Introduction Reducing exposure to cigarette smoke is an imperative for public health and for patients with diabetes. Increasingly, combustion-free nicotine delivery systems (C-F NDS) such as e-cigarettes and heated tobacco products are substituting conventional cigarettes and accelerating the downward trends in smoking prevalence. However, there is limited information about the long-term health impact in patients with diabetes who use C-F NDS. This randomised trial of type 2 diabetic cigarette smokers will test the hypothesis that following a switch from conventional cigarettes to C-F NDS a measurable improvement in metabolic syndrome (MetS) factors will be shown over the course of 2 years.
Methods and analysis The study is multicentre and thus will take place in five locations in four countries in an ambulatory setting. A total of 576 patients with diabetes will be randomised (1:2 ratio) to either a control arm (Study Arm A), in which they will be offered referral to smoking cessation programmes or to an intervention arm (Study Arm B) assigned to C-F NDS use. Participants will be at least 23 years old and of any gender. Patient recruitment will start in February 2021 and is expected to be completed by December 2021. Primary outcome measures include fasting plasma glucose, blood pressure, triglycerides, high-density lipoprotein and waist circumference, while secondary feature absolute change in the sum of the individual factors of MetS and change in each individual factor of MetS measured at each study time point.
Ethics and dissemination The approval of research ethics committee (REC) regarding the trial protocol, informed consent forms and other relevant documents is required to commence the study. Substantial amendments to the study protocol cannot be implemented until the REC grants a favourable opinion. The results of the study are intended to be published as articles in high quality peer-reviewed journals and disseminated through conference papers.
Trial registration number NCT04231838. Pre-results stage.
- diabetes & endocrinology
- lipid disorders
- general diabetes
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
DIASMOKE will be the first study to determine an overall health impact of combustion-free nicotine delivery systems (C-F NDS) in diabetes and its cardiovascular risk.
Adherence to C-F NDS will be strengthened by providing a wide variety of different products to meet patients’ preference.
Compliance to the study protocol will be monitored daily via a mobile application.
Due to the relatively long duration of the study, adequate participants’ retention may be challenging.
Study results cannot be generalised to people with type 1 diabetes mellitus (T1DM) or with unstable T2DM.
Diabetes mellitus (DM) can cause irreversible damage to the blood vessels leading to microvascular (retinopathy, nephropathy and diabetic neuropathy) or macrovascular (coronary artery disease, stroke, peripheral arterial disease) complications,1 the latter cardiovascular complications being most common, and a frequent cause of death. Besides diabetes and hyperglycaemia, obesity, hypertension and dyslipidaemia are well established cardiovascular risk factors, all of which come under the umbrella definition of metabolic syndrome. Other cardiovascular risk factors may also coexist in these patients, the most important being smoking.
Cigarette smoking is a strong cardiovascular risk factor not included in the definition of metabolic syndrome (MetS) but substantially increases the risk of microvascular and macrovascular complications in patients with type 2 DM (T2DM),2–5 whereas quitting smoking substantially reduces this risk.4–7 Given that exposure to cigarette smoke is associated with vascular damage, endothelial dysfunction and activation of coagulation and fibrinolysis,8–10 it is not surprising that smoking enhances the combined harmful effects of elevated blood glucose and other risk factors and accelerates vascular damage in patients with diabetes.
If reducing exposure to cigarette smoke is an imperative for public health, it is even more so for patients with T2DM.11 However, prevalence of smoking among people with DM appears to be similar to that of the general population.12 In the USA, the prevalence of tobacco consumption has decreased substantially, but this beneficial trend has not been observed in patients with DM.13
There is a clear urgent need to target patients with T2DM to successful smoking cessation therapies, such as nicotine-containing preparations.14 15 Unfortunately, there is no convincing demonstration of effective cessation interventions in patients with diabetes16 and, in general, most smokers are reluctant to seek formal treatment for stopping smoking with the vast majority making attempts to quit without assistance.17 18 Consequently, the need for novel and more efficient approaches is required.
Combustion-free technologies for nicotine delivery such as e-cigarettes (ECs) and heated tobacco products (HTPs) are substituting conventional cigarettes globally19 and are thought to be less harmful alternative to tobacco smoking.20–22 However, there are no long-term studies assessing cardiovascular risk or effect on cardiovascular risk factors in patients with diabetes who use these technologies.
The DIASMOKE collaborators seek to determine whether T2DM cigarette smokers who switch to combustion-free nicotine delivery systems (C-F NDS) experience measurable improvements in their cardiovascular risk parameters.
DIASMOKE (Assessing the impact of combustion-free nicotine delivery technologies in DIAbetic SMOKErs) is an international, multicentre, open-label, randomised controlled study analysing two parallel groups of participants, designed to determine whether T2DM cigarette smokers switching to C-F NDS experience measurable improvement in cardiovascular risk parameters as a consequence of avoiding exposure to cigarette smoke toxicants.
The inclusion and exclusion criteria are summarised in box 1. Participants will be recruited from a group of cigarette smokers with a clinical diagnosis of T2DM. Only regular cigarette smokers will be considered for inclusion (criteria mentioned in box 1). Smoking status will be verified by an exhaled carbon monoxide (CO) measurement (exhaled CO ≥7 ppm) at the screening visit. Each participant will be offered access to local free smoking cessation programmes, and only those who refuse participation in cessation programmes and are willing to switch to a C-F NDS will be randomised following informed consent. Participants included will be willing to refrain from eating/drinking prior to the screening visit and check-in at each study visit.
Inclusion and exclusion criteria
Written, informed consent signed before any study-specific procedure.
Men or women aged 23 years and older.
Regular smokers of at least 10 cigarettes/day (maximum of 30 cigarettes/day) for at least 5 consecutive years prior to the screening visit.
Type 2 diabetes mellitus (as defined by the American Diabetes Association).
Body mass index between 18.5 kg/m2 and 34.9 kg/m2, both inclusive.
Body weight exceeding at least 50 kg (men) or 40 kg (women).
Exhaled carbon monoxide level of at least 7 ppm (parts per million).
History of recent acute decompensation of their disease requiring treatment within 4 weeks prior to visit 1.
Known clinically-significant neurological, gastrointestinal, renal, hepatic, cardiovascular, psychiatric, respiratory, metabolic, endocrine, haematological or other major disorder that, in the opinion of the investigator or their appropriately qualified designee, would jeopardise the safety of the participant or impact on the validity of the study results.
Any other condition or therapy that would make the patient unsuitable for the studies and will not allow participation for the full planned study period (eg, active malignancy or other condition limiting life expectancy to <12 months).
A significant history of alcoholism or drug/chemical abuse within 24 months prior to screening.
Regular use of any nicotine or tobacco product other than their own cigarettes within 14 days of screening.
Pregnant or breast feeding or intention to become pregnant during the studies.
Previous (within 90 days prior to randomisation) or concomitant participation in another clinical study involving administration of an investigational drug.
Close affiliation with the investigational site; for example, a close relative of the investigator, dependent person (eg, employee or student of the investigational site).
The study design flow of DIASMOKE is illustrated in figure 1. The project will take place in five locations in four different countries (UK, Italy, Poland and Moldova) in an ambulatory setting.
Participants will attend a screening visit within 28 days prior to visit 1 (table 1A) and undergo demographic assessments including socio-demographic data, detailed medical history (including medication use), detailed smoking, vaping and HTPs use history and their intention to quit. Modification in their diet and/or anti-diabetic medication will be recorded regularly throughout the study. All patients will be offered smoking cessation programme as per local guidelines. Participants will be offered a further second opportunity to enrol in the free local smoking cessation programme prior to enrolment.
Following baseline assessments on day 1 (table 1B), participants will be randomised to either the control (A) arm or the intervention (B) arm. The randomisation sequence will be computer generated, with an allocation ratio of 1:2 (arm A:arm B) in order to compensate for the estimated 50% drop-out rate. The patient will be allocated to one of the study arms automatically after the staff will access the web-based application entering their participant identification number, a month and a year of birth and initials into the programme. Patients randomised into arm B will be allowed to choose the product of their preference from the given pool of most popular C-F NDS. The participants will be trained and counselled on the chosen device and given a full 1-week supply of tobacco sticks/EC cartridges/e-liquids refill bottles prior to check-out on day 1. After randomisation, a dedicated tracker application will be installed on patients’ smartphones. The application is designed to track patients behaviour (physical activity, adherence to sugar testing, cigarette smoking frequency, daily C-F NDS usage) to identify protocol violations that will generate flagging events and alerts, to collect adverse events and to send reminders (next scheduled appointment, study restrictions, instructions and so on) throughout the whole duration of the study.
Subsequently, participants will be invited to attend four further clinical visits conducted in an ambulatory setting (visits 2–5) to undergo a range of measurements and blood tests (table 1A,B). Following each visit participants will be supplied with an appropriate amount of consumables (tobacco sticks, EC cartridges, e-liquid refill bottles). Participants will fast overnight (from midnight) prior to each study visit at which clinical laboratory evaluations will be performed. Patients will be instructed to refrain from consuming alcohol for 24 hours prior to clinic visits and instructed not to consume more than 14 units of alcohol per week for the entire duration of the study.
For patients randomised into arm B, between those clinical visits, additional non-clinical visits aiming to replace the used consumables are planned. At non-clinical visits, study investigators will also have the opportunity to stimulate retention and check compliance. In order to perform an evaluation of the habitual pattern of use of the C-F NDS and to verify product adherence, patients randomised into arm B will return all empty, part-used and unused consumables at each visit.
At each visit, all participants will be advised and encouraged to completely quit smoking (cigarette or C-F NDS). They will explicitly be told about the risks associated with smoking and at every contact time point offered referral to local free smoking cessation programmes.
Premature withdrawal from the study may occur if a participant: (1) experiences a severe adverse event (SAE); (2) sustains any protocol deviations occurred during the conduct of the study, which cannot be corrected; (3) is uncooperative, including non-attendance; (4) decides to stop his/her participation at any moment of the study; (5) becomes pregnant.
DIASMOKE is an unblinded study due to its specification.
It is not possible to blind participants to the intervention they will be receiving as well as trial staff when providing the interventions and collecting data.
Source data and source documents will be managed according to the Good Clinical Practice guidelines.
The trial will formally end on the date of the last visit of the last patient in the last country undertaking the trial.
In order to provide an adequate data collection each individual patient will be allocated a case report form (CRF). CRF will be an electronic document. The CRF data will be used to perform statistical analysis for the trial. Anonymised data from each study visit will be entered directly onto the CRF as it will then become a source document. The CRFs will be web-based and all study sites will have access to their information. In order to promote data quality the study will use standardised instruments such as Diabetes Quality of Life questionnaire or Fagerström Test For Nicotine Dependence. Personal data will be protected as each participant will be allocated a unique study identification number (patient ID). Participants’ personal details will not be attached to the research results and the decoding list will only be available to a limited number of members of the research team. All information obtained during the study procedures will be treated as private and confidential.
Patient and public involvement
A focus group of smokers with diabetes was organised on 25 February 2020 and feedback from smokers was used in the trial design. Further, the study has been reviewed by Ashford and St Peter’s Hospitals NHS Foundation Trust’s Research and Development Committee, which includes a patient representative.
Objectives and endpoints
The primary objective of DIASMOKE is to assess the impact of sustained use of C-F NDS on the proportion of patients with metabolic syndrome, as defined by National Cholesterol Education Program (NCEP) MetS score23 below the diagnostic threshold (<3). The primary outcome of the study will be change in prevalence of an NCEP MetS score <3 between baseline and 2 years follow-up, with comparison being made between patients with T2DM randomised to each arm of the study.
Change in prevalence will also be assessed at 3 months, 6 months and 1 year, as secondary outcomes. All assessments at each time point will be undertaken in all participants in both arms. Considering the results of a number of lifestyle modification interventions, the absolute reduction in MetS prevalence following substantial smoking cessation is expected to be no less than 15%.24–28
The main prespecified secondary endpoint is an absolute change in the sum of the individual factors of the MetS (as defined by NCEP criteria) measured at each study time point (between and within study groups). Other secondary endpoints include change in each individual factor of the MetS (as defined by NCEP criteria) measured at each study time point (between and within study groups) and change of the variables given in table 1B measured at each study time point (between and within study groups).
Powering and sample size calculation
For this study, the following input assumptions were considered:
The absolute reduction in MetS prevalence following substantial smoking cessation is expected to be 15%, based on the results of a range of lifestyle modification interventions.24–28
The baseline prevalence of MetS in T2DM is expected to be 70%.29–32
Sample size was calculated on the basis of demonstration of superiority, using an assumption of normal distribution, as described by Pocock.33 Significance level was set at 5% (α=0.05), with a power of 80% (β=0.20). On this basis, the minimum number of patients with analysable data required is 160 per treatment arm (N).
Further assumptions at the planning stage included an estimated 50% proportion of patients randomised to C-F NDS who are expected to achieve sustained reduction in cigarette consumption of at least 80% for the duration of the study (%SusRed).34–38
The adjusted number of patients in the intervention arm (N2) was therefore increased to 320:
N2=N/%SusRed=320 (N2 indicating the final number of patients required after taking into consideration the 50% sustained reduction figure).
Additionally, the expected number of patients in both arms withdrawing from the trial over 2 years is estimated at 20%.39–41 The total number of patients recruited to each treatment arm was therefore increased by this amount:
Intervention arm: 320×1.2=384.
Control arm: 160×1.2=192.
Total patients both arms=576.
In order to reach the target sample size patients with diabetes will be informed about the potential benefits of switching to C-F NDS as well as the ability to report their health problems to their site investigator via a mobile application.
The primary endpoint for the statistical analysis is defined as the between-groups difference in calculated prevalence of MetS after at least 24 months of follow-up. The full analysis set (FAS) comprises all patients randomised to the intervention arm who achieve a sustained reduction in cigarette consumption of at least 80% across the full duration of follow-up combined with all patients randomised into the standard care control group.
The FAS will be the primary analysis set for all efficacy analysis. Two approaches to the primary analysis will be used: (a) Unadjusted analysis, based on a direct comparison of the change in prevalence. Z test will be used to assess the significance of difference between the two groups in the prevalence percentage changes from baseline to 24-month visit. (b) Adjusted analysis. Baseline demographics, clinical and concomitant therapeutic characteristics will be analysed to identify potential confounders for the primary outcome that are unbalanced between treatment groups. The primary outcome will then be re-analysed using a generalised linear model adjusting for all identified confounders.
Any difference between groups will be assessed for statistical significance at a two-sided alpha of 0.05.
An independent data monitoring and safety committee (DMC) will be established for this study before the first participant is randomised and will overview the safety of the study. The DMC will review safety data on a periodic basis, and make recommendations to continue, modify or stop the study. The DMC will evaluate the efficacy and safety results of the primary analysis after 6 months (or otherwise if determined by the committee) and make a recommendation regarding early termination based on observed results of the study on grounds of an unfavourable risk-benefit profile. In the event that the assumptions underlying the sample size calculation are seen to be incorrect at the time of the interim analysis, they will have the option to advise further recruitment to the study, without disclosing the interim results to the study investigators. The DMC will be independent from the sponsor and competing interests.
A trial monitoring plan will be developed and agreed by the trial steering committee and chief investigator based on the trial risk assessment which may include on-site monitoring. The contact research organisation (Metanoic Health Ltd) will arrange a monitor independent from investigators and the sponsor. The processes reviewed can relate to participant enrolment, consent, eligibility and allocation to trial groups; adherence to trial interventions and policies to protect participants, including reporting of harm and completeness, accuracy and timeliness of data collection. Monitoring will be done by exploring the trial data set or performing site visits.
Adverse events (AEs) and SAEs will be noted during the whole duration of the study. AEs and SAEs will be recorded at baseline and at each subsequent study visit in the adverse event page of the CRF. Signs or symptoms will be investigated at each visit by interviewing the participants. Patients will also be encouraged to report AEs/SAEs at any time during the study. The investigator must pursue and obtain information adequate both to determine the outcome of the AE and to assess whether it meets the criteria for classification as a SAE requiring immediate notification to the competent authority. Sufficient information should be obtained to assess causality. Follow-up of the AE/SAE after the date of study discontinuation is required if the AE/SAE or its sequelae persist.
Ethics and dissemination
The study will be conducted according to the principles of Good Clinical Practice and Declaration of Helsinki. All five local ethics committees reviewed and approved the study and—where appropriate—translated relevant documentation (informed consent form, patients information sheet and so on). A list of the ethics committees that reviewed and approved the study is attached as online supplemental file 1. If any amendments to this protocol are required the chief investigator will be responsible for the decision to amend the protocol and for deciding whether an amendment is substantial or non-substantial. Any substantial amendments will be submitted to the research ethics committee for approval before implementation. Any amendments will apply to all sites.
The informed consent or assent from potential trial participants will be obtained by site investigators through relevant forms (see online supplemental file 2).
In the UK all investigators and trial site staff will comply with the requirements of the Data Protection Act 2018, with regards to the collection, storage, processing and disclosure of personal information and will uphold the Act’s core principles. In other countries any equivalent local data protection regulations will be complied with.
The trial steering committee (TSC) will have access to the full trial data set. A formal access request from site investigator(s) will require TSC and sponsor approval. All committee members are independent and have no conflict of interest.
The intention of the TSC is to disseminate the results of the study through journal articles in high quality peer-reviewed journals and through conference papers. A summary of results will be available on the Ashford and St Peter’s Hospitals website where patients and members of the public will be able to access it. The sponsor institution has open data access policy and the anonymised data will be available on request to any researcher following approval from the established scientific committee.
The informed consent materials (consent form and patient information sheet) are attached as online supplemental files 2 and 3.
Patient recruitment will start in February 2021 and enrolment is expected to be completed by December 2021. Results will be reported between 2023 and 2024.
Little is known about the impact of C-F NDS on patients with T2DM who smoke. Products that do not require combustion to deliver nicotine, such as ECs and HTPs are substituting conventional cigarettes globally.19 They potentially offer substantial reduction in exposure to harmful and potentially harmful chemical constituents compared with conventional cigarettes.20–22 42–44 DIASMOKE will be the first study determining the overall health impact of using such technologies in patients with diabetes. Undoubtedly, it is desirable for patients to avoid consumption of any tobacco-related inhalation products, but in order for governments, health authorities (e.g. European Medicines Agency, Food and Drug Administration) and clinicians to provide guidance about cigarette substitution, robust evidence-based information is required.
We designed this international randomised controlled trial (RCT) to gather such evidence. In particular, we will be testing the hypothesis that avoiding exposure to cigarette smoke toxicants may translate to measurable improvement in cardiovascular risk factors when patients with T2DM who smoke switch to using C-F NDS compared with patients with T2DM who continue to smoke conventional cigarettes. Several parameters measured in this study are associated with the development of cardiovascular diseases (such as high blood pressure, elevated blood cholesterol and BMI >25) and some of these indicators have been shown to improve relatively soon after smoking cessation.34 45 Consequently, the profile of these changes after switching to C-F NDS could provide valuable insights into the overall potential of C-F NDS to reduce the risk of cardiovascular disease.
The decision for a switching study design in DIASMOKE has been guided by the notion that C-F NDS have been promoted as substitutes for tobacco cigarettes. In a switching study of smokers the reference product is their own brand tobacco cigarette. The length of the study was based on the consideration that changes in the primary endpoint could be reasonably observed as early as 6 months. It is however possible that a much longer follow-up period could be necessary to firmly establish findings consistency over time, hence study duration was extended to 24 months. The RCT study design will provide a robust answer to determine the health impact of C-F NDS use on patients with diabetes. Clearly, randomisation will equalise variation in smoking history and other variables between study arms, thus ensuring high quality data. Importantly, the entire study is designed keeping the welfare of all participants at its centre; at every contact smokers will be asked to stop all types of smoking and provided with free local referrals for smoking cessation programmes.
Compliance to the study protocol is critical as failure to fully or largely replace conventional cigarettes with C-F NDS would reduce or nullify the expected changes in study endpoints. Participants will be reminded on the importance of adhering to their randomised product allocation and of abstaining from or greatly reducing conventional cigarette consumption (by at least 80% from their baseline value of cigarette smoked in a day) at every contact. They will also be informed that biochemical verification of compliance as well as assessments of adherence will be conducted at each clinic visit. In addition, any non-compliance will be recorded in the study diary after counting all empty, part-used and unused consumables returned at each visit, and tracked by the application. Although not expected that compliance for this study will be materially different compared with other comparable studies, our power calculations are overestimated to take account of a non-compliance rate of 50%. Thus, the C-F NDS population will be oversampled by adopting a 1:2 randomisation ratio scheme (ie, for every patient randomised in the control population, two will be randomised in the C-F NDS population). Lastly, trial attendance and retention of the C-F NDS population will also be improved by asking participants to return to the clinic for their regular re-supply of tobacco sticks/EC cartridges/e-liquids refill bottles.
This study has several innovative features. To improve adherence to C-F NDS (and maximise overall compliance to the study protocol), patients randomised to switching to C-F NDS use will be offered a wide selection of different products (reflecting the most popular of those commercially available in each participating country) in order to choose the C-F NDS of their preference. Given that the population sample in DIASMOKE is mostly made of elderly patients, we will only offer devices that can ensure a likely user-friendly experience (ie, easy to refill consumables, prefilled consumables and heated tobacco devices). We expect that when participants are freely provided C-F NDS of their choosing they will be more likely to adopt the new technology and switch away from their own conventional cigarettes. Moreover, the study findings will not be product specific and unlikely to be limited in generalisability.
Our study has limitations.
First, due to the relatively long duration of the study (24 months), maintaining a sufficient level of subject retention may be a challenge. Nonetheless, trial attendance and retention is likely improved by inviting participants to return to the clinic for their free supply of study products and by offering a dedicated fast track approach for their outpatient clinic appointments.
Second, DIASMOKE results cannot be generalised to all patients with diabetes who smoke. We will recruit a (ambulatory) population of diabetic smokers who have been stably treated for T2DM. Therefore, the study protocol excludes smokers with untreated disease and T1DM smokers.
Last but not least, COVID-19 restrictions may slow down recruitment in some countries. A competitive recruitment strategy and staggered activation of clinical sites less impacted by the pandemic, will be implemented to minimise the possible negative effect.
Substantiation of the reduced risk potential of long-term C-F NDS use is virtually unexplored. Data from DIASMOKE will be an important addition to the growing body of evidence in the field of understanding the health impact of combustion-free nicotine delivery technologies and will provide valuable insights into the overall potential of these products to reduce the risk of cardiovascular disease in individuals, particularly patients with diabetes.
Collaborators The Steering Committee will take responsibility for the scientific validity of the study protocol, assessment of study quality and conduct as well as for the scientific quality of the final study report. All Committee members are independent of the funder and have no conflicts of interest. Committee members will be: Prof Pankaj Sharma, Chief Investigator, UK, Chair; Dr Chong Lim, Principal Investigator, UK; Prof Edward Franek, Principal Investigator, Poland, Deputy Chair; Dr Prof Francesco Purrello, Principal Investigator (Site 1), Italy; Prof Maurizio Di Mauro, Principal Investigator (Site 2), Italy; Prof Lorina Vudu, Principal Investigator, Moldova; Prof Farrukh Iqbal; Dr David Crook, Research Design Service, University of Brighton, UK; Data Monitoring & Safety Committee (see page 10 of this manuscript); Dr Jonathan Belsey, JB Medical (UK); Prof Aldo Calogero (Italy); Prof Sebastiano Battiato (Italy); Dr David Fluck (UK).
Contributors AK—manuscript drafting and revision. CR—study design, literature review, manuscript drafting and revision. SJ—manuscript drafting and revision. JB—sample size and statistical analysis plan. DC—study design, literature review, manuscript drafting and revision. PC—study design, manuscript drafting and revision. LV—manuscript revision. CWL—manuscript revision. FP—manuscript revision. MDM—manuscript revision. FI—manuscript revision. DF—manuscript revision. EF—manuscript drafting and revision. RP—manuscript drafting and revision. PS—study design, manuscript drafting and revision.
Funding This work was supported by ECLAT srl grant number COE1-05.
Disclaimer The contents, selection and presentation of facts, as well as any opinions expressed in the paper are the sole responsibility of the authors and under no circumstances shall be regarded as reflecting the positions of the Foundation for a Smoke-Free World. The Grantor had no role in the selection of the research topic, study design, or the writing of the paper or the project.
Competing interests None declared.
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.
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.