Article Text

Associations of ophthalmic and systemic conditions with incident dementia in the UK Biobank
Free
  1. Xianwen Shang1,2,
  2. Zhuoting Zhu1,2,
  3. Yu Huang1,2,
  4. Xueli Zhang1,2,
  5. Wei Wang3,
  6. Danli Shi3,
  7. Yu Jiang3,
  8. Xiaohong Yang1,
  9. Mingguang He1,3,4
  1. 1 Guangdong Eye Institute, Department of Ophthalmology, Guangdong Provincial People’s Hospital, Guangdong Academy of Medical Sciences, Guangzhou, Guangdong, People's Republic of China
  2. 2 Guangdong Cardiovsacular Institute, Guangzhou, Guangdong, People's Republic of China
  3. 3 State Key Laboratory of Ophthalmology, Sun Yat-Sen University Zhongshan Ophthalmic Center State Key Laboratory of Ophthalmology, Guangzhou, China
  4. 4 Centre for Eye Research Australia, Melbourne, Victoria, Australia
  1. Correspondence to Dr Mingguang He, Department of Ophthalmology, Guangdong Academy of Medical Sciences, Guangzhou, China; mingguang_he{at}yahoo.com; Dr Xiaohong Yang; syyangxh{at}scut.edu.cn

Abstract

Aims To examine independent and interactive associations of ophthalmic and systemic conditions with incident dementia.

Methods Our analysis included 12 364 adults aged 55–73 years from the UK Biobank cohort. Participants were assessed between 2006 and 2010 at baseline and were followed up until the early of 2021. Incident dementia was ascertained using hospital inpatient, death records and self-reported data.

Results Over 1 263 513 person-years of follow-up, 2304 cases of incident dementia were documented. The multivariable-adjusted HRs (95% CI) for dementia associated with age-related macular degeneration (AMD), cataract, diabetes-related eye disease (DRED) and glaucoma at baseline were 1.26 (1.05 to 1.52), 1.11 (1.00 to 1.24), 1.61 (1.30 to 2.00) and (1.07 (0.92 to 1.25), respectively. Diabetes, heart disease, stroke and depression at baseline were all associated with an increased risk of dementia. Of the combination of AMD and a systemic condition, AMD-diabetes was associated with the highest risk for incident dementia (HR (95% CI): 2.73 (1.79 to 4.17)). Individuals with cataract and a systemic condition were 1.19–2.29 times more likely to develop dementia compared with those without cataract and systemic conditions. The corresponding number for DRED and a systemic condition was 1.50–3.24. Diabetes, hypertension, heart disease, depression and stroke newly identified during follow-up mediated the association between cataract and incident dementia as well as the association between DRED and incident dementia.

Conclusions AMD, cataract and DRED but not glaucoma are associated with an increased risk of dementia. Individuals with both ophthalmic and systemic conditions are at higher risk of dementia compared with those with an ophthalmic or systemic condition only.

  • glaucoma
  • contact lens
  • macula
  • epidemiology

Data availability statement

Data are available in a public, open access repository.

Statistics from Altmetric.com

Introduction

Globally 43.8 million people lived with dementia in 20161 and this number is estimated to increase to 152 million by 2050.2 Dementia accounted for 2.4 million deaths and 28.8 million disability-adjusted life-years in 2016.1 The epidemic of dementia imposes a tremendous burden on the healthcare systems globally. As there is no effective treatment to stop the progression of dementia, it is imperative to identify modifiable factors for the prevention of dementia.3 4

Systemic conditions including obesity, depression, hypertension, diabetes, heart disease and stroke have been identified as the leading risk factors of dementia.4–7 Recent evidence has suggested that vision impairment is one of the first manifestations of the development of dementia.8 9 Meanwhile, loss of visual input may lead to reduced activation in central sensory pathways, resulting in an increased risk of cognitive load and brain structure damage,10 11 which accelerates the development of dementia.12 Recent studies have linked main causes of vision impairment including age-related macular degeneration (AMD), cataract, diabetic retinopathy (DR) and glaucoma to dementia or cognitive impairment;13–18 however, the results are inconsistent and the studies are limited by small sample size or case–control design. Ophthalmic and systemic conditions are frequently clustered in pairs,4 19 20 as they are all heavily age-related conditions. Therefore, an increased risk of dementia associated with ophthalmic conditions may be due to their positive associations with systemic conditions such as obesity and cardiovascular disorders. However, whether ophthalmic conditions are associated with a higher incidence of dementia independent of these systematic conditions and whether this association is mediated by newly developed systematic conditions is unclear.21–23 It is also unknown whether an ophthalmic condition with a systemic condition as a comorbidity is a stronger predictor of dementia than an ophthalmic or systemic condition only is.

Using the data of the UK Biobank study, we sought to determine the independent associations between ophthalmic conditions and incident dementia. We then examined whether the clustering of ophthalmic and systemic conditions was associated with an increased risk of dementia. The potential mediation effects of newly developed systemic conditions on the association between ophthalmic conditions and incident dementia were also evaluated.

Methods

Study population

The UK Biobank is a population-based cohort of more than 500 000 participants aged 40–73 years who attended 1 of the 22 assessment centres throughout the UK between 2006 and 2010.24 The UK Biobank study design and population has been detailed previously elsewhere.24

Ophthalmic conditions

Ophthalmic conditions were defined based on the question: ‘Has a doctor ever told you that you have any of the following problems with your eyes?’ Individuals were classified as AMD, cataract, diabetes-related eye disease (DRED) and glaucoma for those who selected the corresponding item from a predefined list of answers to the question. DERD includes DR, diabetic AMD, diabetes-related cataract and diabetes-related glaucoma. Additional prevalent cases of ophthalmic conditions were identified using hospital inpatient records.

Systemic conditions

Systemic conditions that are well-known risk factors of dementia were defined based on self-reported data or interviews. Participants were asked whether they had ever been told by a doctor that they had certain common medical conditions, including heart attack, angina, stroke, high blood pressure and diabetes. Heart attack and angina were combined as heart disease. Depression was recorded during the interview with a research nurse. Body mass index (BMI) was computed based on measured height and weight and obesity was defined as BMI ≥30 kg/m2.25 Although depression is not a systemic condition, we included depression as an important non-ophthalmic condition in the analysis.

Newly developed systemic conditions during follow-up (before the onset of dementia) were identified using hospital inpatient records and self-reported data during follow-up (the codes for each condition can be seen in online supplemental table e-1).

Ascertainment of incident dementia

Dementia cases in the UK Biobank Study were ascertained using hospital inpatient records, death registers and self-reported data. The codes for international classification diseases were used to identify dementia (detailed in online supplemental table e-2). The earliest recorded code date regardless of sources was used as the onset date of dementia. Person-years were calculated from the date of baseline assessment to the date of onset dementia, date of death or the end of follow-up (31 December 2020 for England and Wales and 18 January 2021 for Scotland), whichever came first.

Covariates

Sociodemographic factors including age, sex, education and income were self-reported. Participants completed a detailed questionnaire on a touch screen computer about their lifestyle, including smoking status, alcohol consumption and dietary intakes. Questions about walking, moderate physical activity and vigorous physical activity, which were similar to those used in the short form of the International Physical Activity Questionnaire, were used to estimate excess metabolic equivalent hours per week of physical activity during work and leisure time.

Statistical analysis

Data were expressed as frequency (percentage) and means±SDs by groups of individuals who had AMD, cataract, DRED and glaucoma and who were free of any ophthalmic conditions at baseline.

Cox proportional hazard regression models were used to examine the association between ophthalmic conditions and incidence of all-cause dementia, Alzheimer’s disease and vascular dementia. The following models were tested: (1) age and gender and (2) model 1 plus education, income, smoking, alcohol consumption, physical activity, BMI, cholesterol, glucose and intakes of cooked vegetables, raw vegetables, fresh fruits and dried fruits at baseline. We then examined associations of the clustering of ophthalmic and systemic conditions with incident dementia. For example, HRs (95% CIs) for dementia associated with AMD only, stroke only and both AMD and stroke were estimated with individuals free of AMD and stroke as the reference. Dietary intake, lifestyle, education and BMI were adjusted for in the model because these factors have been linked to dementia in previous studies.4 26 The association between DRED and incident dementia was conducted among patients with diabetes.

The potential mediation effects of newly developed systemic conditions during follow-up (before onset dementia) on the association between ophthalmic conditions and incident all-cause dementia were estimated using Cox proportional hazards regression models. Percentages of total effects mediated by systemic conditions were computed.

We then tested whether DR was a risk factor of dementia in patients with diabetes. DR was defined as individuals with DRED but not with AMD, cataract or glaucoma (those with DERD and any of AMD, cataract or glaucoma were excluded from the analysis). Sensitivity analysis was conducted to examine the association between ophthalmic conditions and incident dementia by excluding individuals who developed dementia in the first 5 years of follow-up.

Data analyses were conducted using SAS V.9.4 for Windows (SAS Institute) and all p values were two sided with statistical significance set at <0.05.

Results

Baseline characteristics

At baseline, 502 505 participants were assessed. After excluding individuals missing data on ophthalmic conditions (n=3 27 891), those with prevalent dementia (n=112) or those aged <55 years (n=62 138), 112 364 adults (54.0% females) aged 55–73 years (mean±SD: 62.4±4.1) were included in the final analysis (online supplemental figure e-1). The baseline characteristics of the participants are presented in table 1. As shown in online supplemental table e-3, individuals with incident dementia were more likely to be older and males and have APOE4 compared with those without incident dementia.

Table 1

Baseline characteristics of participants according to ophthalmic conditions

Incidence of dementia

Over 1 263 513 person-years of follow-up (median (IQR) length of follow-up: 11.0 (10.7–11.7) years), 2304 new cases of incident all-cause dementia, 945 new cases of Alzheimer’s disease and 513 new cases of vascular dementia were documented.

Ophthalmic conditions and incident dementia

Of participants with AMD at baseline, 2.94 cases per 1000 person-years developed all-cause dementia compared with 1.78 of those without AMD (multivariable-adjusted HR (95% CI): 1.26 (1.05 to 1.52)). The multivariable-adjusted HRs (95% CI) for dementia associated with cataract and DRED were 1.11 (1.00 to 1.24) and 1.61 (1.30 to 2.00), respectively. Glaucoma was associated with a higher risk of vascular dementia (HR (95% CI): 1.48 (1.13 to 1.94)), but not with Alzheimer’s disease (0.89 (0.70 to 1.14)) or all-cause dementia (1.07 (0.92 to 1.25)) after adjustment for confounders (figure 1, online supplemental figures e–2 and 3).

Figure 1

The risk for dementia associated with ophthalmic conditions. Incidence of dementia represents cases per 1000 person-years. HR (95% CI) for incident dementia associated with ophthalmic conditions was estimated using Cox proportional regression models adjusted for age, gender, education, income, cooked vegetables intake, raw vegetables intake, fresh fruits intake, dried fruits intake, smoking, alcohol consumption, physical activity, BMI, cholesterol and glucose at baseline. The analysis for the association between DRED and dementia was conducted among patients with diabetes. AMD, age-related macular degeneration; BMI, body mass index; DRED, diabetes-related eye disease.

Clustering of ophthalmic and systemic conditions and incident dementia

Stroke only, heart disease only, diabetes only, hypertension only and depression only but not obesity only at baseline were independently associated with an increased risk of dementia. AMD with any systemic condition as a comorbidity was associated with an increased risk of dementia and the risk was larger than AMD only or a systemic condition only. AMD and diabetes among the combination of AMD and a systemic condition had the highest risk for incident dementia (HR (95% CI): 2.73 (1.79 to 4.17)), which was followed by AMD and obesity (1.88 (1.35 to 2.63), figure 2.

Figure 2

The risk for incident dementia associated with age-related macular degeneration (AMD) and systemic conditions. Incidence of dementia represents cases per 1000 person-years. HR (95% CI) for incident dementia associated with AMD and systemic conditions was estimated using Cox proportional regression models. Multivariable analysis was adjusted for age, gender, education, income, cooked vegetables intake, raw vegetables intake, fresh fruits intake, dried fruits intake, smoking, alcohol consumption, physical activity, body mass index, cholesterol and glucose at baseline. Central squares of each horizontal line represent the HR for each subgroup. Horizontal lines indicate the range of the 95% CI and the vertical dash lines indicate the HR of 1.0.

Cataract only was not independently associated with an increased risk of incident dementia. Cataract with a comorbidity being stroke (HR (95% CI): 2.00 (1.40 to 2.84)), heart disease (1.77 (1.43 to 2.21)), hypertension (1.36 (1.17 to 1.58)), diabetes (2.19 (1.78 to 2.70)), depression (2.29 (1.74 to 3.01)) or any systematic condition (1.57 (1.36–1.81)) was associated with an increased risk of dementia (figure 3).

Figure 3

The risk for incident dementia associated with cataract and systemic conditions. Incidence of dementia represents cases per 1000 person-years. HR (95% CI) for incident dementia associated with cataract and systemic conditions was estimated using Cox proportional regression models. Multivariable analysis was adjusted for age, gender, education, income, cooked vegetables intake, raw vegetables intake, fresh fruits intake, dried fruits intake, smoking, alcohol consumption, physical activity, body mass index, cholesterol and glucose at baseline. Central squares of each horizontal line represent the HR for each subgroup. Horizontal lines indicate the range of the 95% CI and the vertical dash lines indicate the HR of 1.0.

Of ophthalmic conditions, DRED only had the largest risk for incident dementia. Individuals with DRED and a systemic condition as a comorbidity were 1.50–3.24 times more likely to develop dementia compared with those without DRED and the corresponding systemic condition. These associations were independent of glucose and duration of diabetes (figure 4).

Figure 4

The risk for incident dementia associated with diabetes-related eye disease (DRED) and systemic conditions. Incidence of dementia represents cases per 1000 person-years. HR (95% CI) for incident dementia associated with DRED and systemic conditions was estimated using Cox proportional regression models. The analysis was conducted among patients with diabetes. Multivariable analysis was adjusted for age, gender, education, income, cooked vegetables intake, raw vegetables intake, fresh fruits intake, dried fruits intake, smoking, alcohol consumption, physical activity, body mass index, cholesterol, glucose and duration of diabetes at baseline. Central squares of each horizontal line represent the HR for each subgroup. Horizontal lines indicate the range of the 95% CI and the vertical dash lines indicate the HR of 1.0.

Glaucoma only was not significantly associated with incident all-cause dementia. However, when it co-occurred with stroke, heart disease, diabetes, hypertension or depression, the risk for dementia increased significantly and was larger than the risk associated with the corresponding systemic condition only (online supplemental figure e-4).

Mediation analysis

The percentage (95% CI) for the association between cataract and incident dementia explained by diabetes, hypertension, heart disease, depression and stroke was 9.2% (95% CI: 2.9% to 25.4%), 7.1% (2.7% to 17.0%), 9.1% (3.3% to 22.5%), 12.0% (4.5% to 28.4%) and 22.4% (8.0% to 48.9%), respectively. This association was fully mediated by these significant mediators combined (65.4% (5.3% to 98.5%). Hypertension, stroke, heart disease, depression and diabetes together explained 10.0% (9.4% to 31.5%) of the association between DRED and incident dementia (online supplemental figure e-5).

Sensitivity analysis

Individuals with DR than those without had an increased risk of dementia in those with diabetes at baseline (HR (95% CI): 1.65 (1.25 to 2.16), online supplemental figure e-6). Among individuals by excluding those who developed dementia in the first 5 years of follow-up, the multivariable-adjusted HRs (95% CIs) for dementia associated with AMD, cataract, DRED and glaucoma were 1.28 (1.06 to 1.55), 1.13 (1.01 to 1.25), 1.56 (1.24 to 1.95) and 1.09 (0.94 to 1.28), respectively (online supplemental figure e-7).

The HRs (95% CI) for dementia associated with 1 and ≥2 ophthalmic conditions were 1.11 (1.01 to 1.22) and 1.55 (1.31 to 1.82), respectively (online supplemental figure e-8).

The incidence of dementia increased with the increasing number of ophthalmic conditions across the number of systemic conditions (online supplemental figure e-9).

Discussion

This prospective cohort study with a large sample size demonstrated that AMD, cataract and DRED but not glaucoma were independently associated with an increased risk of all-cause dementia. Participants with an ophthalmic condition and a systemic condition had a higher risk for dementia compared with those with an ophthalmic or a systemic condition only. Newly developed hypertension, diabetes, stroke, heart disease and depression mediated the association between cataract/DRED and dementia.

Our study is consistent with several cohort studies showing that AMD was associated with an increased risk of dementia. Data from the Adult Changes in Thought (ACT) Study of 3877 participants aged ≥65 years showed that AMD (HR (95% CI): 1.20 (1.02 to 1.40)) was a risk factor of dementia.21 A longitudinal case–control study of 29 958 participants aged ≥65 years reported that AMD was independently associated with an increased risk of subsequent Alzheimer’s disease (HR (95% CI): 1.44 (1.26 to 1.64)) over 4 years.27 In a cohort study of 308 340 Korean participants aged ≥50 years, those with AMD had a higher risk for Alzheimer’s disease (HR (95% CI): 1.48 (1.25 to 1.74)) over 8 years compared with those without.22 Age is an important marker for dementia,28 however, these studies included participants with a wide range of age. Our further analysis demonstrated that the positive association between AMD and incident dementia was even significant when the analysis was restricted to those aged 60–70 years. In contrast, a prospective population-based study of 1438 Dutch adults aged ≥75 years demonstrated that AMD at baseline was associated with an increased risk of incident Alzheimer’s disease before, but not after adjustment for smoking and atherosclerosis.13 Two case–control cohort studies did not find a significant association between AMD and incident dementia.15 16 The case–control design and diagnosis of AMD using hospital admission records only may partly explain why the results of these studies are inconsistent with our findings.22

The mechanisms for the positive association between ophthalmic conditions and dementia are largely unknown, but there are several potential pathways for this association. First, ophthalmic conditions are associated with well-known risk factors of dementia including diabetes, stroke, heart disease, hypertension and depression.4 19 20 Second, ophthalmic conditions and dementia have many share risk factors including older age, low levels of education, smoking and physical inactivity.19 20 29 30 Importantly, vision deprivation may result in reduced activation in central sensory pathways, which is associated with a higher risk of cognitive load and brain structure damage.10 11 Of note, there are aetiological differences in types of dementia including Alzheimer’s disease and vascular dementia, such that the potential effects of ophthalmic conditions might differ among dementia types.

Only several cohort studies have reported the association between cataract and incident dementia. A case–control prospective study showed that cataract was associated with an increased risk of dementia over 14 years.15 Data from the ACT Study showed that cataract was not significantly associated with dementia (HR (95% CI): 1.21 (0.93 to 1.57)) over 8 years.21 In the Shanghai Aging Study of 1659 community residents aged ≥60 years, no significant association between cataract and incident dementia was observed.23 The non-significant association may be attributed to the small number of cataract cases in these two studies.21 We found that hypertension, diabetes, depression, stroke and heart disease identified during follow-up combined fully mediated the association suggesting that the risk for dementia associated with cataract may depend on these conditions.

Diabetes is a well-known risk factor for dementia,31 but DRED has been linked to dementia in only two prospective studies.15 21 DR was associated with an increased risk of dementia in the ACT Study.21 Similarly, a retrospective case–control study of 126 650 participants showed that those with DR were 1.42 (95% CI: 1.30 to 1.55) times more likely to develop dementia than those without.15 Our study is consistent with these studies showing that DRED was an independent risk factor of dementia. The association between DRED and dementia may reflect the association between diabetes and dementia. However, these previous studies cannot distinguish whether the risk for dementia associated with DRED was due to diabetes or eye disease. Our further analysis demonstrated that DRED was positively associated with incident dementia in individuals with diabetes. We also observed that hypertension, stroke, heart disease and depression combined did not fully mediate the association highlighting the importance of DRED as an independent risk factor for dementia.

The association between glaucoma and incident dementia has been investigated in some previous studies with conflicting results.14 15 21 23 32 33 A positive association between glaucoma and dementia has been reported in a case–control study in Taiwan32 and one in Korea.33 Data from the Shanghai Aging Study showed that individuals with glaucoma had a higher risk of dementia over 5 years than those without (HR (95% CI): 2.38 (1.08 to 5.23)).23 However, a retrospective cohort study of older American adults aged ≥68 years demonstrated that glaucoma was inversely associated with the risk of dementia over 14 years.15 Another prospective cohort study in the USA showed that glaucoma was not significantly associated with the incidence of dementia.21 A case–control study of 11 721 Danish participants aged 61–77 years did not find a significant association between glaucoma and dementia over 5 years.14 Our study is in line with the previous studies from Europe and the USA demonstrating that glaucoma was not an independent risk factor for dementia. Further analysis had demonstrated that glaucoma may have the potential to increase the risk of vascular dementia but not Alzheimer’s disease.

No previous study has examined the association between combinations of ophthalmic and systemic conditions with incident dementia. As both ophthalmic and systemic conditions examined in our study are aged-related conditions, they are highly likely to be clustered in pairs.4 19 20 29 We found that all combinations of an ophthalmic condition and a systemic condition were associated with an increased risk of dementia. This risk was greater than that associated with an ophthalmic or a systemic condition only suggesting that there might be an additive effect of ophthalmic and systemic conditions on the development of dementia. Underlying mechanisms are unclear, but impairment of multiple organs than a single is more likely to enlarge the risk of dementia. Only one previous study examined multiple ophthalmic conditions and showed that individuals with both glaucoma and cataract were over three times more likely to develop dementia compared with those without ophthalmic conditions.23 However, this study is limited by the small sample in the subgroup of glaucoma and cataract (n=29). Our study with a much larger sample size demonstrated that individuals with ≥2 ophthalmic conditions had a higher risk of dementia than those with one ophthalmic condition. Notably, individuals with ≥2 ophthalmic and ≥2 systemic conditions were almost three times more likely to develop dementia in our study. This indicates that individuals with both ophthalmic and systemic conditions should be targeted for the prevention and screening of dementia in older adults.

To our knowledge, this is the first prospective study to examine the association between the clustering of ophthalmic and systemic conditions and incident dementia. Our study has several potential limitations. First, ophthalmic conditions were defined based on self-reported and inpatient record data, which may result in misclassification errors. However, these data are more likely to underestimate the prevalence of ophthalmic conditions,34–36 which is more likely to have biased these findings towards the null. Second, some cases of dementia might not be captured in the medical records or death registers. However, previous research has demonstrated good agreement of dementia case ascertainment with primary care records.37 Third, dementia documented during follow-up may have occurred before eye diseases, as the prodromal period of dementia can last decades. Therefore, our study may not infer causal relationships although the results by excluding individuals who developed dementia in the first 5 years of follow-up are consistent with those for dementia during the whole follow-up period. Fourth, this sample was restricted to a subgroup of the UK Biobank cohort who had self-reported data on ophthalmic conditions. Individuals excluded from the analysis were more likely to be younger and have a lower prevalence of diabetes, cardiovascular disease and depression compared with those included in the analysis (online supplemental table e-4). However, a previous study has demonstrated that findings regarding exposure–disease relationships may be generalised to other populations.38

In conclusion, AMD, cataract and DRED but not glaucoma are associated with an increased risk of dementia. Individuals with both ophthalmic and systemic conditions are at higher risk of dementia compared with those with an ophthalmic or systemic condition only.

Data availability statement

Data are available in a public, open access repository.

Ethics statements

Ethics approval

The UK Biobank Study’s ethical approval has been granted by the National Information Governance Board for Health and Social Care and the NHS North West Multicenter Research Ethics Committee (REC reference: 16/NW/0274). All participants provided informed consent through electronic signature at baseline assessment. The present study was conducted under application number 62443 of the UK Biobank resource.

Acknowledgments

This research was conducted using the UK Biobank resource. The authors thank the participants of the UK Biobank.

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

  • Contributors XS, XY and MH conceived the study. XS and ZZ did the literature search, data analysis and data interpretation. DS and YJ did data acquisition. XS drafted the initial manuscript. All authors critically revised the manuscript.

  • Funding This work was supported by the Fundamental Research Funds of the State Key Laboratory of Ophthalmology, Project of Investigation on Health Status of Employees in Financial Industry in Guangzhou, China (Z012014075) and Science and Technology Program of Guangzhou, China (202002020049). MH received support from the University of Melbourne at Research Accelerator Program and the CERA Foundation. The Centre for Eye Research Australia received Operational Infrastructure Support from the Victorian State Government.

  • Disclaimer The sponsor or funding organisation had no role in the design, conduct, analysis or reporting of this study. The funding sources did not participate in the design and conduct of the study; collection, management, analysis and interpretation of the data; preparation, review or approval of the manuscript and decision to submit the manuscript for publication.

  • 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.

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.