18 (table 3). These studies should help to elucidate the natural history of DCIS and how to identify patients with a low risk of invasive recurrence.

What to do in the light of the uncertainty•  The benefit of WBRT in DCIS in many cases is

small, and patients who wish to avoid WBRT can be supported and reassured in this decision.

•  Discuss the uncertainty over the risks and benefits of WBRT treatment with patients fully.

•  It is likely that a subset of patients does not need WBRT, but we have yet to develop the best way to identify this group.6

•  A reasonable approach is to risk stratify patients using one of the scoring systems in table 1 (see bmj.com). Table 4 lists clinical and pathological features shown to be associated with recurrence. Where possible, use shared decision making aids to help the conversation with patients.32 Offer counselling with specialist nurses and provide information leaflets that display all treatment options.

Competing interests: We have read and agreed to the BMJ's conflict of interest policy. JB was a clinical editor at BMJ from 2015 to 2017.Cite this as: BMJ 2018;361:k1410Find the full version with references at http://dx.doi.org/10.1136/bmj.k1410

EDUCATION INTO PRACTICE• How has

reading this article changed the way you might approach discussions of radiotherapy with patients with DCIS?

• How might you better support patients who have received a diagnosis of DCIS?

RECOMMENDATIONS FOR FUTURE RESEARCH• Development of better predictive markers/tools to identify a group of patients in

whom WBRT can be safely omitted• Explore how current evidence is understood by clinicians and delivered to patients to

reduce geographical variations in the proportion of patients being offered WBRT• Explore how patients feel about making decisions around WBRT

SEARCH STRATEGYWe searched the Cochrane Library (including the Cochrane Central Database of Controlled Trials), Ovid Medline, and clinical trial registers (clinicaltrials.gov, controlled-trials.com, who.int/trialsearch), PROSPERO, National Cancer Research Institute portfolio, Cancer Research UK, and Macmillan websites from 1990 until May 2017. We also cross referenced bibliographies. We ran multiple searches using the terms:

“Ductal carcincoma,” “DCIS” AND “RADIOTHERAPY,” “DCIS” AND “Treatment” in combination and alone. We selected manuscripts and trials that were most relevant to the article through discussion between the authors.

Light micrograph of a section from

an affected breast

STEV

E G

SCH

MEI

SSN

ER/S

PL

the bmj | 19 May 2018 249

CASE REVIEWSubtle skin changes that suggest severe diseaseA 53 year old woman presented to the dermatology department with insect bites to the dorsum of the hands. On examination, yellow skin lesions were found on both sides of her neck (fig 1). The lesions, which were confluent papules 2-4 mm in diameter, had appeared several years before. The woman had no other symptoms.

She was taking duloxetine and clonazepam for depression, and dexketoprofen, calcium, and vitamin D supplements for chronic cervical spine pain.

The patient did not recall any other similar cases of skin lesions in her family.1 What condition has caused the coalescent yellow

papules?2 What regions of the body are predominantly affected by

this disease?3 How should this patient be managed?Submitted by Juan Antonio Moreno Romero, Alba Alvarez Abella, Daniel Lorenzo, and Ramon Grimalt

Patient consent obtained.

Cite this as: BMJ 2018;361:k1336

ENDGAMES For long answers go to the Education channel on bmj.com

answ

ers

If you would like to write a Case Review for Endgames, please see our author guidelines at http://bit.ly/29HCBAL and submit online at http://bit.ly/29yyGSx

Fig 1

Fig 2

You can record CPD points for reading any article. We suggest half an hour to read and reflect on each.

Articles with a “learning module” logo have a linked BMJ Learning module at http://learning.bmj.com.READING

0. 5 H O U RS

LEARNINGMODULE

READING

0.5 HOURS

LEARNINGMODULE

READING

0.5 HOURS

LEARNINGMODULE

READING

LEARNINGMODULE

0.5 HOURSREADING

0.5 HOURS

READING

0.5 HOURS

CASE REVIEWSubtle skin changes that suggest severe disease

1 Pseudoxanthoma elasticum (PXE), a genetic disease causing yellowish papular lesions and redundant folds in flexural areas. Primary PXE skin lesions are yellowish papules 1 to 5 mm in diameter. They tend gradually to coalesce to form plaques, which have a cobblestone appearance.

2 PXE mainly affects the cutaneous, ocular, and cardiovascular systems. PXE has an estimated prevalence of 1:25 000-100 000. The characteristic ocular defects of PXE are angioid streaks of the retina, characterised by reddish brown curvilinear bands that radiate from the optic disk (arrows, fig 2).

3 There is no specific treatment for PXE. Management involves prevention and monitoring of complications and involves the multidisciplinary team. Early recognition and lifestyle adjustments are important to reduce morbidity. Genetic counselling might be helpful and PXE support groups can be valuable.

Occupational hazards of professional footballersThe prevalence of knee pain and radiographic evidence of osteoarthritis is two to three times higher among retired professional footballers than in the male population generally, according to a survey in the British Journal of Sports Medicine. Footballers were also more likely to have had a total knee replacement. These differences persisted after adjusting for a history of knee injury, which might suggest that repeated episodes of minor trauma end up being more destructive than a single severe event. Perhaps the surprising thing is that the excess of knee symptoms found in footballers isn’t a lot larger.

Fish oil and the brainFish oils appear to have been discredited as a preventive treatment for coronary heart disease. The idea that they can slow decline in cognitive function seems to be going the same way, despite the fact that polyunsaturated fatty acids are essential constituents of brain phospholipids. A randomised controlled trial in Australia tested supplements of docosahexanoic and eicosapentaenoic acids against supplements of olive oil over an 18 month period in community-living adults aged 65 to 90. It found no differences between the two groups in tests of higher mental function (Am J Clin Nutr).

Breast feeding and intelligenceA large randomised trial of an intervention to promote breastfeeding took place in the Republic of Belarus in the late 1990s. The investigators reported that children in the intervention group showed better neurocognitive development when tested at the age of six. But longer term follow-up finds that these benefits aren’t sustained (PLoS Med). At age 16, children in the intervention group scored no more highly than children in the control group in a range of cognitive tests. A subgroup of children who had been exclusively breast fed for 3 months showed a small increase in verbal function but not in any other domains.

Funding for cancer researchAn analysis of public and philanthropic funding for cancer research in the UK between 2000 and 2013 shows that 69% of grants were awarded to male principal investigators (BMJ Open). What’s more, male principal investigators tended to receive larger grants than female principal investigators. The reasons underlying this gender difference aren’t clear, but there’s no doubt that female cancer researchers consistently

receive less funding than their male counterparts in terms of total investment, number of awards, and average value of awards.

Nut consumption and cardiovascular diseaseIf you can persuade a large number of people to complete a food frequency questionnaire and then observe them over a long period, you’re almost certain to find positive associations between some quirk of diet and health outcomes. Nut consumption, for example, turned out to be inversely associated with several cardiovascular conditions including myocardial infarction, heart failure, atrial fibrillation, and abdominal aortic aneurysm in a study of 60 000 Swedish adults followed for 17 years (Heart). However, adjustment for likely confounders weakened these associations, which probably means that the sort of people who eat nuts do many other things that reduce their risk of vascular disease.Cite this as: BMJ 2018;361:k2072

MINERVA A wry look at the world of research

An 88 year old woman presented with progressive stridor, dysphonia, and dysphagia. She had severe osteoporosis, with loss of height and thoracic kyphosis. Fibreoptic laryngoscopy showed a right vocal cord palsy. Differential diagnosis included laryngeal malignancy, and laryngeal nerve compression. Contrast computed tomography revealed narrowing of the thoracic inlet anatomy caused by thoracic kyphosis (figure). Distortion of the cervical spine had caused compression of the trachea between the right brachiocephalic artery and vertebral column, leading to pressure on the recurrent laryngeal nerve. Tracheostomy was not possible because of the level of obstruction, and

the patient declined the option of a radiologically inserted gastrostomy feeding tube. Management was conservative.

This is a rare cause of stridor and unilateral vocal cord palsy, but could become a more important differential diagnosis given the rising prevalence of osteoporosis in an ageing population.Vanushia Thirumal (vanushia@gmail.com); Kim To; Athena Togo, Department of Otolaryngology, Head & Neck Surgery, Raigmore Hospital, Inverness, Scotland, UK; Jonathan Brodie, Department of Radiology, Raigmore Hospital, Inverness, Scotland, UKPatient consent obtainedCite this as: BMJ 2018;361:k1905

Severe osteoporosis and stridor

250 19 May 2018 | the bmj