Article Text

Concomitant use of oral glucocorticoids and proton pump inhibitors and risk of osteoporotic fractures among patients with rheumatoid arthritis: a population-based cohort study
  1. Shahab Abtahi1,2,3,
  2. Johanna H M Driessen1,2,3,4,
  3. Andrea M Burden1,5,
  4. Patrick C Souverein2,
  5. Joop P van den Bergh6,7,8,
  6. Tjeerd P van Staa2,9,
  7. Annelies Boonen6,10,
  8. Frank de Vries1,2,3,11
  1. 1Department of Clinical Pharmacy and Toxicology, Maastricht University Medical Centre, Maastricht, The Netherlands
  2. 2Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands
  3. 3Cardiovascular Research Institute Maastricht (CARIM), Maastricht University, Maastricht, The Netherlands
  4. 4NUTRIM School for Nutrition and Translational Research in Metabolism, Maastricht University Medical Centre, Maastricht, The Netherlands
  5. 5Institute of Pharmaceutical Sciences, Department of Chemistry and Applied Biosciences, ETH-Zurich, Zurich, Switzerland
  6. 6Department of Internal Medicine, Division of Rheumatology, Maastricht University Medical Centre, Maastricht, The Netherlands
  7. 7Faculty of Medicine and Life Sciences, Hasselt University, Hasselt, Belgium
  8. 8Department of Internal Medicine, VieCuri Medical Centre, Venlo, The Netherlands
  9. 9Centre for Health Informatics, Division of Informatics, Imaging and Data Science, School of Health Sciences, Faculty of Biology, Medicine and Health, The University of Manchester, Manchester, UK
  10. 10Care and Public Health Research Institute (CAPHRI), Maastricht University, Maastricht, The Netherlands
  11. 11MRC Epidemiology Lifecourse Unit, Southampton General Hospital, Southampton, UK
  1. Correspondence to Professor Frank de Vries, Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, PO BOX 80082, 3508 TB, Utrecht, The Netherlands; f.devries{at}


Background Patients with rheumatoid arthritis (RA) commonly use oral glucocorticoids (GCs) and proton pump inhibitors (PPIs), both associated with osteoporotic fractures. We investigated the association between concomitant use of oral GCs and PPIs and the risk of osteoporotic fractures among patients with RA.

Methods This was a cohort study including patients with RA aged 50+ years from the Clinical Practice Research Datalink between 1997 and 2017. Exposure to oral GCs and PPIs was stratified by the most recent prescription as current use (<6 months), recent use (7–12 months) and past use (>1 year); average daily and cumulative dose; and duration of use. The risk of incident osteoporotic fractures (including hip, vertebrae, humerus, forearm, pelvis and ribs) was estimated by time-dependent Cox proportional-hazards models, statistically adjusted for lifestyle parameters, comorbidities and comedications.

Results Among 12 351 patients with RA (mean age of 68 years, 69% women), 1411 osteoporotic fractures occurred. Concomitant current use of oral GCs and PPIs was associated with a 1.6-fold increased risk of osteoporotic fractures compared with non-use (adjusted HR: 1.60, 95% CI: 1.35 to 1.89). This was statistically different from a 1.2-fold increased osteoporotic fracture risk associated with oral GC or PPI use alone. Most individual fracture sites were significantly associated with concomitant use of oral GCs and PPIs. Among concomitant users, fracture risk did not increase with higher daily dose or duration of PPI use.

Conclusions There was an interaction in the risk of osteoporotic fractures with concomitant use of oral GCs and PPIs. Fracture risk assessment could be considered when a patient with RA is co-prescribed oral GCs and PPIs.

  • osteoporosis
  • rheumatoid arthritis
  • glucocorticoids
  • epidemiology

Data availability statement

No additional data are available.

Statistics from

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.

Key messages

What is already known about this subject?

  • Oral glucocorticoids (GCs) and proton pump inhibitors (PPIs) are among the commonly prescribed medications for patients with rheumatoid arthritis (RA).

  • Oral GCs increase the risk of osteoporotic fractures by well-established biological mechanisms, including effects on bone, muscle strength and vision.

  • The association between PPIs and fracture risk is less well established, although biological mechanisms such as hypochlorhydria and reduced calcium absorption have been proposed.

What does this study add?

  • This is the first study to evaluate the association between concomitant use of oral GCs and PPIs and the risk of osteoporotic fractures in patients with RA, using a large primary care database.

How might this impact on clinical practice or future developments?

  • As we observed a 1.6-fold increased risk of osteoporotic fractures with concomitant use of oral GCs and PPIs in patients with RA, fracture risk assessment could be considered for these patients.


Rheumatoid arthritis (RA) is a common chronic musculoskeletal inflammatory disease with many complications, including an elevated risk of osteoporotic (OP) fractures.1–3 The contributors to increased fracture risk include the inflammatory process of RA and the pharmacological treatment of the disease, most importantly oral glucocorticoids (GCs). About one-quarter of patients with RA in the UK are current users of oral GCs.1 Patients with RA taking oral GCs have reduced bone mineral density (BMD) at the hip and vertebrae and up to a 35% increased 5-year fracture risk.1 4 This higher fracture risk with GCs is independent of the disease process and by known mechanisms, such as decreased bone formation, elevated bone resorption and ultimately reduced bone density.5–9

Apart from GCs, patients with RA frequently use other medications that could also be associated with fragility fractures. Non-steroidal anti-inflammatory drugs (NSAIDs) are routinely prescribed for patients with RA as analgesics, and proton pump inhibitors (PPIs) may be co-prescribed to reduce the gastrointestinal side effects. A randomised, double-blind, crossover trial showed that fractional 45calcium absorption was significantly decreased among elderly women using omeprazole (3.5%) versus placebo (9.1%), possibly because of hypochlorhydria.10 Observational studies have reported conflicting results. Some reported an increased risk of hip and vertebral fractures with PPI use, suggesting a causal effect,11–15 whereas others could not match the shape of the hazard function of PPI-induced fracture risk to calcium absorption hypothesis.16 17 Other mechanisms such as an increased fall risk due to hypomagnesaemia or explanations such as unmeasured confounding were also proposed to explain this association.16–20

A population-based study reported a 2.4-fold increased risk of hip fracture among concomitant users of both PPIs and high-dose oral GCs (≥15 mg prednisolone equivalent dose (PED)).16 But, to our knowledge, no studies have evaluated the effects of simultaneous use of both drugs on fracture risk in patients with RA, particularly in elderly patients who are regular users of PPIs.21 22 Thus, we sought to investigate the association between concomitant use of oral GCs and PPIs and the risk of OP fractures among patients with RA.


Data source

This was a retrospective cohort study based on the Clinical Practice Research Datalink (CPRD) GOLD database ( The CPRD is one of the largest databases of primary care data in the world, which contained medical records of 674 practices in the UK in 2013, representing 4.4 million active patients, which equalled 6.9% of the total population.23 It includes data on patient demographics, clinical diagnoses, prescription details, laboratory test results, specialist referrals and major outcomes since 1987, with continuing data collection. The CPRD has been well validated for a wide range of diseases, including hip and vertebral fractures.24 25

Study population

The study cohort included adults aged 50+ years and diagnosed with RA between 1 January 1997 and 31 December 2017. We used a validated algorithm to identify definite RA cases in the CPRD, which can detect 86% of the true RA cases (online supplementary table 1).26 27 The date of the first RA diagnosis during valid data collection defined the index date. Patients were followed until the occurrence of the outcome, the end of the study period, a patient’s transfer out of practice, death or the end of data collection, whichever came first. Following a new-user design, patients with a history of GC/PPI use during the 1 year before the index date and those with an OP fracture prior to the index date were excluded.

Exposure and outcome

Oral GCs and PPIs were the exposures of interest. From the RA index date, follow-up was divided into consecutive 30-day periods and exposure status was assessed time-dependently at the start of each period. A period was defined as current, recent or past use when the most recent prescription of oral GCs/PPIs was issued within 6 months, 7–12 months and >12 months before a period, respectively.7 11 12 16 28 Follow-up time was defined as non-use if no oral GC/PPI had ever been prescribed. Patients were allowed to move between exposure states during follow-up. Once a non-user had taken oral GCs/PPIs, he could never become a non-user again.

To evaluate a dose–response relationship and to replicate previous similar studies,11 17 28 current use of both drugs was further stratified in average daily and cumulative dose, and duration of treatment. All oral GC and PPI prescriptions were retrieved, and the prescribed quantity was extracted and converted into PED for GCs and omeprazole equivalent dose (OED) for PPIs using the WHO Anatomical Therapeutic Chemical classification system .29 Values for missing data on prescribed quantity were assigned the median value of all prescriptions. The cumulative amount of the drug prescribed in each follow-up period was estimated by summing all consecutive prescriptions since the index date. The average daily dose in each follow-up period was calculated by dividing the cumulative amount prescribed by the treatment time (ie, the time between the first oral GC/PPI prescription and the start date of a period of current use). Continuous duration of PPI use was determined at each period of current use using the prescribed quantity and written dosage information, allowing a gap of 30 days after the expected end date of a prescription.30 The outcome in this study was a first OP fracture after the RA index date, which included hip, clinically symptomatic vertebral, humerus, forearm, pelvic and rib fractures.1 28 31 32

Potential confounders

Body mass index (BMI), smoking status and alcohol use were determined at the index date. Age and history of comorbidities and comedications were determined time-dependently. Comorbidities included asthma, chronic obstructive pulmonary disease, ischaemic heart disease (including myocardial infarction), cerebrovascular disease, congestive heart failure, anaemia, peripheral vascular disease, gastro-oesophageal reflux disease, peptic ulcer disease, inflammatory bowel diseases (Crohn’s disease and ulcerative colitis), coeliac disease, hyperthyroidism, hypothyroidism, type 1 and 2 diabetes mellitus, chronic renal failure, ankylosing spondylitis, dementia, Parkinson’s disease, major infections (ie, sepsis, meningitis, and upper and lower respiratory tract infections) and malignant neoplasms (excluding non-melanoma skin cancers).33 Falls were measured in the 7–12 months prior to a period. Use of comedications in the 6 months prior included antihypertensives, anticoagulants, calcium/vitamin D, bisphosphonates, hormone replacement therapy, anticonvulsants, hypnotics/anxiolytics, antidepressants and antipsychotics. The following proxy indicators of RA severity were included: use of non-selective NSAIDs, cyclo-oxygenase-2 selective inhibitors, paracetamol, tramadol, opioids (stronger than tramadol) or conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) in the past 6 months.

Statistical analysis

Time-dependent Cox proportional-hazards models estimated the risk of OP fracture in patients with RA with concomitant current use of oral GCs and PPIs versus non-use. Also, the use of oral GCs alone and PPIs alone, and the recent and past use of oral GCs and PPIs (regardless of the use of the other drug) were compared with non-use. Individual exposure categories were statistically compared with a Wald test to detect between-group significance. Stratified analyses were conducted for various OP fracture sites. Potential confounders were incorporated into the model if the beta coefficient of the association changed by >5% or based on expert opinion.

Secondary analyses focused on average daily and cumulative dose of current GC use in relation to average daily dose and continuous duration of PPI use. Furthermore, three sensitivity analyses were performed. First, calcium/vitamin D and bisphosphonates were added to the model as confounders. They were not considered in the main analysis because of the accompaniment of their prescriptions with those of oral GCs and as we expected them to lie in the causal pathway of the intended association as mediators.34–36 Second, the main association was re-evaluated by including the prevalent users of GCs and PPIs. Finally, the association between PPI use and OP fractures was assessed among the primary cohort of patients with RA, by excluding only those with PPI use during the 1 year before the index date. Data were analysed using SAS V.9.4 (SAS Institute).


The study population included 12 351 patients with RA (figure 1). The mean age of concomitant users of oral GCs and PPIs at the index date was 67.5 years, 1.5 years younger than non-users (table 1). The mean duration of follow-up was 9.1 years for concomitant users and 5.1 years for non-users. About two-thirds of patients with RA were women (concomitant users: 67%; non-users: 70%). More than one-third of concomitant users were overweight, whereas 34% of non-users had a normal BMI. In the 6 months before the index date, 54% of concomitant users and 48% of non-users had taken non-selective NSAIDs. The average duration of drug use was 3.3 years for concomitant and single GC users, and 4.1 years for single PPI users.

Figure 1

Flowchart on establishment of patient population. CPRD, Clinical Practice Research Datalink; GC, glucocorticoid; OP,osteoporotic; PPI, proton pump inhibitor; RA, rheumatoid arthritis; TOD, transfer out of database date (ie, date the patient was transferred out of the practice); UTS, up to standard time (ie, date at which the practice data are deemed to be of research quality).

Table 1

Baseline characteristics of study population at index date, stratified by oral GC and PPI therapy status during follow-up (N=12 351)

Concomitant current use of oral GCs and PPIs in patients with RA was associated with a 1.6-fold increased risk of OP fractures compared with non-use of both drugs (adjusted HR (adj. HR): 1.60, 95% CI: 1.35 to 1.89; table 2). Both oral GC and PPI use alone had a 1.2-fold increased risk of OP fracture (adj. HR: 1.23, 95% CI: 1.03 to 1.47 (oral GC use alone); adj. HR: 1.22, 95% CI: 1.05 to 1.42 (PPI use alone)). The OP fracture risk associated with the current use of oral GCs or PPIs alone was statistically different from concomitant use. There was no significant increase in OP fracture risk in those patients who had stopped taking oral GCs or PPIs for more than 6 months (recent and past users) versus non-use. Considering calcium/vitamin D and bisphosphonates as confounders reduced the association to a 1.4-fold increased fracture risk for concomitant users and to a statistically non-significant risk for oral GC use alone versus non-use (online supplementary table 3).

Table 2

OP fracture risk by concomitant use of oral GCs and PPIs in patients with rheumatoid arthritis

Table 3 shows that among patients with RA, most OP fracture sites were statistically significantly associated with concomitant current use of oral GCs and PPIs versus non-use. With concomitant current use of oral GCs and PPIs, we observed a 1.5-fold increased risk of hip fracture, a 2.8-fold increased risk of clinical vertebral fracture, a 2.5-fold increased risk of pelvic fracture and a 4-fold increased risk of rib fracture. Risks of fracture of the humerus or forearm were not increased.

Table 3

OP fracture risk by concomitant use of oral GCs and PPIs in patients with rheumatoid arthritis, stratified by fracture type

Table 4 shows the stratification of concomitant oral GC and PPI use by average daily doses of GCs and then substratification by average daily doses and continuous duration of PPI use. There was no increased fracture risk with increasing PPI daily doses. Under all strata of GC use, short-term PPI use (≤1 year) was associated with higher fracture risk, but there was no association between long-term PPI use (>1 year) and OP fractures. When concomitant use of GCs and PPIs was stratified by cumulative GC use and then substratified by PPI use, similar associations were observed (online supplementary table 2).

Table 4

OP fracture risk by average daily dose of oral GC use in patients with rheumatoid arthritis, stratified by average daily dose and continuous duration of PPI use

The second sensitivity analysis including prevalent users of GCs and PPIs (N=21 650) resulted in similar estimates to the main analyses (online supplementary table 4). In the third sensitivity analysis (N=14 602), current PPI use was associated with a 1.3-fold increased risk of OP fractures (adj. HR: 1.30, 95% CI: 1.15 to 1.47) versus non-use (online supplementary table 5).


Concomitant use of oral GCs and PPIs was associated with an increased risk of OP fractures compared with non-use in patients with RA. This was significantly higher when compared with the single use of oral GCs or PPIs. Increased fracture risk associated with concomitant GC and PPI use was observed for fractures of the hip, clinical vertebrae, pelvis and ribs, but not for those of the humerus or forearm. Among concomitant users, there was no increased OP fracture risk with higher daily dose or longer duration of PPI use.

This is the first study, to our knowledge, that looked into the association between concomitant use of GCs and PPIs and the risk of OP fracture in patients with RA . A Dutch population-based study found a 1.3-fold to 2.4-fold increased risk of hip/femur fracture with concomitant use of PPIs and various daily doses of oral GCs.16 This is in line with our finding for the concomitant current use of GCs and PPIs (adj. HR: 1.60) and most of the strata of concomitant use in table 4. However, their reference group was different and limited to never PPI users. Moreover, they focused on 18+ general population, whereas we included patients with RA aged 50+ years with higher baseline fracture risks.

Our results regarding the higher fracture risk with PPI use are partly in line with several previous observational studies.11 12 15–17 A meta-analysis of observational studies in non-RA patients reported increased risk of hip and spine fracture with PPI use (relative risks of 1.30 and 1.56, respectively),14 which is comparable with adj. HR of 1.30 for current PPI use and OP fractures in our study. However, a recent study in patients with RA did not reveal a higher risk of OP fractures with PPI use, which was attributed to higher use of bisphosphonates among PPI users.22 Previous studies found stronger associations with higher daily doses of PPIs11 or with 7 years of PPI use and fracture risk,12 whereas another older study that used the same data source but a different reference group did not report any dose–response or duration–response relationships at all.17 Our findings (ie, no specific trend with longer duration or higher daily doses of PPI use) are comparable to the latter study.

Our findings in the single GC use group were generally consistent with the literature. Previous observational studies have reported increased OP fracture risks in patients with RA with current GC use between 43% and 70%, higher than the 23% increased risk that we found.1 31 We used a different reference group (non-users of both GCs and PPIs), which may also explain the unexpected lack of statistical significance for a higher risk of clinical vertebral fracture with current GC use alone.

The magnitude of the association between concomitant GC and PPI use and the risk of OP fracture may indicate an additive effect of the individual drugs rather than a synergistic effect. This was suggested by a significantly higher fracture risk with concomitant GC and PPI use compared with monotherapy with either drug and as the observed HRs seem to be additive. This may be related to different biological mechanisms of GCs and PPIs acting on osteoporosis or falling. The effect of GCs on bone is mostly via decreased bone formation and interference with active bone remodelling sites.1 6 8 9 But additionally, GCs might increase the fracture risk by inducing muscle atrophy or cataract especially with higher doses and in long-term use.37–39 Previous studies have shown that the onset and offset of the effects of GCs on fracture risk are rather rapid, which is supported by our results.7 31 40 Similar to GCs, the positive association of fracture risk with PPI use quickly subsided when the patient discontinued the treatment (after 6 months). But for PPIs, underlying pharmacological effects on fracture are not well understood.41 42 The US Food and Drug Administration published a drug safety communication for a possible increased fracture risk with PPI use in 2011, which remained unchanged to date and was based on evidence from observational studies.43 This was later criticised for not being supported by a clear biological mechanism.44

Various pharmacological mechanisms have been proposed to explain the PPI use and fracture risk association. Reduced intestinal absorption of calcium was previously suggested due to induced hypochlorhydria by PPI therapy and the effect on bone quality.10 However, a more recent trial found no BMD changes after 52 weeks and non-significant changes in bone turnover markers after 26 weeks with dexlansoprazole or esomeprazole use.45 An alternative mechanism is an increased falling risk due to muscle weakness and drowsiness, caused by malabsorption of magnesium or vitamin B12.18–20 46 47 Long-term PPI therapy (≥1 year) in elderly women was shown to significantly reduce serum vitamin B12 levels and double the 5-year risk of injurious falling-related and fracture-related hospitalisation.46 But the design of this study did not consider proper timing of the exposure and outcome, which limits its interpretation. A third mechanism is effects on osteoclasts to increase bone resorption by PPIs.48 Finally, methodological explanations for the observed associations include selection bias and/or unknown confounding.16 17 44 Significant association only with short-term PPI use and no specific trend with increasing daily doses do not fit into any of the proposed mechanisms mentioned above. As we used different strategies in design and analysis to avoid potential sources of bias and to adjust for confounding, and when the GC findings are supported by previous literature with well-known biological mechanisms, the mere explanation of the PPI results by unmeasured confounding would be difficult. Hence, more research is recommended to elaborate on the exact biological mechanism of PPIs on bone.

This study had several strengths. We used data from the CPRD, which is one of the world’s largest primary care databases. Our study had a substantial mean duration of follow-up (9.1 years for concomitant users). To bring more insight into the observed association, we stratified GC and PPI use by recency of use, average daily and cumulative dose, and duration of treatment. Furthermore, all analyses were performed time-dependently, incorporating all follow-up times, to avoid time-related biases. There were also several limitations. Biological therapies, especially during hospitalisation, and some RA severity indices (eg, the disease activity score using 28 joints (DAS-28)) were not adequately captured in the CPRD as a general practice database, which might have introduced confounding by indication or disease severity. Patients with higher disease activity may have an elevated risk of fracture and be more prone to receive oral GCs/PPIs. Also, an improved clinical status might have led to both discontinuation of drug(s) and lower fracture rates. To partly overcome this, we statistically adjusted our analyses for six indicators of RA severity, including analgesics and csDMARDs. We cannot confirm the actual use of medications as we only had prescribing information, and GCs and PPIs are often prescribed on an as-needed basis. The over-the-counter use of PPIs was also not captured. However, with an average duration of use of >3 years, repeated prescriptions are indicators of actual use. Finally, the number of vertebral fractures might be underestimated, as some of them might not immediately come into clinical attention.49 50 This might virtually increase the HRs for vertebral fractures due to detection bias.35

In conclusion, there was an interaction in the risk of OP fracture with concomitant use of oral GCs and PPIs. This increased risk seems to emerge from separate mechanisms of action of GCs and PPIs on bone or falling risk. Considering the increasing life expectancies and high consumption of PPIs among elderly patients, fracture risk assessment could be considered when a patient with RA is co-prescribed oral GCs and PPIs.

Data availability statement

No additional data are available.

Ethics statements

Ethics approval

This study was reviewed and approved by the Independent Scientific Advisory Committee of the Clinical Practice Research Datalink (reference 19_018R), which is responsible for reviewing protocols for scientific quality.


The authors would like to acknowledge Keele University’s Prognosis and Consultation Epidemiology Research Group who have given us permission to use the morbidity definitions/categorisations lists (©2014). The copyright of the morbidity definitions/categorisations lists (©2014) used in this paper is owned by Keele University, the development of which was supported by the Primary Care Research Consortium. For access/details relating to the morbidity definitions/categorisation lists (©2014), please go to The authors would also like to acknowledge Ard van Veelen from Maastricht University Medical Centre for double-checking the accuracy of tables and figures.


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.

  • Lay summary

    Disclaimer : This is a summary of a scientific article written by a medical professional (“the Original Article”). The Summary is written to assist non medically trained readers to understand general points of the Original Article. It is supplied “as is” without any warranty. You should note that the Original Article (and Summary) may not be fully relevant nor accurate as medical science is constantly changing and errors can occur. It is therefore very important that readers not rely on the content in the Summary and consult their medical professionals for all aspects of their health care and only rely on the Summary if directed to do so by their medical professional. Please view our full Website Terms and Conditions.
    Copyright © 2021 BMJ Publishing Group Ltd & European League Against Rheumatism. Medical professionals may print copies for their and their patients and students non commercial use. Other individuals may print a single copy for their personal, non commercial use. For other uses please contact our Rights and Licensing Team.

  • 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.


  • Handling editor Josef S Smolen

  • Twitter @Sh_Abtahi, @ETH_PharmEpi

  • Correction notice This article has been corrected since it published Online First. The corresponding address has been corrected.

  • Contributors SA, AMB, AB and FV contributed to intellectual concept and design of the research. JD and PS carried out the data acquisition. SA and JD conducted the data analysis. All authors contributed to data interpretation. SA wrote the manuscript. All authors were involved in editing or quality control. All authors are accountable for their own contribution, in addition to all aspects of the manuscript. All authors have approved the (re)submitted version of the manuscript.

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

  • Competing interests JvB reports grants from Amgen, Eli Lilly and UCB, outside the submitted work. AB received research grants to her department from Celgene and AbbVie, and speakers’ fees or honoraria from Sandoz, UCB, Lilly, Novartis and Biogen, and Gilead also to her department, not related to this study. FV supervises three PhD students who are employed with F. Hoffmann La Roche (Basel, Switzerland/Welwyn Garden City, UK). He has not received any fees or reimbursements for this, and the topics of their PhD thesis are not related to the current study proposal. SA, JD, AMB, PCS and TPvS have no competing interests.

  • 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.

Linked Articles