Article Text

Contemporary management of calcified coronary lesions
  1. Emanuele Barbato1,2
  1. 1 Cardiovascular Research Center, OLV Hospital, Aalst, Belgium
  2. 2 Department of Advanced Biomedical Sciences, Universita degli Studi di Napoli Federico II, Napoli, Italy
  1. Correspondence to Prof. Emanuele Barbato, Cardiovascular Research Center, OLV Hospital, Aalst 9300, Belgium; barba22{at}

Statistics from

Learning objectives

  • To understand the clinical relevance of calcified coronary lesions.

  • To identify the most reasonable management strategy of patients with calcified stenoses.

  • To learn the tools and techniques facilitating safe percutaneous treatment in this challenging setting.


Management of patients with calcified coronary lesions represents one of the last unmet clinical needs in interventional cardiology. The age of patients referred for invasive coronary angiography with indication to revascularisation is steadily increasing.1 This leads to a higher proportion of patients with complex coronary artery disease (CAD) and with heavily calcified coronary stenosis.2 Beyond the mere epidemiological finding, the presence of calcified coronary stenosis has a significant clinical impact. In fact, the presence of moderate-to-severe calcification portends to increased cardiovascular mortality and major adverse cardiovascular events in patients treated with percutaneous revascularisation.3 4 Severe coronary calcification is an independent predictor of 1-year stent thrombosis and target lesion revascularisation.3 An increased cardiovascular mortality has been reported also for patients with heavily calcified coronary stenosis treated with surgical revascularisation as compared with patients without calcified coronary arteries.5 This can be partly explained by the fact that complete revascularisation is seldom achieved in patients with heavily calcified coronary arteries, given the additional challenge in performing the distal anastomosis of the graft to the target coronary artery.6 Advanced age, the possible association of porcelain aorta and the frequent presence of comorbidities are additional factors considered within the heart team when selecting percutaneous coronary intervention (PCI) as the revascularisation option in these patients. This article, mainly directed to trainees and general cardiologists, focuses on the management of patients with heavily calcified coronary stenoses undergoing PCIs.

Diagnostic challenges of coronary calcifications

One of the first challenges represented by calcified coronary stenosis is that they are often underdiagnosed with routine coronary angiography. Operators report moderate-to-severe calcifications in about 10% of coronary angiograms.7–9 When coronary angiograms are more carefully evaluated, for example, by central core laboratories, this rate increases up to 30%–40%.10 Yet, angiographic estimation of coronary calcification is still suboptimal as compared with that of intravascular imaging like intravascular ultrasound (IVUS) or optical coherence tomography (OCT), where the rate of calcium detection goes up to nearly 80%.11 12

Why identifying coronary calcifications is important? First, crossing calcified lesions might be difficult with regular balloons and sometimes even with regular guidewires. Calcified lesion is the most important determinant of reduced coronary compliance, resulting into failure of balloons to adequately dilate these lesions.12 Second, calcified stenoses are associated with failure to deliver a stent, possible polymer disruption of the drug-eluting stent (DES) implanted and stent underexpansion.13–15 Stent asymmetry and malapposition are also frequent findings in calcified stenoses and might portend to increased target vessel failure and adverse clinical outcomes.16 Therefore, early identification of moderate-to-severe coronary calcification is of paramount importance to anticipate procedural challenges and to predispose optimal lesion preparation and stent implantation.

Practical tips and algorithms, based on non-randomised clinical data and operator’s experiences, might help in anticipating the severity of calcification triggering the need of specific tools and/or techniques aiming at plaque modification to facilitate optimal PCI (figure 1). Ideally, moderate-to-severe calcified coronary stenoses should be identified before embarking into PCI. Before considering the diagnostic invasive angiography, useful hints that might raise suspicion of calcified CAD are provided by several clinical predictors and by coronary CT angiography (CCTA). In a large registry of more than 220 000 PCIs, clinical predictors of moderate-to-severe calcified coronary stenoses requiring rotational atherectomy (RA) were advanced age, diabetes, hypertension, peripheral vascular disease, reduced left ventricular function, previous coronary artery bypass graft and high syntax score.17 CCTA is being increasingly performed before coronary angiogram and might easily disclose and characterise severe calcified coronary lesions. In fact, independent predictors of calcified stenoses requiring RA were a lesion calcium score of ≥450 Hounsfield units (HU), a lesion length of >20 mm and a CCTA scoring grade of 5 in calcified plaque (figure 2 and table 1).18

Figure 1

Practical algorithm for considering plaque modification techniques. Plaque modification should be considered in the presence of one or more of the clinical, angiographic or procedural predictors of severe calcified/fibrotic coronary stenoses. *Liberal use of IVUS or OCT is advisable in the presence of moderate coronary calcification at the CAG. CABG, coronary artery bypass graft; CAG, coronary angiography; CCTA, coronary CT angiography; IVUS, intravascular ultrasound; LVEF, left ventricular ejection fraction; OCT, optical coherence tomography; PCI, percutaneous coronary intervention; PVD, peripheral vascular disease.

Figure 2

Calcium severity based on CCTA.18 Classification based on calcium severity on cross sections by CCTA and on the number of quadrants involving calcium: grade 1, non-calcified (no calcification);) grade 2, single calcification with an arc of <90°; grade 3, calcification arc≥90° and <180° (a) or multiple calcifications with visible residual lumen (b); grade 4, calcification arc≥180° and <270°; and grade 5, calcification arc ≥270° (a), calcification arc=360° with visible residual lumen (b) or without visible lumen (c). CCTA, coronary CT angiography.

Table 1

Coronary calcification severity according to the imaging technique

In the catheterisation laboratory, plaque modification techniques should be considered in the presence of severe angiographic coronary calcification (table 1). Intravascular imaging should be liberally adopted if high suspicion of moderate calcified lesions is raised at the coronary angiogram.15 19 A practical calcium score based on OCT has been proposed to predict stent underexpansion (figure 3).20 Adequate stent expansion is achieved with a score of up to 3, whereas calcified stenoses with a score of 4 had poor stent expansion and therefore might require appropriate plaque modification. At times, the IVUS or OCT probes fail to cross the calcified lesions. This latter might be another practical hint of limited vessel compliance; hence, some experienced operators are keeping the bar very low to use RA in these situations.

Figure 3

OCT-based calcium score.20 OCT-based calcium score based on three parameters derived from prestent OCT: (1) maximum calcium angle, (2) maximum thickness and (3) calcium length. One or two points is assigned for each parameter, and the total score (0–4) calculated. A score of 4 indicates the need for adequate plaque modification. OCT, optical coherence tomography.

During the course of PCI, additional predictors of the need for more aggressive plaque modification are derived from randomised studies comparing conventional PCI with RA in heavily calcified stenoses.21 22 Failure of the balloon or of the stent to cross the stenosis and suboptimal balloon expansion represent the main factors determining the crossover to RA in 10%–15% of the patients initially allotted to the conventional PCI strategy. Of importance, a balloon:artery ratio of 1 should be used to confirm a suboptimal expansion. This latter should be best attempted with a non-compliant balloon to minimise the risk of vessel perforation. To summarise (figure 1), it is best to consider plaque modification in the presence of

  • Clinical predictors of heavily calcified lesions.

  • High calcification degree on CCTA.

  • Severe coronary calcification as identified both on coronary angiography or at intravascular imaging (table 1).

  • Failure to cross or to dilate the calcified lesion.

Treatment challenges of calcified coronary stenosis

A potentially straightforward procedure can unexpectedly turn into a nightmare of complications when procedural setting and strategy have not been carefully prepared in order to face the multiple options offered by the available tools and techniques for optimal lesion preparation. With moderate-to-severe calcified coronary stenosis, the experienced operator might even decide to attempt conventional balloon PCI first, but having everything ready to escalate to higher procedural complexity.

Procedural setting in case of PCI of heavily calcified coronary stenosis cannot be the one routinely used. These are often time-consuming procedures that are best rescheduled electively if the patient is clinically stable and in case of a busy schedule in the catheterisation laboratory. This latter option offers also the advantage to calmly review the case within the team, to make sure that all necessary tools are available on the shelf, and that other experienced colleagues might be available to support as second operator or as backup.

PCI can be performed both from radial or femoral artery, though 5 F guiding catheters should not be used in calcified coronary stenosis. Even though contemporary RA is often performed with 6 F guiding catheters (that can accommodate up to 1.75 mm burr but with some friction), 7 F guiding catheters (or sheathless guiding if desirable in case of radial) provide stronger back-up support and facilitate the use of more bulky devices.23 Choice of guiding catheter type normally follows the vessel anatomy, even if in calcified stenosis single-curve guiding catheters (eg, extra back up (EBU, XB) or Judkins catheters) are associated with less friction and resistance to the passage of large-bore devices. Double-curve catheters (eg, Amplatz 0.75 or 1.0) provide the best support for stenosis located in the right coronary artery. Additional back-up support can be provided by guiding extension catheters (1.25 mm burr can be accommodated in a 6 F extension catheter with some friction, while a 1.5 would definitely require a 7 F extension catheter).

Hydrophilic guidewires might be helpful in heavily calcified lesions that can be difficult to cross with workhorse guidewires. In addition, preference should go to more supportive guidewires that can facilitate crossing the lesion with balloons and other devices. As a general rule, one should have in place the best back-up support in order to make sure that failure to cross the lesion really depends on the highly resistant lesion.

Finally, procedural antithrombotic therapy in patients with heavily calcified stenosis undergoing PCI should follow the clinical presentation of the patient. In general, unfractionated heparin is administered at a dose to target an activated clorring time of 250–300 s. Dual antiplatelet therapy is composed of aspirin (low dose) and clopidogrel or ticagrelor/prasugrel/cangrelor, depending on whether the patient is stable or presenting with acute coronary syndromes. In general, IIbIIIa inhibitors are best avoided electively in patients with heavily calcified stenoses, and if needed can be considered as a bailout option (eg, in case of no reflow).

Plaque modification and lesion preparation

Contemporary PCI in heavily calcified stenosis aims at plaque modification and adequate lesion preparation for optimal stent implantation. Plaque modification is achieved by creating multiple fractures in the calcified lesion, therefore changing the vascular compliance and increasing the likelihood of maximal luminal gain, in order to facilitate complete and circumferential stent expansion. Plaque modification is achieved by the means of multiple adjunctive tools such as, but not limited to, (1) specialty balloons (cutting or scoring balloons); (2) atherectomy devices (rotational, orbital and laser); and (3) intravascular lithotripsy (IVL). In contrast to what is often believed, a decrease in procedural and fluoroscopy times, as well as reduced resource consumption, is observed when plaque modification is attempted in the first place, rather than in a bailout situation.24

Specialty balloons

Specialty balloons have been developed to address some limitations of plain old balloon angioplasty (POBA). Lesion dilatation with POBA occurs by applying radial forces in random directions, resulting in multiple intima–media dissections very often in vascular segments of lower resistance. In the presence of highly fibrotic or calcified vascular segments, POBA results in an asymmetric balloon expansion in the direction of the most compliant vascular segments opposite to the calcified segments. This is the reason why non-compliant balloons should be used to prevent rupture at the vessel side of lower resistance. In fact, non-compliant balloons will only have a minimal increase in diameter when inflated beyond the nominal pressure as compared with the semicompliant balloon. In some cases, the balloon might also appear angiographically fully expanded in one angiographic view, giving the false impression of a successful balloon dilatation. Nevertheless, an asymmetrical balloon dilatation is often associated with eccentric and frequently malapposed stents.

The cutting balloon (Wolverine; Boston Scientific, Marlborough, Massachusetts, USA) is a non-compliant balloon catheter with three to four microblades mounted longitudinally on the balloon surface.25 The cutting balloon is inflated slowly (with a progressive increase of 2 atmospheres) until the nominal pressure to allow deep radial incision into the fibrocalcific stenosis, resulting in larger luminal gain as compared with conventional POBA.26 27 Although improved with the latest-generation Wolverine, the drawback of the cutting balloon is in the stiffness and in the crossing profile that might hamper its delivery, especially in tortuous coronary anatomy and distal lesion location.

The scoring balloons (AngioSculpt; Spectranetics, Colorado Springs, Colorado, USA, and ScoreFlex, OrbusNeich, Hong Kong) are semicompliant balloons surrounded by external nitinol spiral scoring wires. These devices have smaller crossing profiles and more flexibility than the cutting balloon. In 299 patients undergoing IVUS-guided PCI, DES implantation after scoring balloon was associated with better stent expansion compared with direct stenting or with DES implantation after POBA.28 In an observational series of patients mostly with calcified coronary stenosis, the scoring balloon was associated with 98% procedural success rate and limited rate of complications (mostly dissections).29 30

There are no randomised studies comparing cutting and scoring balloons. Based on operators’ experiences, cutting balloons could be used in proximal lesions and in-stent restenosis, while they should be avoided in tortuous vascular segments and tight stenosis. Scoring balloons could be preferred instead in tortuous segments and tight stenosis or more distal lesion location. After angioplasty with cutting or scoring balloon, further inflations with slightly larger non-compliant balloons could be considered both to confirm effective plaque modification and to increase lumen enlargement.

Atherectomy devices

The available atherectomy devices consist of RA (introduced in 1988),31 orbital atherectomy (OA, introduced in 2013) and excimer laser coronary atherectomy (ELCA, introduced in 1991).32 Both RA and OA apply the same principle of the differential cutting or sanding, respectively. Practically, they operate through a mechanical ablation of inelastic non-compliant fibrocalcific plaques while sparing adjacent elastic tissue that deflects away from the ablating burr. In addition, both RA and OA share the principle of orthogonal displacement of frictions that enable such bulky devices to easily advance through the most challenging inelastic vascular segments, minimising the vascular injury and reducing the rate of dissections.

RA is performed by using the Rotational Atherectomy System or Rotablator (RotaPro; Boston Scientific, Massachusetts, USA), which is composed of an advancer, a console and a burr. The RotaPro represents the recent iteration of the previous Rotablator system that has significantly simplified its use. More details are provided in the online supplemental material.

Supplemental material

Contemporary RA technique has been described in an European Association of Percutaneaous Cardiovascular Interventions consensus document and reflects previous large experience accumulated over the years.33 Choice of the burr size follows a burr:artery ratio of less than 0.7, and in general, a single 1.5 mm burr is recommended to tackle most of the lesions, being able to achieve good plaque modification without exceeding in resource consumption. A clinical case of RA is shown in figure 4. In the Study to determine Rotablator And Transluminal Angioplasty Strategy (STRATAS) and Coronary Angioplasty and Rotablator Atherectomy Trial (CARAT), an RA strategy aiming at a burr:artery ratio of <0.7 was associated with similar efficacy (in terms of acute angiographic results and 6 months clinical outcomes) and less procedural complications (eg, perforation, no-reflow and large dissection) compared with a burr:artery ratio of >0.7.34 35

Figure 4

Clinical case of RA. Heavily calcified stenosis of the mid-right coronary artery at baseline (A) and after percutaneous coronary intervention with RA (1.25 and 1.5 mm burrs) and drug-eluting stent implantation (B). OCT catheter could not cross the tightest part of the lesion. At baseline, a wedging OCT catheter shows heavy calcification involving three vascular quadrants (C). (D) OCT result after 1.25 and 1.5 mm burr ablation. (D) Final stenting OCT result. Yellow arrows indicate the lesion site before (A) and after treatment (B). OCT, optical coherence tomography; RA, rotational atherectomy.

The San Antonio Rotablator Study (SARS) has also prompted the adoption of a lower ablation speed in the range of 135 000–180 000 rpm.36 Importantly, burr spinning deceleration of >3000 rpm should be avoided to reduce the risk of burr lodging. The latter is also achieved by a slow forward–backward movement of the burr (pecking motion) that reduces the ablation time and the contact of the burr with the calcified plaque. The objective is to avoid engaging the calcified stenosis all at once, but rather negotiating it bit by bit. These modifications to the RA protocol were associated with lower platelet activation and less thrombotic complications.37

With the current RA technique, transient atrioventricular blocks are seldom observed and are easily managed with boluses of atropine. Elective positioning of a temporary pacemaker could be considered, especially when rotablating right coronary or dominant left circumflex arteries, and should be balanced against the risk of right ventricular perforation induced by the tip of the pacemaker.

Several registries demonstrate higher PCI success rate and favourable clinical outcomes when DES implantation followed RA in calcified stenoses. In the EURO4C registry, including >1000 European patients with high clinical and angiographic risk profiles, a very low in-hospital complication rate was reported.38 All-cause death and major adverse cardiovascular events (MACE) at 1 year were 12% and 17%, respectively. Of interest, in an Australian patient population with high clinical and angiographic risk profile, a similar MACE rate was reported up to 1 year of follow-up when DES implantation followed RA compared with lower-risk patients treated with conventional DES implantation,39 suggesting that RA could function as risk equaliser. In randomised clinical trials including patients with moderate-to-severe calcified coronary stenoses, DES implantation after RA was associated with higher procedural success as compared with DES implantation after both conventional balloon or modified balloon (eg, with cutting/scoring) PCI.21 22 Importantly, up to 12% of the patients initially allotted to conventional or modified balloon PCI crossed over to RA strategy, suggesting that in a sizeable proportion of lesions and patients, PCI cannot be performed unless with RA. Subtraction of this patient from the initially allotted strategy has led to a risk dilution that can explain the lack of significant difference observed between the two comparative strategies in terms of late lumen loss and clinical endpoints.

These data are consolidated and reflected into the guideline recommendations granting a class IIa (level of evidence C) to RA for the preparation of heavily calcified or severely fibrotic lesions that cannot be crossed by a balloon or adequately dilated before planned drug-eluting stenting.40

OA (Diamondback 360°; Cardiovascular System, St. Paul, Minnesota, USA) is characterised by a 1.25 mm diamond coated crown mounted eccentrically proximal to the tip of the OA shaft. More details are provided in the online supplemental material.

Clinical data supporting OA are derived from the safety and feasibility of ORBITal atherectomy for the treatment of calcified coronary lesions (ORBIT) I and II trials, and a registry supporting the safety and efficacy of OA in severely calcified coronary lesions.41–44 The ORBIT II trial showed a target lesion revascularization rate of 7.8% at 3 years in patients with severe calcifications.45 The Evaluation of Treatment Strategies for Severe CaLcifIc Coronary Arteries: Orbital Atherectomy vs Conventional Angioplasty Prior to Implantation of Drug Eluting Stents or ECLIPSE trial is enrolling nearly 2000 patients randomised either to plaque modification with OA or conventional balloon angioplasty in patients with severe calcified coronary stenosis (NCT03108456).

OA was approved in the USA in 2013 for lesion preparation of severely calcified coronary lesions prior to stent implantation. It is currently available in the USA and Japan and recently was approved also in European Union.

There are no head-to-head comparisons between RA and OA. Given the burr design (mounted at the tip in case of RA versus proximal and sided to the tip in case of OA), RA might be preferred in tightest/uncrossable lesions.

ELCA is able through ultraviolet laser light to disrupt intravascular material. ELCA is not as effective as the other two atherectomy devices to modify heavily calcified lesions.46 The ablative effects of ELCA in calcified lesions are minimal on calcium and more pronounced on fibrous tissue. Yet, ELCA might be useful in tight stenoses, uncrossable with balloons or even microcatheters, especially in those cases where specialty wires of the RA or OA cannot be positioned distal to the calcified lesion. In fact, ELCA can be delivered on a standard 0.014-inch guidewire.

Intravascular lithotripsy

IVL is performed by the means of a novel balloon catheter, the Shockwave Medical Coronary Rx Lithoplasty System (Shockwave Medical, Fremont, California, USA) that emits pulsatile mechanical energy to disrupt calcified lesions, with a technology similar to lithotripsy for kidney stones. More details are provided in the online supplemental material.

A clinical case with IVL is shown in figure 5. IVL was tested in multicentre observational studies that demonstrated safety and efficacy of IVL in the treatment of heavily calcified coronary stenoses.47–49 Ever since, several single-centre observational registries and reports have supported the implementation of IVL in different off-label indications.50 Given the bulky profile of the current generation of Shockwave balloon, IVL seems to be best suitable for proximal lesion in rather large vessels. Sometimes, it might be needed to predilate tight stenosis with small 2.0–2.5 mm balloons in order to facilitate crossing of Shockwave balloon. Increasing preference to IVL is given by the operators to treat underexpanded stents to avoid performing the traditional stent ablation, that is, RA within the stent that exposes to higher risk of burr lodging.50

Figure 5

Clinical case of IVL. Heavily calcified stenosis of the mid-LAD at baseline (A) and after percutaneous coronary intervention with IVL (Shockwave balloon 3 mm) and drug-eluting stent implantation (B). At baseline, OCT shows heavy calcification involving three vascular quadrants (C). (D) OCT result after 80 pulses. (E) Final stenting OCT result. Yellow arrows indicate the lesion site before (A) and after treatment (B). IVL, intravascular lithotripsy; OCT, optical coherence tomography.


Complications possibly observed with devices and techniques for the treatment of heavily calcified coronary stenoses are as follows: (1) slow-flow/no reflow, (2) dissection, (3) perforation, (4) guidewire or burr fracture, (5) burr entrapment; (6) IVL-induced ventricular capture. The first three complications are to be expected within a more challenging lesion setting in patients often older and with several comorbidities. Yet, selection of the appropriate device and application of proper technique might prevent these complications, which, when occurring, should be managed not differently from cases of non-calcified coronary stenoses.

Wire and burr fractures or entrapment are rare complications that might occur with RA or OA especially in tortuous lesions or very angulated coronary anatomies.33 51–53 Several management strategies have been proposed, ranging from forcefully pulling the device (sometimes with the aid of an extension catheter) through advancing additional wires or balloons next to the device, up to the surgical removal of entrapped or embolised fragments.33

IVL-induced ventricular capture is frequently observed (up to 40% of the cases) and associated with a transient and generally well-tolerated decrease in systolic blood pressure.49 Rarely, IVL-induced ventricular capture has been associated with tachyarrhythmias.54 55


Invasive percutaneous treatment of calcified coronary stenoses begins with a careful preparation of the procedure. The operator needs to anticipate possible procedural challenges by assessing the clinical and diagnostic information available. In addition, there is no one-size-fits-all solution, but a successful treatment derives from a combination of tools and techniques. One might rely on algorithms proposed by experienced operators to help in decision-making while building its own experience. Implementation of available imaging techniques both non-invasive and invasive, along with judicious adoption of available tools and techniques, is key to a successful treatment of this challenging lesion setting. The bottom line is to gain sufficient confidence to manage the hardest coronary lesion at the lowest risk of complication for the patient.

Key messages

  • Heavily calcified coronary stenoses represent one of the last unmet clinical needs in percutaneous coronary interventions.

  • Plaque modification of the calcified lesion (ie, adequate fracturing of the calcium within the vessel media to increase lumen compliance to balloon dilatation) is paramount before drug-eluting stent implantation.

  • Several tools and techniques available nicely complement and are used in association to facilitate optimal treatment of this challenging lesion setting.

CME credits for Education in Heart

Education in Heart articles are accredited for CME by various providers. To answer the accompanying multiple choice questions (MCQs) and obtain your credits, click on the ‘Take the Test’ link on the online version of the article. The MCQs are hosted on BMJ Learning. All users must complete a one-time registration on BMJ Learning and subsequently log in on every visit using their username and password to access modules and their CME record. Accreditation is valid for only 2 years from the date of publication. Printable CME certificates are available to users who achieve the minimum pass mark.

Supplemental material

Ethics statements


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. *19.
  20. 20.
  21. 21.
  22. *22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. *33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. *38.
  39. 39.
  40. 40.
  41. 41.
  42. *42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. *50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Contributors As the sole author, EB confirms that he drafted, reviewed and submitted the manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests EB declares speaker’s fees from Boston Scientic and Abbott Vascular.

  • Provenance and peer review Commissioned; externally peer reviewed.

  • Author note References which include a * are considered to be key references.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.