Article Text
PDF
Abstract
Background Catheter retention and difficulty in retrieval have been observed during embolization of brain arteriovenous malformations (bAVMs) with the Onyx liquid embolic system (Onyx). The Apollo Onyx delivery microcatheter (Apollo) is a single lumen catheter designed for controlled delivery of Onyx into the neurovasculature, with a detachable distal tip to aid catheter retrieval. This study evaluates the safety of the Apollo for delivery of Onyx during embolization of bAVMs.
Methods This was a prospective, non-randomized, single-arm, multicenter, post-market study of patients with a bAVM who underwent Onyx embolization with the Apollo between May 2015 and February 2018. The primary endpoint was any catheter-related adverse event (AE) at 30 days, such as unintentional tip detachment or malfunction with clinical sequelae, or retained catheter. Procedure-related AEs (untoward medical occurrence, disease, injury, or clinical signs) and serious AEs (life threatening illness or injury, permanent physiological impairment, hospitalization, or requiring intervention) were also recorded.
Results A total of 112 patients were enrolled (mean age 44.1±17.6 years, 56.3% men), and 201 Apollo devices were used in 142 embolization procedures. The mean Spetzler–Martin grade was 2.38. The primary endpoint was not observed (0/112, 0%). The catheter tip detached during 83 (58.5%) procedures, of which 2 (2.4%) were unintentional and did not result in clinical sequelae. At 30 days, procedure related AEs occurred in 26 (23.2%) patients, and procedure-related serious AEs in 12 (10.7%). At 12 months, there were 3 (2.7%) mortalities, including 2 (1.8%) neurological deaths, none of which were device-related.
Conclusion This study demonstrates the safety of Apollo for Onyx embolization of bAVMs.
Clinical trial registration CNCT02378883.
- vascular malformation
- arteriovenous malformation
- catheter
- device
- liquid embolic material
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information. All data relevant to the study are included in the article or uploaded as supplementary information.
Statistics from Altmetric.com
Introduction
Brain arteriovenous malformations (bAVMs) are potentially life threatening and associated with significant morbidity and mortality due to the risk of rupture.1 The widely used Onyx liquid embolization system (Onyx) allows slow and controlled preoperative embolization of bAVMs.2–5 However, several disadvantages of Onyx represent challenges for neurointerventionalists: (1) the requirement of dimethyl sulfoxide-compatible rigid microcatheters and the frequent need to use guidewires may increase the risk of bleeding,6 7 (2) the fear of catheter retention, which has been reported in up to 10% of individuals and may result in thromboembolic complications,3 8–12 and (3) the difficulty in microcatheter withdrawal that may result in hemorrhage from vessel rupture.12 As these factors often preclude aggressive embolization, alternative approaches using devices optimized to prevent catheter retention may reduce the risk of significant morbidity and mortality from Onyx embolization.
The Apollo Onyx delivery microcatheter (Apollo) is a single lumen, end hole microcatheter with a detachable distal tip, which is designed to access the neurovasculature for controlled delivery of Onyx (figure 1). The detachable tip allows for catheter retrieval in the event that the distal 1.5–3 cm of the microcatheter tip becomes implanted within the bAVM or feeding vessel during embolization secondary to reflux of Onyx. This study represents the first clinical trial on the Apollo, with the primary purpose of evaluating its safety during bAVM embolization procedures.
Apollo Onyx delivery microcatheter. The catheter system consists of a 165 cm catheter with proximal and distal markers. The detachable tip containing the distal marker is shown in both attached and detached configurations. The detachable tip feature permits catheter retrieval in the event that the distal tip becomes implanted within the brain arteriovenous malformation or feeding vessel during the embolization procedure secondary to reflux of Onyx.
Methods
Study design and patient selection
This was a prospective, non-randomized, single-arm, post-market study performed at 22 centers across the USA between May 2015 and February 2018. The institutional review board at each center approved the study. After obtaining written informed consent, patients who had a confirmed diagnosis of bAVM were enrolled in the study if they were clinically and neurologically stable for at least 48 hours prior to embolization and had a life expectancy of at least 1 year. Patients were excluded if they (1) were participating in another investigational drug or device study evaluating treatments for arteriovenous malformations (AVMs) or other cerebrovascular disease at the time of the study, (2) had a known coagulation disorder, or (3) had a contraindication for vasodilators.
Baseline data recorded included demographic characteristics, medical history, AVM characteristics, physical or neurological symptoms potentially related to the AVM, and concomitant medication. A review of images at baseline confirmed the diagnosis of AVM, no earlier than 1 month prior to the embolization procedure. Pre-embolization neurological evaluation was assessed using the modified Rankin Scale (mRS), National Institutes of Health Stroke Scale, and Barthel Index.
Procedure and follow-up
Embolization(s) with the Onyx 18 or Onyx 34 liquid embolic system using the Apollo device was performed in a neuroangiography suite under intravenous sedation or general anesthesia. Patients were heparinized after placement of an arterial sheath. The endovascular procedures were performed according to local standards. Some patients subsequently underwent complete or partial surgical resection of their AVM, and information about the surgical procedure, imaging, medications, neurological function, concomitant medication, and adverse events (AEs) were recorded. Patients were monitored in hospital at least 24 hours post-embolization or to discharge, and then followed-up at 30 days and 12 months. Patients not experiencing an AE and without any catheter retained were considered complete at 30 days post-index embolization and did not have a 12 month follow-up visit (30 day cohort). Patients with retained catheter and/or tip and/or those with multiple embolization procedures underwent a 12 month follow-up post-index embolization (12 month cohort). Any patients with AEs at 30 days or 12 months post-embolization were followed until resolution of the AEs.
AEs were defined as any untoward medical occurrence, unintended disease or injury, or adverse clinical signs (including abnormal laboratory findings) and were classified as serious or non-serious. Serious adverse events (SAEs) were AEs that led to death or serious health deterioration, that resulted in (1) a life threatening illness or injury, (2) permanent physiological impairment, (3) inpatient or prolonged hospitalization, or (4) required a medical or surgical intervention to prevent permanent life threatening illness/injury or physiological impairment.
Study endpoints
The primary endpoint was any incidence of catheter-related AEs due to (1) premature (unintentional) catheter tip detachment with clinical sequelae, (2) catheter rupture/break/fracture with clinical sequelae, or (3) retained catheter body in the vasculature.
Acute (30 day) secondary endpoints included (1) rate of premature (unintentional) catheter tip detachment, (2) rate of intentional catheter tip detachment, (3) rate of migration of the retained catheter tip post-embolization, (4) incidence of procedure-related AEs, and (5) rate of leakage from the catheter detachment point without clinical sequelae.
Long term (12 month) secondary endpoints included (1) incidence of AEs and SAEs and (2) rate of migration of the retained catheter tip post-embolization. A steering committee of five neurointerventionalists oversaw the conduct and scientific aspects of the study. A clinical events committee (CEC), comprised of three physicians knowledgeable about the treatment and embolization of bAVMs, reviewed and adjudicated all AEs and SAEs. An imaging core laboratory assessed device migration.
Subgroup analyses assessed the incidence of procedure-related SAEs by sex, AVM location, and catheter tip destination (ie, not left in vasculature, fully resected, not resected, partially resected). Data were divided across time points (day 0 to study exit, days 0–30, and day 31 to study exit) for all patients, for the 30 day cohort, and for the 12 month cohort.
Statistical analysis
Data are presented using descriptive statistics. Discrete variables are presented using frequency distributions and cross tabulations. Continuous variables are presented using means, SD, median, and minimum and maximum values. Categorical variables were analyzed using frequency, incidence, and event rate. All statistical analyses were performed using Statistical Analysis System (SAS) (Windows V.9.2 or higher, SAS Institute Inc, Cary, North Carolina, USA).
Results
Patient enrollment
A total of 112 patients met the inclusion criteria and were enrolled in the study. The 30 day cohort comprised of 66 patients who either had the catheter and/or tip retained in the vasculature during the index procedure but was fully resected before the 30 day follow-up (n=23) or did not have any catheter/tip retained (n=43). The 12 month cohort comprised 46 patients who either had the catheter and/or tip retained in the vasculature during the procedure and did not have any resection performed (n=31) or were partially resected before the 30 day follow-up (n=15). Online supplemental figure 1 shows the patient selection flowchart.
Supplemental material
Patient demographics and bAVM characteristics
Baseline characteristics of the 112 enrolled patients are shown in the online supplemental table 1. Mean patient age was 44.1±17.6 years. The majority of patients were men (63/112, 56.3%). There were 97 (87.4%) patients with an mRS score of 0–2 at baseline, and 14 (12.6%) with an mRS score of 3–5. Baseline National Institutes of Health Stroke Scale score was 0 in 74 (66.7%) patients, 1–4 in 28 (25.2%), 5–15 in 7 (6.3%), and 16–20 in 2 (1.8%). Mean Barthel Index score was 93.3±18.8. It is noteworthy that a relatively high percentage of patients presented with intracranial hemorrhage (ICH) (45.5%), headache/migraine (51.8%), seizures (21.4%), and deep venous drainage (35.7%).
The characteristics of the target AVMs and the presenting neurological symptoms are presented in online supplemental table 2. A total of 59 (52.7%) AVMs were located on the left side of the brain, most frequently in the parietal lobe (39/112, 34.8%). Nidus size was small (<3 cm) in 60 (53.6%) patients, medium (3–6 cm) in 47 (42.0%), and large (>6 cm) in 5 (4.5%). Spetzler–Martin grade was I in 24 (21.4%) patients, II in 37 (33.0%), III in 36 (32.1%), and IV in 15 (13.4%). The rate of Spetzler–Martin grades III and IV was 45.5%, and mean Spetzler–Martin grade was 2.38.
Embolization procedure characteristics
Embolization procedure characteristics are presented in table 1. Overall, 142 study procedures were performed in 112 patients, with 1.3±0.6 procedures per patient. The 1.5 cm tip was used in 69.7% (99/142) of procedures, and the 3.0 cm tip length was used in 40.1% (57/142) of procedures. In some procedures, both models were used. For the majority of procedures (129/142, (90.8%)), the catheter was inserted by right femoral access. Onyx was injected in a total of 130 procedures: Onyx 18 was used in 99 (76.2%) procedures, Onyx 34 was used in 51 (39.2%) procedures, and both Onyx 18 and Onyx 34 were used in 20 (15.4%) procedures. Mean volume of Onyx injected was 2.4±2.4 mL. Mean cumulative fluoroscopy time was 158±227 min (range 10.0–1786.0). Mean total reflux was 3.1±2.8 cm among 128 procedures, and a gap was present between the reflux and proximal marker in 66.2% (94/142) of procedures. The catheter tip detached in 58.5% (83/142) of procedures, of which 2 (2.4%) were unintentional but did not result in clinical sequelae. In one case, the tip detached due to tortuosity and the small caliber of the vessel. In the second case, the tip detached unintentionally during catheter removal; this occurred during the fourth embolization, 157 days after the procedure, which was outside of the 30 day acute secondary endpoint analysis window.
Embolization procedure characteristics
A summary of procedure outcomes is presented in table 2. A total of 5 (4.5%) patients underwent radiosurgery and 62 (55.4%) underwent surgical resection following embolization. The mean number of pedicles embolized was 1.9±2.0. A total of 13 (11.7%) patients achieved 80.0–100.0% AVM obliteration, with the majority (68, 61.3%) achieving 0.0–20.0% obliteration. The AVM was considered cured in 64 (57.1%) patients, as determined by the study physician. The majority of patients (69, 61.6%) did not have additional treatment planned.
Summary of procedure outcomes
Primary endpoints
Unintentional catheter tip detachment with clinical sequelae, catheter rupture/break/fracture with clinical sequelae, or retained catheter body in the vasculature was not observed in any patients (0/112, 0.0%, table 3). The primary endpoint was also analyzed based on 126 Apollo devices deployed within 30 days of the acute endpoint window. The primary endpoint was not observed in any of the microcatheters used in the acute period of this study (0/126, 0.0%).
Primary endpoint analysis
Secondary endpoints
Acute secondary endpoints by patient are presented in online supplemental table 3. Up to the 30 day follow-up, unintentional catheter tip detachment was observed in 1 (0.9%) patient. Intentional catheter tip detachment was observed in 68 (60.7%) patients. Procedure-related AEs were observed in 31 (27.7%) patients, including SAEs in 12 (10.7%), and non-serious AEs in 26 (23.2%). Leakage from the catheter detachment point without clinical sequelae and migration of the retained catheter tip after embolization was not observed in any patients at the 30 day follow-up visits.
Long term secondary endpoints are also presented in online supplemental table 3. Up to the 12 month follow-up, CEC-adjudicated AEs were reported in 68 (60.7%) patients. SAEs were reported in 39 (34.8%) patients, and non-serious AEs in 50 (44.6%). Migration of the retained catheter tip post-embolization was not observed up to 12 months. The relationship of CEC-adjudicated AEs is shown in online supplemental table 4 and a summary of SAEs by site is shown in online supplemental table 5. At 30 days, 43 procedure-related AEs occurred in 26 (23.2%) patients, and 16 procedure-related SAEs occurred in 12 (10.7%) patients. A full description of the procedure-related AEs and SAEs are presented in online supplemental tables 6 and 7. During the study, there were 3 (2.7%) mortalities, including 2 (1.8%) neurological deaths, none of which were device-related.
The first mortality was adjudicated by the CEC as a procedure-related SAE due to air embolism, meeting the definition of neurological death. The second occurred due to an unrelated malignant lung neoplasm, adjudicated as an SAE due to concomitant disease. The third was caused by a hemorrhagic stroke on day 339 following the embolization procedure and was adjudicated as a disease-related SAE meeting the definition of neurological death.
Subgroup analyses
Procedure-related SAEs by sex, AVM location, and catheter tip placement across different time points are presented in online supplemental table 8. The number of procedure-related SAEs up to study exit for all patients was similar between women (5/49, 10.2%) and men (9/63, 14.3%; p=0.52). Parietal AVMs had the highest rate of procedure-related SAEs for day 0 to study exit for all patients at 17.9% (7/39). When comparing between subgroups for catheter tip placement, the highest rates of procedure-related SAEs were found in patients with no catheter/tip resection (22.6%, 7/31) and partial catheter/tip resection (20.0%, 3/15). Those with no catheter tip left in the vasculature had procedure-related SAE rates of 4.7% (2/43) and with full resection of the catheter/tip at 8.7% (2/23) up to study exit (p=0.31 for resection vs. no and partial resection up to study exit). There was no difference in procedure-related SAEs for patients treated with the 1.5 cm catheter tip length (11.9%, 10/84) versus the 3.0 cm tip length (13.2%, 5/38; p=1.00 up to study exit).
Discussion
Prospectively collected post-market data from patients with bAVMs treated with Apollo showed that no primary endpoint events leading to clinical sequelae occurred during the study. Up to the 30 day follow-up, the rate of unintentional catheter tip detachment was low and not associated with clinical sequelae. Leakage from the catheter detachment point without clinical sequelae or migration of the retained catheter tip post-embolization was not observed in any patients. The cure rate of 57.1% falls within the range reported in other studies that deployed Onyx,2 4 13 14 despite the relatively high rates of unfavorable baseline conditions, such as ICH and Spetzler–Martin grades III and IV bAVMs in this cohort, with many patients requiring more than one embolization procedure. In addition, we and others showed that Onyx embolization using Apollo had satisfactory morbidity and mortality rates,2 4 demonstrating the safety of Apollo for delivery of Onyx for bAVM embolization.
Dimethyl sulfoxide-compatible undetachable microcatheters, such as Marathon (Medtronic, Minneapolis, USA) and Echelon (Medtronic) have been used to deploy Onyx for the treatment of bAVMs. However, cases of catheter entrapment and difficulty in retrieval of the undetachable catheters have been observed in a number of AVM embolization studies using Onyx.3 8–12 15–17 Catheter entrapment/abandonment occurred in 0.0–9.7% of patients treated with Onyx, with an average entrapment rate of 5.8%.8–12 16 Withdrawal of the entrapped microcatheter can cause hemorrhage from vessel rupture, or lead to severe morbidity or death.12 Fear of having a conventional catheter retained conduces to timid non-aggressive Onyx embolization. Therefore, many alternative approaches have been proposed to prevent catheter entrapment or aid in the safe removal of entrapped catheters.18–23
The concept of detachable tip microcatheters designed to minimize the risk of catheter entrapment during the embolization of bAVMs with Onyx was first employed by using the Sonic microcatheter (Balt Extrusion, Montmorency, France) which was approved in Europe, but not the USA. Few studies have evaluated the feasibility and safety of the Sonic microcatheter when using Onyx.4 24–26 Similar to the concept of Sonic, the Apollo was designed with a detachable tip that allows for catheter retrieval and was approved by the Food and Drug Administration in 2014. Many studies have evaluated the safety and efficacy of Apollo.2 14 27 28
In our study, 40.2% of patients were treated by Onyx embolization only, while in the remaining patients, Onyx embolization was performed as an adjunctive treatment to either radiosurgery or surgical resection, and the majority (80.4%) underwent a single embolization. There was no further treatment planned for 61.6% of patients and the bAVM was cured in more than half of the patients, suggesting the aggressive nature of treatment when using the Apollo. This type of aggressive treatment may explain the high rate of detachment (58.5%) during embolization procedures, similar to previous investigations that reported high rates of microcatheter tip detachment ranging from 20% to 50%.14 28 29 Indeed, Flores and colleagues suggested the notion of more aggressive treatment when using a detachable microcatheter tip.14 In the present study, unintentional catheter tip detachment with clinical sequelae, catheter rupture/break/fracture with clinical sequelae, or retained catheter body in the vasculature were not observed in any patient, and rates were lower than previously published studies.8–12 16 Additionally, there were no cases of migration of the retained catheter tip after embolization, in accordance with a previous study using Apollo for embolization of pediatric bAVMs.27
Morbidity and mortality rates of this trial agree with previously published data. In a large retrospective study of 131 patients with bAVMs, rates of temporary morbidity, permanent morbidity, and mortality post-embolization were 10%, 8%, and 1%, respectively.30 Several series have reported treatment-related morbidity rates ranging from 3% to 11%.2 8 9 The BRAVO trial showed that embolization with Onyx resulted in 4.3% mortality and 5.1% morbidity.31 In a retrospective study examining the overall neurologic complication rate in patients undergoing AVM embolization, transient neurologic deficits occurred in 11.5%, with permanent non-disabling complications and permanent disabling complications or deaths in 2.6% and 1.6% of patients, respectively.10 Baharvahdat et al specifically investigated hemorrhagic complications in 408 patients undergoing embolization for bAVMs and observed an 11% rate of procedure-related hemorrhage.32 In the current study, all-cause mortality rates and procedural mortality rates were 2.7% and 1.8%, respectively. The majority of procedure-related AEs observed in this study were ischemic strokes, cerebral hemorrhage, and ICH, and are comparable with rates in recent reports.
The subgroup analyses assessing procedure-related SAEs revealed indicators of particularly challenging AVMs. Subgroups of sex, laterality, and AVM location showed modest differences in SAE rates, with parietal AVMs having the highest number of SAEs. Parietal AVMs are associated with symptoms and cognitive deficits, including localized headache, homonymous hemianopsia, hemiparesis, ataxia, hemisensory changes, agnosia, weakness, rotational vertigo, seizures, and coma.33 34 Parietal AVMs are particularly challenging to treat surgically due to limited access to the interhemispheric fissure, and the presence of the bridging veins, arterial feeders, and deep venous drainage.34 The subgroups with no or partial catheter tip resection (which may indicate extra-nidal and/or off-target embolization, or embolization performed with radiation treatment) had higher SAE rates compared with those with no catheter tip left in the vasculature or full resection.
Risk factors associated with new neurological deficits following bAVM embolization include increasing patient age, larger number of embolizations, higher injected volume of embolic agent, and deposition of embolic material in the venous outflow before complete nidal occlusion.32 35 The AVM embocure score (AVMES) has been used to characterize AVMs based on the number of arterial pedicles and draining veins, size of the AVM nidus, and vascular eloquence.36 Higher AVMES has been associated with reduced complete occlusion rates and higher risk of complications; thus, embolization with the intent to fully cure in the setting of high AVMES may result in an increased risk of neurological complications compared with partial embolization.36 37 ICH is the most common cause of neurological disability following endovascular treatment of AVMs.38 In addition, small bAVMs (<3 cm)39 and deep venous drainage40 have been suggested as risk factors for AVM bleeding. Many patients in this study had a history of ICH, headache/migraine, seizures, and deep venous drainage, which may have contributed to the high incidence of procedure-related AEs. Despite the relatively high percentage of (1) patients presenting with ICH (45.5%), (2) grade III and grade IV Spetzler–Martin (45.5%), and (3) patients requiring more than one embolization procedure (19.6%), AVMs were deemed to be cured in a relatively high percentage (57.1%) of patients, and further treatment was not planned in the majority of patients at the discretion of the treating physicians. In accordance with our results, studies using detachable microcatheters, including Sonic and Apollo, to deploy Onyx in patients presenting with high rates of ICH and high grades of Spetzler–Martin demonstrated cure rates ranging from 42.9% to 55.0%.2 4 Importantly, a literature review of 18 large case series of Onyx embolization without the use of a detachable tip microcatheter reported an endovascular cure rate ranging from 0.0% to 51.1% (mean 20.3%), major morbidity ranging from 1.4% to 23.9% (mean 8.8%), and mortality rate ranging from 0.0% to 4.3% (mean 2.2%),14 which are less optimal than the outcomes reported using a detachable tip in our study and others.2 4
Collectively, our results confirm the safety of Onyx delivery by the Apollo, which is a clear example of an industry response to a clinical need, by allowing longer injections, deeper penetration of the bAVM nidus, and higher volumes of Onyx that can be safely injected. The major limitation of this study was its single-arm nature, as all patients were treated with the Apollo, with no matched comparison group. Because the study was not randomized, selection bias cannot be completely excluded. However, there were independent committees in place to design and monitor the study and adjudicate outcome measurements.
Conclusion
This study provides evidence that the Apollo Onyx delivery microcatheter is safe for delivery of Onyx during bAVM embolization procedures, with little risk of unintentional catheter detachment or migration causing clinical sequelae.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information. All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Ethics approval
The institutional review board at each center approved the study.
Acknowledgments
The authors acknowledge Medtronic for editorial support.
References
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.
Footnotes
Twitter @justinSMD, @fegomd, @henrywoomd
Contributors All authors made substantial contributions to the conception and design, analysis, and interpretation of data; drafted or critically revised the article; and gave final approval of the version to be published.
Funding This study was sponsored by Medtronic, Inc.
Competing interests JF reports grants and non-financial support from Stryker, grants and other funding from Penumbra, grants from Microvention, and equity in Cerebrotech and The Stroke Project. CAG serves on the speaker’s bureau and as a consultant and proctor for Medtronic and Stryker, and serves on the speaker’s bureau for Genentech. MM serves as a consultant for Toshiba/Canon Medical. PK serves as a consultant for Stryker Neurovascular, Medtronic, and Cerenovus. TY serves as a consultant and proctor for Medtronic. OD serves as a proctor for Microvention/Terumo. DF serves as a consultant for Balt, Marblehead Medical, Medtronic, Stryker, Microvention, Penumbra, and Cerenovus, receives research support from Cerenovus, Medtronic, Stryker, Siemens, Microvention, and Penumbra, receives royalties from Codman, and is a stockholder in Marblehead, Neurogami, and Vascular simulations outside the submitted work. RVC has received research funding from Microvention, Cerenovus, and Medtronic. AKW has received research grants from Philips Medical, serves as a consultant for Stryker and Phenox, is a stockholder in InNeuroCo, EpiEP, Neural Analytics, Rist, Analytics 4 Life, and ThrombX, and serves on the speaker’s bureau for SCENT trial presentations. ASP serves as a consultant for and has received research grants from Medtronic Neurovascular and Stryker Neurovascular. RAH serves as a consultant for Medtronic, Stryker, Codman, and Microvention, and is a stockholder in InNeuroCo. LFG holds stock in Culvert Therapeutics and NeuroCool, serves as a proctor for Medtronic Neurovascular, and participates in training other physicians in the use pipeline embolization device. MAA-S serves as a proctor for Medtronic Neurovascular, and participates in training other physicians in the use of the embolic agent Onyx and the pipeline embolization device.
Provenance and peer review Not commissioned; externally peer reviewed.
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.