We use cookies to improve our service and to tailor our content and advertising to you.More infoYou can manage your cookie settings via your browser at any time. To learn more about how we use cookies, please see our cookies policy.
We would like to thank Hakim et al for the opportunity to continue the debate around the prevalence and definitions of conditions associated with joint hypermobility. Coming to anything like a true population prevalence for these conditions is still a long way off due to both the huge under-diagnosis (1) and ongoing evolution of the clinically-based criteria (2-4) and terminology for patients who do not have one of the known genetic mutations for classical, vascular or the other so-called rare subtypes of Ehlers-Danlos Syndrome (EDS).
Hakim et al assert in their letter that diagnosed Joint Hypermobility Syndrome (JHS) is known to be ‘common’. We have searched the literature for reported whole population prevalence rates for JHS and have been unable to find any. Therefore our paper is the first to report a diagnosed prevalence for this condition, although this figure can also be derived for the population of Sweden, as we pointed out in our work. (5)
We completely agree with the letter’s authors that at this point in time it is not possible to know what proportion of people who met the Brighton Criteria for JHS also meet the 2017 hEDS criteria. We hope further research may reveal this data in the future. What we can say, however, is that for the decade or so prior to 2017, experts in the field considered JHS and EDS-HT to be clinically indistinguishable (1, 6-10), indeed many clinicians used the term JHS/EDS-HT in correspondence in recognition of this. Indeed,...
We would like to thank Hakim et al for the opportunity to continue the debate around the prevalence and definitions of conditions associated with joint hypermobility. Coming to anything like a true population prevalence for these conditions is still a long way off due to both the huge under-diagnosis (1) and ongoing evolution of the clinically-based criteria (2-4) and terminology for patients who do not have one of the known genetic mutations for classical, vascular or the other so-called rare subtypes of Ehlers-Danlos Syndrome (EDS).
Hakim et al assert in their letter that diagnosed Joint Hypermobility Syndrome (JHS) is known to be ‘common’. We have searched the literature for reported whole population prevalence rates for JHS and have been unable to find any. Therefore our paper is the first to report a diagnosed prevalence for this condition, although this figure can also be derived for the population of Sweden, as we pointed out in our work. (5)
We completely agree with the letter’s authors that at this point in time it is not possible to know what proportion of people who met the Brighton Criteria for JHS also meet the 2017 hEDS criteria. We hope further research may reveal this data in the future. What we can say, however, is that for the decade or so prior to 2017, experts in the field considered JHS and EDS-HT to be clinically indistinguishable (1, 6-10), indeed many clinicians used the term JHS/EDS-HT in correspondence in recognition of this. Indeed, in guidance released by The Ehlers-Danlos Society alongside the 2017 reclassification, it is stated that “if someone was diagnosed with hEDS before the 2017 criteria, there’s no cause to seek a new diagnosis unless they decide to participate in new research or need to be reassessed for some other reason”. (11(p.1)) For these reasons, we would maintain that grouping patients with JHS and EDS-HT is legitimate. However, in order to accurately reflect the current classification, notwithstanding the fact that the hEDS criteria are being revised at the present time, we chose to recognise the term Hypermobility Spectrum Disorders (HSD) as a possible diagnosis for an unknown proportion of these patients. The same 2017 guidance states that “treatment is more important than labels” (11( p.1)) With this sentiment we wholeheartedly agree. We cannot know how many of the patients in this historical cohort meet which clinical criteria now, nor indeed at the time they were given these labels. But we can say that there are many more people than previously realised who are suffering from symptomatic generalised joint hypermobility and its associated multisystemic features (8), by whatever definition. It is patients and their symptoms which we believe we should be focusing on until such time as we have a clearer understanding of the presumed underlying genetic factors.
To test Hakim et al’s hypothesis that the characteristics of the JHS and EDS groups could differ, we decided to repeat our case-control calculations, combining all ages but separating the two diagnostic groups, and we present this data in figures 1 and 2. We find it interesting that the two groups show such homogeneity, with no significant differences across the 35 Read chapter categories, other than in additional musculoskeletal diagnoses and musculoskeletal medications for those with JHS within the 12 months before and after diagnosis. We believe this adds to the previously held opinion that JHS and the hypermobile form of EDS as assessed by the previous Villefranche or Berlin criteria are clinically indistinguishable. We did find a difference in the age and gender characteristics of the two populations (see figures 3 and 4), possibly related to patterns of referral depending on age and symptomatology, which could potentially explain the observed difference in musculoskeletal findings. This is speculation, however, and we would encourage exploration of our cohorts by other researchers who may be able to elucidate this further. The speculation by Hakim et al that it is possible that no one with a previous diagnosis of JHS would meet the 2017 hEDS criteria seems unlikely to us, based on our clinical experience and the observations of others since 2017 (12-14) (Eccles, J A, written communication, 6th and 14th November 2019), but others must draw their own conclusions until such time as we have further research data on this question. We thank the respondents for their intriguing calculation that if just 15% of people with a previous diagnosis of JHS now meet the 2017 hEDS criteria, this would be sufficient to find EDS not to be rare. We ourselves made no attempt to model or guess at the proportions, but simply reported the available coded data.
Where Hakim et al refer to the proportion of ‘the EDS cohort’ which have hEDS versus the other subtypes, it is unclear whether they are referring to our data or to people with EDS in general. We were careful to set out in detail where patients had a subtype of EDS recorded and where they did not. Whilst we might all presume that the overwhelming majority of patients with the header code EDS in fact have the hypermobile form, we cannot make this assumption without supporting evidence. Given how few individuals had a coded subtype other than hypermobile, we made the decision to include all people with EDS in our study. We have been very careful to consistently talk about a combined prevalence of EDS and JHS in our paper and stated that other subtypes of EDS are genuinely rare and regret if this has been misquoted on social media.
We completely agree with Hakim at al that ‘further studies are required to determine how common HSD, hEDS, and other forms of EDS are’, however at the present moment, after the 2017 reclassification but whilst we are awaiting a revision and validation of the hEDS criteria, such studies, at least at an epidemiological level, are virtually impossible, especially when one considers the inconsistency between these diagnostic labels and the clinical codes available. No new clinical codes have been introduced post the 2017 classification to distinguish between JHS/EDS-HT and HSD/hEDS or to expand on the sub-coding of EDS. (15) The definitions of ‘Hypermobility Spectrum Disorders’ remain confusing, and to date, there has been no research which has been able to demonstrate a difference between those who meet the hEDS criteria and those who fall into HSD12 (16-18).
We reject the assertion that we combined JHS and EDS populations in order ‘to identify a larger absolute value for prevalence’. At all times we worked from the clinical guidance about these conditions, although obviously constrained by the historical nature of the coded data. We do not seek to remove the protected ‘rare’ status from all subtypes of EDS, but from the data available to us, we made what we felt to be the most accurate calculations we could, and our paper passed through the full rigours of open peer review prior to publication. It should not be forgotten, again, that the vast majority of patients remain without a diagnosis. (1) In terms of NHS England hEDS is already effectively excluded from the provision of specialist commissioning for rare diseases. (19(p.26)) But because it is still believed to be ‘rare’, there is also a lack of appropriate locally commissioned services tailored to the needs of this group. Our work in no way intends to disrupt the pathways or support available for people who require genetic sequencing to confirm subtypes of EDS with known mutations, and nor should it. We do hope that as a result of our work, clinicians, commissioners, educators and others will come to realise the significance, and higher prevalence than previously realised, of hypermobility-related conditions.
Figure 1: see here https://blogs.bmj.com/bmjopen/files/2019/11/figure-1.jpg
Odds ratios of Read chapter diagnoses within 12 months before and after EDS or JHS diagnosis. Presented are all results that affect at least 5 cases or 20 controls. Chapters are as follows: A: Infectious/parasitic diseases; B: Neoplasms; C: Endocrine.butr./metab./immun. Disease; D: Blood/blood forming organs disease; E: Mental disorders; F: Nervous system / sense organ disease; G: Circulatory system disease; H: Respiratory system disease; J: Digestive system disease; K: Genitourinary system disease; L: Pregnancy / childbirth /puerperium; M: Skin/ subcutaneous tissue disease; N: Musculoskeletal / connective tissue; P: Congenital anomalies; R: Symptoms, signs, ill-defined conditions; S: Injury and poisoning; T: Causes of injury and poisoning; U: External causes of morbidity and mortality; Z: Unspecified conditions
Figure 2: see here https://blogs.bmj.com/bmjopen/files/2019/11/figure-2.jpg
Odds ratios of Read chapter prescriptions 12 months before and after EDS or JHS diagnosis. Presented are all results that affect at least 5 cases or 20 controls. Chapters are as follows: a: Gastro-intestinal drugs; b: Cardiovascular drugs; c: Respiratory drugs; d: Central Nervous System drugs; e: Drugs used in Infections; f: Endocrine drugs; g: Obs/Gynae/Uti drugs; h: Chemotherapy/ Immunosupp drugs; i: Heamatology/Dietetic drugs; j: Musculoskeletal drugs; k: Eye drugs; l: ENT drugs; m: Skin drugs; n: Immunology drugs and vaccines; o: Anaesthetic drugs; p: Appliances and reagents (dressing etc.)
1. Grahame R. Hypermobility: an important but often neglected area within rheumatology. Nature Clinical Practice Rheumatology 2008;4(10):522-24. doi: 10.1038/ncprheum0907
2. De Wandele L, Rombaut L, Coussens M, et al. A suggestion for refining the new diagnostic criteria for hypermobile EDS. International Symposium on the Ehlers-Danlos Syndromes. Ghent, Belgium, 2018.
3. Rombaut L, De Wandele L, Coussens M, et al. The 2017 diagnostic criteria for hEDS: the Ghent data and experience, and a suggestion for refining the new diagnostic criteria for hEDS. International Symposium on the Ehlers-Danlos Syndromes. Ghent, Belgium, 2018.
4. Grahame R. Should comorbidities be included in the hEDS criteria? International Symposium on the Ehlers-Danlos Syndromes. Ghent, Belgium, 2018.
5. Cederlöf M, Larsson H, Lichtenstein P, et al. Nationwide population-based cohort study of psychiatric disorders in individuals with Ehlers-Danlos syndrome or hypermobility syndrome and their siblings. Bmc Psychiatry 2016;16 doi: 10.1186/s12888-016-0922-6
6. Gazit Y, Jacob G, Grahame R. Ehlers-Danlos Syndrome-Hypermobility Type: A Much Neglected Multisystemic Disorder. Rambam Maimonides Medical Journal 2016;7(4) doi: 10.5041/rmmj.10261
7. Tinkle B, Castori M, Berglund B, et al. Hypermobile Ehlers-Danlos Syndrome (a.k.a. Ehlers-Danlos Syndrome Type III and Ehlers-Danlos Syndrome Hypermobility Type): Clinical Description and Natural History. American Journal of Medical Genetics Part C-Seminars in Medical Genetics 2017;175(1):48-69. doi: 10.1002/ajmg.c.31538
8. Castori M, Tinkle B, Levy H, et al. A Framework for the Classification of Joint Hypermobility and Related Conditions. American Journal of Medical Genetics Part C-Seminars in Medical Genetics 2017;175(1):148-57. doi: 10.1002/ajmg.c.31539
9. Castori M, Dordoni C, Valiante M, et al. Nosology and Inheritance Pattern(s) of Joint Hypermobility Syndrome and Ehlers-Danlos Syndrome, Hypermobility Type: A Study of Intrafamilial and Interfamilial Variability in 23 Italian Pedigrees. American Journal of Medical Genetics Part A 2014;164(12):3010-20. doi: 10.1002/ajmg.a.36805
10. Tinkle BT, Bird HA, Grahame R, et al. The Lack of Clinical Distinction Between the Hypermobility Type of Ehlers-Danlos Syndrome and the Joint Hypermobility Syndrome (a.k.a. Hypermobility Syndrome). American Journal of Medical Genetics Part A 2009;149A(11):2368-70. doi: 10.1002/ajmg.a.33070
11. The Ehlers-Danlos Society. Hypermobile Ehlers-Danlos syndrome (hEDS) vs. Hypermobility Spectrum Disorders (HSD): What’s the Difference? no date [Available from: https://ehlers-danlos.com/wp-content/uploads/hEDSvHSD.pdf.
12. Huynh J, Saboda K, Sanoja A, et al. Patient-reported symptoms do not differ between people who do and do not meet hEDS criteria. International Symposium on the Ehlers-Danlos Syndromes. Ghent, Belgium, 2018.
13. McGillis L, Clarke H, Hussey L, et al. Experiencing utilizing the 2017 criteria for diagnosis of Ehlers-Danlos syndrome: a Canadian perspective. International Symposium on the Ehlers-Danlos Syndromes. Ghent, Belgium, 2018.
14. Rosen C, O'Connor C, Schwartz S, et al. Application of the 2017 hypermobile EDS diagnostic criteria in a paediatric population. International Symposium on the Ehlers-Danlos Syndromes. Ghent, Belgium, 2018.
15. Facts and Dimensions. New UK Health Dimensions Database For Use by all NHS Organisations no date [Available from: https://factsanddimensions.co.uk/.
16. Chiarelli N, Ritelli M, Zoppi N, et al. Cellular and Molecular Mechanisms in the Pathogenesis of Classical, Vascular, and Hypermobile Ehlers-Danlos Syndromes. Genes 2019;10(8) doi: 10.3390/genes10080609
17. Zoppi N, Chiarelli N, Binetti S, et al. Dermal fibroblast-to-myofibroblast transition sustained by alpha v beta 3 integrin-ILK-Snaill/Slug signaling is a common feature for hypermobile Ehlers-Danlos syndrome and hypermobility spectrum disorders. Biochimica Et Biophysica Acta-Molecular Basis of Disease 2018;1864(4):1010-23. doi: 10.1016/j.bbadis.2018.01.005
18. Copetti M, Morlino S, Colombi M, et al. Severity classes in adults with hypermobile Ehlers-Danlos syndrome/hypermobility spectrum disorders: a pilot study of 105 Italian patients. Rheumatology 2019;58(10):1722-30. doi: 10.1093/rheumatology/kez029
19. NHS England. Highly specialised services 2018 2018 [Available from: https://www.england.nhs.uk/commissioning/wp-content/uploads/sites/12/201....
We are grateful to Levy et al for their comments on our paper.
Much of the response to the paper has been to the media coverage and potential consequences of this. That MDIs have a large carbon footprint is hardly news within technical and academic literature. (1) It’s fair to say we were taken aback by the media response. Any guilt induced by headlines which focussed on individual change is deeply regrettable. Our paper was focussed on modelling at the NHS level. In-line with other previous reports we included an individual level comparison to provide context to our findings. (1-3) The “180-mile car journey” ascribed to us by Levy et al. originates from a story about NICE asthma inhalers decision aid.(4) Clearly we can’t control the media and have tried to correct errors in media reporting where possible. It is our opinion that it would be unethical and paternalistic to withhold significant information about treatment options from patients. As academic authors we are reflecting on how best to communicate information of the environmental impact of healthcare to the public and media, and considering how we might improve this in the future.
Levy et al. highlight concerns “that all patients can be summarily switched from pMDIs to DPIs” or that patients are “deprived of access to pMDI therapy”. We do not propose this in the paper. In fact we make suggestions on how to reduce the greenhouse gas emissions from MDIs where their continuing use is necessary by prio...
We are grateful to Levy et al for their comments on our paper.
Much of the response to the paper has been to the media coverage and potential consequences of this. That MDIs have a large carbon footprint is hardly news within technical and academic literature. (1) It’s fair to say we were taken aback by the media response. Any guilt induced by headlines which focussed on individual change is deeply regrettable. Our paper was focussed on modelling at the NHS level. In-line with other previous reports we included an individual level comparison to provide context to our findings. (1-3) The “180-mile car journey” ascribed to us by Levy et al. originates from a story about NICE asthma inhalers decision aid.(4) Clearly we can’t control the media and have tried to correct errors in media reporting where possible. It is our opinion that it would be unethical and paternalistic to withhold significant information about treatment options from patients. As academic authors we are reflecting on how best to communicate information of the environmental impact of healthcare to the public and media, and considering how we might improve this in the future.
Levy et al. highlight concerns “that all patients can be summarily switched from pMDIs to DPIs” or that patients are “deprived of access to pMDI therapy”. We do not propose this in the paper. In fact we make suggestions on how to reduce the greenhouse gas emissions from MDIs where their continuing use is necessary by prioritising smaller volume HFA134a inhalers. We would join with Levy et al. in opposing universal switches to DPI therapy or switching inhalers without patient consent. We are grateful to Levy et al. for raising the important issue of SABA overuse as well, and agree there are further potential opportunities here to improve asthma control and reduce greenhouse gas release. For patients who need MDIs for rapid relief of symptoms in emergency situations, including a MDI + spacer in a rescue pack has been proposed as an effective strategy.(5)
Whilst our paper focusses on population level effects, Levy et al. are quite right to raise the issue of what happens at an individual level and the potential risks when switching inhalers. Levy et al. provide anecdotal reports of chaos, discontinuation of inhalers and an increase in exacerbations following inhaler switches. In 2016 it is estimated that 6% of asthma patients and 10% of COPD patients had their inhalers switched purely for cost reasons.(6) We are aware of two relevant UK studies examining the impact of inhaler switches on disease control. Thomas et al. researched the impact of switching ICS for asthma patients without a clinical consultation.(7) Most of these switches were from DPI to MDI and none were from MDI to DPI. They found increased SABA use in the switched group compared to the control group, but no increase in asthma consultations, hospitalisations or oral corticosteroid use. Their findings illustrate the importance of inhaler technique training and assessment for any inhaler switch, but the increased SABA use seen may have been due to inappropriately selecting MDIs, which the authors describe as especially difficult to use. The switches were likely made for financial reasons (without considering greenhouse gas impact) and this may help to explain the near universal use of MDI for ICS inhalers in England. A more recent and more comprehensive study by Bloom et al. found that, contrary to Levy et al’s concerns, adherence improved and exacerbations decreased after switching inhalers.(6) Their analysis included MDI to DPI switches and 95% of patients continued with the inhalers they were switched to. Whilst we agree switching inhalers potentially risks disrupting care, it also represents an opportunity to improve adherence and inhaler technique provided it is accompanied by a clinical consultation.
Levy et al. report the large carbon footprint of aspects of emergency medical care. The healthcare sector needs to find ways to address these issues. It is entirely feasible that we could reduce the need for emergency care and improve disease control whilst switching to lower global warming potential inhalers. It is perhaps also worth noting that the carbon footprint of the using a single dose of treatment from an electronic nebuliser it is roughly half that of a dose of a small-volume HFA134 reliever inhaler.(8)
Levy et al. argue that our paper is “heavily biased in favour of DPIs”. They reference one paper comparing Seretide Evohaler and Seretide Accuhaler as evidence that patients with stable asthma do better on MDIs than DPIs.(9) We referenced this paper in our article, along with other evidence favouring the use of DPIs. We would argue that UK prescriptions as a whole are heavily biased towards MDI. We note 97% of prescriptions for SABAs and 94% of ICS in England are MDI.(10) The UK is a clear outlier compared to the rest of Europe in this regard (11) and the fact that UK has one of the worst asthma mortality rates in Europe makes it hard to argue that our high rates of MDI use bring any systematic benefits to patients. We recognise that differences in demographics and disease profiles between different countries, such as between the UK and Sweden, could influence the need for different therapies. The Netherlands as another example has some of the best asthma mortality rates in Europe has approximately 75% DPI usage overall and 50% DPI use for reliever inhalers in 2011. (11) We note also significant geographical variations within the UK in DPI prescription rates (12) and that some CCGs don’t include any DPI options for their recommended inhalers in some steps of asthma treatment.(13) Levy et al. are concerned about depriving patients of MDIs, but the reality on the ground is that patients are effectively being deprived of DPI treatments in the UK. Research consistently shows that a significant proportion of patients are unable to use MDIs correctly.(14) This suggests to us that prescribers follow the default path of prescribing MDIs a lot of the time, but better care could be offered by more carefully selecting the correct inhaler device to match patients’ abilities and preferences. If prescribers did carefully verify the ability of individual patients to use inhalers effectively before prescribing, we anticipate that we would see large increases in DPI prescription rates, improvements in asthma control and reduction in greenhouse gas emissions.
Our analysis did not include the cost of retraining patients to use different types of inhaler. This is not a straightforward calculation and depends which strategy is adopted to reduce greenhouse gas emissions, which we were not prescriptive about. Further research will attempt to address this issue. Changing the types of inhalers started de novo incurs no additional training cost, neither does switching to a near identical inhaler containing less propellant. There could be additional costs involved with retraining some patients although these would be partly offset by the fact that all patients should be having their inhaler technique regularly assessed as part of good routine practice.
Levy et al. point to other potential environmental impacts of inhalers. We focused on the greenhouse gas impact of inhalers, as we believe it represents the most pressing environmental issue. It is also the one issue that differs most dramatically between different types of inhalers.
Levy et al. commend the analysis of Jeswani & Azapagic (3) which we agree has many strengths, particularly its comprehensive life-cycle analysis. Levy et al. argue in favour of their more comprehensive assessment of impacts and whilst we would agree with this, it can be difficult to know the clinical relevance of information such as the greater proportion of potentially carcinogenic compounds in DPI manufacture. The greenhouse gas findings of Jeswani & Azapagic are largely in agreement with our own, in particular the conclusion that the overwhelming majority of the carbon footprint of MDIs derives from the HFA propellant. We have concerns about their method of relying on information from patient information leaflets to estimate the volume of propellants within MDIs. For instance, Jeswani & Azapagic report that a Salamol MDI contains 16.54g of HFA propellant which can’t be correct as the combined weight of the canister and contents of a new Salamol inhaler is only 15.6g. (15)
Levy et al. point to weaknesses in our analysis of the carbon footprint of DPIs and whilst we agree there is limited information in this area, the information that is available is concordant.(3,16) We are reliant on manufacturers for much of this information and would encourage more companies to release life cycle analyses into the public domain.
Levy points out that SMART/MART regimens have potential to improve patient care, reduce SABA use limit greenhouse gas emissions. If as Levy et al. argue this strategy is a safer and more effective strategy then this is to be welcomed. This strategy has been recommended by GINA guidelines, but we would like to see more real-world evidence of effectiveness before whole-heartedly supporting it. We are also less confident that it will be adopted worldwide, particularly in low and middle income countries as combination inhalers may be considered too expensive.
Looking beyond MART, we see many opportunities in the UK to improve care alongside a switch to low global warming potential inhalers. For SABAs there is currently near universal (97%) use of MDIs with opportunities to improve care and reduce MDI use. This could be achieved by using DPIs instead where appropriate, promoting regular preventer therapies such as MART, but also low GWP LAMA/LABA inhalers in COPD. For SAMAs, regular long-acting controller therapies have better evidence for improved disease control, preventing exacerbations and with lower greenhouse gas emissions. As described above ICS inhalers in the UK are almost universally (94%) MDIs. This is not compatible with the idea that we are carefully matching inhalers to patients’ preferences and abilities. None of these MDI SABA, SAMA or ICS inhalers have dose counters, leading to the dangerous and common phenomenon of patients using MDIs that are effectively empty (17). DPI counterpart inhalers all include dose counters. For ICS+LABA inhalers, in addition to MART, there is very widespread overuse of high dose ICS in COPD, with subsequent risks of pneumonia, oral thrush, and probably diabetes and bone fractures.(18) We see huge scope for improvements in patient care alongside financial and greenhouse gas savings by switching to more appropriate inhalers.(19) Reducing greenhouse gas emissions could be a happy coincidence of these improvements, an opportunity seized alongside improving disease control, or a catalyst providing renewed impetus to drive improvements in patient care.
Levy et al. raise concerns about the contribution of lactose monohydrate to the carbon footprint of DPIs. Lactose was included in the carbon footprint analyses of GSK’s DPIs (16) and by Jeswani and Azapagic.(3) Even applying the comparatively high GWP figures quoted by Levy et al. of 13.1kg CO2e/kg of Whey powder, the contribution to the carbon footprint is still very small. Seretide Accuhaler contains 12.5mg lactose per blister, or 0.75g per inhaler.(19) The carbon footprint contribution from lactose is therefore less than 10g CO2e per inhaler (13.1 x 0.75 = 9.83gCO2e per inhaler). This amounts to approximately 1% of the total carbon footprint of the inhaler.
We’d agree with the conclusion of Levy et al. that “asthma and COPD management should focus on prescribing appropriate inhaler devices …. that individual patients are able to use efficiently after appropriate education and in the context of management plans agreed with appropriately trained health care professionals.” As we’ve argued above, if we did this we’d see large switches to low global warming potential inhalers. We would additionally suggest that information on environmental impact should be considered and shared with patients whenever inahlers are being reviewed, or switched for other reasons, and that where all other considerations are equipoised then the treatment with the lowest carbon footprint should be selected.
1.Montreal Protocol On Substances that Deplete the Ozone Layer Report of the UNEP Medical Technical Options Committee 2014 Assessment. Nairobi, Kenya, 2010.
2.NICE Patient decision aid. Inhalers for asthma. London, 2019.
3.Jeswani HK, Azapagic A. Life cycle environmental impacts of inhalers. Journal of Cleaner Production. 2019;237:117733.
4.https://www.telegraph.co.uk/news/2019/04/08/asthma-inhalers-bad-environm...
5.Emergency MDI and spacer packs for asthma and COPD Keeley, Duncan et al.The Lancet Respiratory Medicine, Volume 7, Issue 5, 380 - 382
6.Bloom CI, Douglas I, Olney J, et al. Cost saving of switching to equivalent inhalers and its effect on health outcomes Thorax 2019;74:1078-1086.
7.Thomas, M., Price, D., Chrystyn, H. et al. Inhaled corticosteroids for asthma: impact of practice level device switching on asthma control. BMC Pulm Med 9, 1 (2009) doi:10.1186/1471-2466-9-1
8.Goulet B, Olson L, Mayer B. A Comparative Life Cycle Assessment between a Metered Dose Inhaler and Electric Nebulizer. Sustainability 2017; 9: 1725.
9.Price D , Roche N , Christian Virchow J , et al . Device type and real-world effectiveness of asthma combination therapy: an observational study. Respir Med 2011;105:1457–66.doi:10.1016/j.rmed.2011.04.010
10.NHS England. NHS digital. 2017. https://digital.nhs.uk/prescribing (accessed July 17, 2018).
11.Lavorini F, Corrigan CJ, Barnes PJ, et al. Retail sales of inhalation devices in European countries: So much for a global policy. Respir Med 2016; 105: 1099–103.
12.https://openprescribing.net/measure/environmental_inhalers/
13.http://www.gpref.bedfordshire.nhs.uk/media/209947/BCCG%20Asthma%20Guidel...
14.Sanchis J, Gich I, Pedersen S. Systematic review of errors in inhaler use: Has patient technique improved over time? Chest 2016; 150: 394–406.
15.Sellers WFS. Asthma pressurised metered dose inhaler performance: propellant effect studies in delivery systems. Allergy Asthma Clin Immunol 2017; 13: 30.
16.Janson C, Henderson R, Löfdahl M, et al Carbon footprint impact of the choice of inhalers for asthma and COPD Thorax Published Online First: 07 November 2019. doi: 10.1136/thoraxjnl-2019-213744
17.Conner JB, Buck PO. Improving asthma management: the case for mandatory inclusion of dose counters on all rescue bronchodilators. J Asthma. 2013;50(6):658–663. doi:10.3109/02770903.2013.789056
18.Cataldo D, Derom E, Liistro G, et al. Overuse of inhaled corticosteroids in COPD: five questions for withdrawal in daily practice. Int J Chron Obstruct Pulmon Dis. 2018;13:2089–2099. Published 2018 Jul 5. doi:10.2147/COPD.S164259
19.d’Ancona G, Patel I, Saleem A, et al P29 Impact Of Respiratory Virtual Clinics In Primary Care On Responsible Respiratory Prescribing And Inhaled Corticosteroid Withdrawal In Patients With Copd: A Feasibility Study Thorax 2014;69:A90.
20.https://www.medicines.org.uk/emc/files/pil.5504.pdf
Many thanks for bringing this to our attention. Unfortunately that report has been removed from their website and we are unable to add it now as a supplementary file. Here is an alternative reference published more recently which contains the same information:
Janson C, Henderson R, Löfdahl M, et al. Carbon footprint impact of the choice of inhalers for asthma and COPD Thorax doi: 10.1136/thoraxjnl-2019-213744
The findings presented here on the carbon footprint of inhalers are entirely concordant with other previous reports.
The paper says "...specialty and associate specialist (SAS) doctors (doctors who have completed specialist training but do not have a staff position)...".
That's not correct. Firstly while SAS doctors may have completed specialist training, they more often have not. Secondly, a permenent SAS doctor does have a "staff position" in ordinary language. If this is a reference to US-specific terminology then I suspect the problem is that there is no direct equivalent.
Dear Professor Cullum
Thank you for your comments regarding the of Cullum et al (2016) and Madden & Morley (2016). These two publications were identified through our search strategy, however, they did not meet our inclusion criteria. The a priory criteria set for inclusion and exclusion of primary studies are listed below:
Inclusion criteria: All steps from James Lind Alliance, list of Top 10 priorities, adults (aged > 18 years or older)
Exclusion criteria: Unpublished literature, articles not written in English, priority setting partnership without James Lind Alliance, James Lind Alliance without priority setting partnership, protocols, errata, editorial, thesis, comments, review, guidelines, randomized controlled trials.
1] Cullum N, Buckley H, Dumville J, Hall J, Lamb K, Madden M, Morley R, O'Meara S, Goncalves PS, Soares M, Stubbs N. Wounds research for patient benefit: a 5 year programme of research. NIHR Journals Library; 2016. This report does not describe prioritized the Top 10 list.
2] Madden M, Morley R. (2016). Exploring the challenge of health research priority setting in partnership: reflections on the methodology used by the James Lind Alliance Pressure Ulcer Priority Setting Partnership. This article does not describe prioritized the Top 10 list.
In England I have used the output area classification with success for over 10 years to identify social groups with higher rates of admission. I have found deprivation to be a very crude measure. Alas much of this is not published. but several studies regarding admission to the CCU. Also a couple of studies looking at outbreaks of a mystery disease. See below.
Hope this helps.
Beeknoo N, Jones R. Using Social Groups to Locate Areas with High Emergency Department Attendance, Subsequent Inpatient Admission and Need for Critical Care. British Journal of Medicine and Medical Research 2016; 18(6): 1-23. doi: 10.9734/BJMMR/2016/29208
Beeknoo N, Jones R. Using social groups to locate areas of high utilization of critical care. Brit J Healthcare Manage 2016; 22(11): 551-560.
Jones R. Role of social group and gender in outbreaks of a novel agent leading to increased deaths, with insights into higher international deaths in 2015. Fractal Geometry and Nonlinear Analysis in Medicine and Biology 2017; 3(1): 1-7. doi: 10.15761/FGNAMB.1000146
In England I have used the output area classification with success for over 10 years to identify social groups with higher rates of admission. I have found deprivation to be a very crude measure. Alas much of this is not published. but several studies regarding admission to the CCU. Also a couple of studies looking at outbreaks of a mystery disease. See below.
Hope this helps.
Beeknoo N, Jones R. Using Social Groups to Locate Areas with High Emergency Department Attendance, Subsequent Inpatient Admission and Need for Critical Care. British Journal of Medicine and Medical Research 2016; 18(6): 1-23. doi: 10.9734/BJMMR/2016/29208
Beeknoo N, Jones R. Using social groups to locate areas of high utilization of critical care. Brit J Healthcare Manage 2016; 22(11): 551-560.
Jones R. Role of social group and gender in outbreaks of a novel agent leading to increased deaths, with insights into higher international deaths in 2015. Fractal Geometry and Nonlinear Analysis in Medicine and Biology 2017; 3(1): 1-7. doi: 10.15761/FGNAMB.1000146
Jones R. Different patterns of male and female deaths in 2015 in English and Welsh local authorities question the role of austerity as the primary force behind higher deaths. Fractal Geometry and Nonlinear Analysis in Medicine and Biology 2017; 3(1): 1-4. doi: 10.15761/FGNAMB.1000145
Saez, Vidiella-Martin, and López Casasnovas1 investigate the impact of the Great Recession on self-assessed health in Spain analyzing data from four waves—2005, 2008, 2011 and 2014—of a Survey of Household Finances by the Bank of Spain. The surveys included repeated observations of self-assessed health and other variables measured in the same individuals.
The statistical model of Saez et al. is a mixed logistic regression in which the log of the odds of poor health (log [P / (1 - P)], where P is the probability of declaring poor health) is computed as a linear combination of a random effect for the year of the survey, random effects for individuals and families, and a large set of control variables at both family level (gross wealth, total debt, family income, savings rate, family size, number of family members who work, and type of family residence—owned or rented) and individual level (sex, age—stratified in six groups—, educational level, occupation, and marital status). Saez et al. computed the model with and without adjustment for the control variables and both for the whole sample and for 12 subsamples stratified by sex and age. The observations were weighted, to compensate for the fact that the original survey oversampled the wealthiest households, and the sample was trimmed to eliminate outliers.
Saez et al. found a downward change in self-perceived health during the third wave of the survey, i.e., that of 2011, which coincides with the most severe per...
Saez, Vidiella-Martin, and López Casasnovas1 investigate the impact of the Great Recession on self-assessed health in Spain analyzing data from four waves—2005, 2008, 2011 and 2014—of a Survey of Household Finances by the Bank of Spain. The surveys included repeated observations of self-assessed health and other variables measured in the same individuals.
The statistical model of Saez et al. is a mixed logistic regression in which the log of the odds of poor health (log [P / (1 - P)], where P is the probability of declaring poor health) is computed as a linear combination of a random effect for the year of the survey, random effects for individuals and families, and a large set of control variables at both family level (gross wealth, total debt, family income, savings rate, family size, number of family members who work, and type of family residence—owned or rented) and individual level (sex, age—stratified in six groups—, educational level, occupation, and marital status). Saez et al. computed the model with and without adjustment for the control variables and both for the whole sample and for 12 subsamples stratified by sex and age. The observations were weighted, to compensate for the fact that the original survey oversampled the wealthiest households, and the sample was trimmed to eliminate outliers.
Saez et al. found a downward change in self-perceived health during the third wave of the survey, i.e., that of 2011, which coincides with the most severe period of the Great Recession in Spain (2009–2011, according to Saez et al.). However, they also found significant differences when they stratified the analysis by gender and age groups. According to Figure 1 of the paper, the unadjusted odds of declaring poor health dropped in 2008 to 0.8 and rose in 2011 to about 1.2, but the adjusted odds changed much less. This is interpreted by Saez et al. as an indication that the variation in health status can be captured by either demographic or socioeconomic controls. Figure 2 of the paper reveals extreme heterogeneity in the evolution of the unadjusted odds of declaring poor health in the year 2011. For example, the odds of reporting poor health in 2011, that is, at the worst time of the economic crisis, are about 5 for females aged 35-44 while just slightly above 1.0 for males in that age group. However, when the model includes the control variables the adjusted odds for the year 2011 dramatically drop down to about 1.0 for females and 1.2 for males, as shown in Figure 3. Comparing Figures 2 and 3 of the paper reveals major differences in the profiles of adjusted and unadjusted odds for specific groups of individuals.
Overall, it is hard to make sense of the large heterogeneities found by Saez et al. in this investigation. In the long discussion section of the paper the authors comment on characteristics of the Spanish social security system, on the impact of the Great Recession on income and wealth of pensioners, and other issues, but no clear rationale to explain the extreme heterogeneity of results is presented. It is known that self-assessed health is to some extent a measure of psychological wellbeing and thus it is not surprising that it deteriorated during the Great Recession, as indicators of mental health have been repeatedly found deteriorating during recessions. However, that does not explain the extreme heterogeneity of the results found by Saez et al. To this respect, it may be perhaps important to consider sample size. The investigated sample included 28,678 individuals which Saez et al. claim to be representative of a population of some 20 million Spanish individuals. Now, a general sample of some 29,000 individuals, when stratified in 12 subsamples might yield some very small subsamples in which almost any result is possible. Unfortunately, the sizes of the stratified subsamples are not reported in the paper. At any rate, this investigation tends to reinforce the view that self-assessed health is an indicator quite hard to interpret. For purposes of assessing population health, it would not resist a comparison for instance with mortality indicators.
Saez et al. assert that “it is widely acknowledged that the Great Recession (…) had an adverse impact on health.” This is a very arguable statement. Evidence has piled up that during recessions mortality tends to decline faster than during economic expansions and during the Great Recession in particular, mortality accelerated its decline with respect to the previous years of expansion, both in Europe in general2,3 and in Spain in particular.4,5 As shown in Table 1 here, life expectancy at birth for the Spanish population either grew or kept steady during 2009-2011, that is, the years which Saez et al. consider to be the worst of the Great Recession; that indicates that mortality continued declining during those years. Contrarily, both before the crisis, in the expansion years of 2003, 2005 and 2007, and in 2015, when the recession was finished and the Spanish GDP grew at 3.4%, life expectancy at birth declined indicating increasing mortality. There is now plenty of evidence provided by population statistics rather than survey samples, that health tends to deteriorate in economic expansions and improve in recessions.2-7 However, based perhaps on the flimsy evidence provided by Harvey Brenner in the 1980s, preconceptions about a presumable harmful effect of recessions on health are common. Probably supported by peer reviewers with these preconceptions, a paper was published in 2018 “demonstrating” that in terms of mortality, the Great Recession in Spain had been even more damaging that the Spanish civil war, as it had caused more than a half a million deaths.8 This report has been recently retracted,9 as it was fully based in wrong data.10 The pooled evidence clearly suggests that in the experience of the present and the past century, recessions, on average, have beneficial short‐term effects on mortality, and therefore, on health, of the general population.2-7, 10 There is little evidence either that they might increase health inequalities.4,11
REFERENCES
1. Saez M, Vidiella-Martin J, Casasnovas GL. Impact of the great recession on self-perceived health in Spain: a longitudinal study with individual data. BMJ Open 2019;9:e023258. doi: 10.1136/bmjopen-2018-023258
2. Toffolutti V, Suhrcke M. Assessing the short term health impact of the Great Recession in the European Union: A cross‐country panel analysis. Preventive Medicine 2014, 64(7):54– 62.
3. Tapia Granados JA , Ionides EL. Population health and the economy: mortality and the great recession in Europe. Health Econ 2017;26:e219235.doi:10.1002/hec.3495.
4. Regidor E, Vallejo F, Tapia Granados JA, Viciana-Fernández FJ, de la Fuente L, Barrio G. Faster mortality decline in low socioeconomic groups during the economic crisis in Spain: a cohort study of 36 million people. Lancet 388: 2642–52.
5. Regidor E, Ronda E, Tapia Granados JA, Pulido J, de la Fuente L, Barrio G. Reversal of upward trends in mortality during the Great Recession in employed and unemployed individuals at baseline in a national longitudinal study. Am J Epidemiol 2019; 188(11):2004–2012, doi.org/10.1093/aje/kwz150.
6. Tapia Granados JA, Ionides EL. Statistical evidence shows that mortality tends to fall during recessions: a rebuttal to Catalano and Bruckner. Int J Epidemiol 2016, 45(5):1683–1686, doi.org/10.1093/ije/dyw206.
7. Tapia Granados JA, House JS, Ionides EL, Burgard S, Schoeni RS. Individual joblessness, contextual unemployment, and mortality risk. Am J Epidemiol 2014; 180(3):280-7, doi: 10.1093/aje/kwu128.
8. Cabrera de León A, Rodríguez IM et al. Austerity policies and mortality in Spain after the financial crisis of 2008. Am J Public Health 2018;108(8):1091-1098, doi: 10.2105/AJPH.2018.304346.
9. Morabia A. Notice of Retraction of “Austerity Policies and Mortality in Spain After the Financial Crisis of 2008”. Am J Public Health 2019, 109(7):e17. doi:0.2105/AJPH.2019.305147.
10. Regidor E, Mateo A, Barrio G, de la Fuente L. Mortality in Spain in the Context of the Economic Crisis and Austerity Policies. Am J Public Health Published Online: June 05, 2019, ajph.aphapublications.org/doi/10.2105/AJPH.2019.305075
11. Bartoll X, Palència L, Malmusi D, Suhrcke D, Borrell C. The evolution of mental health in Spain during the economic crisis. Eur J Public Health 2014, 24(3):415–418.
Table 1. Life expectancy at birth (years) for males and females and the whole population in Spain, 2000-2015
Annual change
Year All Females Males All Females Males
2000 79.5 83.0 76.0 0.7 0.7 0.7
2001 79.8 83.3 76.3 0.3 0.3 0.3
2002 79.9 83.4 76.4 0.1 0.0 0.1
2003 79.8 83.1 76.4 -0.1 -0.2 0.0
2004 80.5 83.9 77.0 0.7 0.7 0.6
2005 80.4 83.8 77.1 0.0 -0.1 0.0
2006 81.2 84.5 77.9 0.8 0.7 0.8
2007 81.2 84.5 77.9 0.0 -0.1 0.0
2008 81.5 84.7 78.3 0.3 0.2 0.4
2009 81.9 85.0 78.8 0.4 0.3 0.4
2010 82.3 85.4 79.2 0.4 0.4 0.4
2011 82.4 85.5 79.4 0.1 0.0 0.2
2012 82.6 85.5 79.5 0.1 0.1 0.2
2013 83.2 86.2 80.2 0.7 0.6 0.7
2014 83.3 86.2 80.4 0.1 0.1 0.2
2015 83.0 85.8 80.1 -0.4 -0.5 -0.3
Data source: HFA database of the European Regional Office of the World Health Organization. Shaded cells correspond to the years considered by Saez et al. as the worst of the economic crisis.
We thank Dr Pollock and colleagues for their interest in our recent study, which suggested that among patients with non-valvular atrial fibrillation in the UK, inappropriate underdosing was more than twice as common among patients starting on apixaban than those starting on dabigatran or rivaroxaban.
In our analyses, we assumed that patients with missing data on renal function were likely to have unimpaired renal function. Pollock et al expressed their concern regarding the possibility of significant bias resulting from misclassification of renal function among these patients, which could have resulted in the percentage of patients inappropriately prescribed a reduced dose NOAC being overestimated. Among our study population of 30,467 patients, 3856 (12.7%) had missing data on renal function (eGFR values).
Pollock and colleagues also queried the absence of bodyweight data in our results. We acknowledge that it may have been useful for the reader to see these data, although we presented data on BMI in Table 1 of our article as a proxy measure. Nevertheless, we can confirm that in our dataset very few patients had missing data on bodyweight (2.6% of patients starting on apixaban, 3.0% of those starting on dabigatran and 2.3% of those starting on rivaroxaban). Among patients with a recorded bodyweight, the mean bodyweight was 81.4 kg for patients starting on apixaban, 82.6 kg for those starting on dabigatran, and 82.0 kg for those starting on rivaroxaban. While...
We thank Dr Pollock and colleagues for their interest in our recent study, which suggested that among patients with non-valvular atrial fibrillation in the UK, inappropriate underdosing was more than twice as common among patients starting on apixaban than those starting on dabigatran or rivaroxaban.
In our analyses, we assumed that patients with missing data on renal function were likely to have unimpaired renal function. Pollock et al expressed their concern regarding the possibility of significant bias resulting from misclassification of renal function among these patients, which could have resulted in the percentage of patients inappropriately prescribed a reduced dose NOAC being overestimated. Among our study population of 30,467 patients, 3856 (12.7%) had missing data on renal function (eGFR values).
Pollock and colleagues also queried the absence of bodyweight data in our results. We acknowledge that it may have been useful for the reader to see these data, although we presented data on BMI in Table 1 of our article as a proxy measure. Nevertheless, we can confirm that in our dataset very few patients had missing data on bodyweight (2.6% of patients starting on apixaban, 3.0% of those starting on dabigatran and 2.3% of those starting on rivaroxaban). Among patients with a recorded bodyweight, the mean bodyweight was 81.4 kg for patients starting on apixaban, 82.6 kg for those starting on dabigatran, and 82.0 kg for those starting on rivaroxaban. While it may be true that, among the general population as a whole, weight may not be recorded in many patient records in primary care, this is unlikely to be the case for specific patient populations, such as patients with NVAF, who will be having more regular contact with their primary care practitioner.
In response to these concerns, and to validate our assumption that patients with missing data on renal function were likely to have unimpaired renal function, we have performed a sensitivity analysis excluding patients with missing data on renal function and/or bodyweight (4418 patients; 14.5% of our study cohort). The results of the sensitivity analysis are shown in the Table and Figures below. Only very minimal differences were seen for apixaban (21.6% were inappropriately underdosed in the main analysis vs. 21.4% in the sensitivity analysis), for dabigatran (8.3% in the main analysis vs. 8.2% in the sensitivity analysis), and for rivaroxaban (9.1% in the main analysis vs. 8.4% in the sensitivity analysis)(Table and Figure 1). When dosing was evaluated according to degree of renal function (Figure 2), the data still showed that reduced doses were prescribed to patients with no evidence of renal impairment, with minimal or no difference seen between results of the main analysis and the sensitivity analysis: 26.7% vs. 26.6% for apixaban, 51.1% vs. 50.5% for dabigatran, and 10.3% in both analyses for rivaroxaban. Furthermore, in another ongoing study conducted by our team we have found that when we looked at eGFR values recorded further back in patients’ medical records (further than within the 1 year before the index date), among patients with no recorded value within the year before the index date, about 10% of these had an eGFR value indicating suboptimal renal function. This would translate into about 1% of patients in our current study being misclassified as having normal renal function based on this older eGFR reading. Therefore, we do not consider potential misclassification of renal function, or the inclusion of patients with missing data on bodyweight, to be significant sources of bias, and further analysis is highly unlikely to alter the conclusions of the study in this respect.
In response to the final point raised by Pollock et al, we would like to clarify that the aim of our study was not to ascertain whether a reduced dose such as 2.5 mg/day apixaban was the correct dose for an individual patient, but whether patients were prescribed an appropriate dose as specified on the EU drug label, based on the data recorded in THIN and CPRD databases. We provided a clear definition of appropriate and inappropriate dosing with respect to the EU label in our article.
Considering this further evidence from our study – that any effect of bias due to misclassification of renal function or bodyweight is likely to be very minimal – we are of the opinion that the inferences made in our article remain valid and justifiable.
Thank you for this critical note about vitamin D and B12 testing. The aim of our study was to explore the barriers and facilitators for reducing the number of unnecessary ordered vitamin D and B12 laboratory tests. We found that GPs experienced difficulty to request laborotory tests only for evidenced based indications; often vitamin testing was performed to satisfy patients' requests. We acknowledge the presence of certain medical indications to test vitamin D or B12 bloodlevels and we also performed a training for participating GPs of our study on vitamin D and B12 deficiency and people at risk of such deficiency. The purpose of our study was not to reduce the number of vitamin D and B12 tests to zero, but to explore the barriers and facilitators related to vitamin D and B12 testing in order to improve properly indicated vitamin testing in general practice.
Re. Diagnosed prevalence of Ehlers-Danlos syndrome and hypermobility spectrum disorder in Wales, UK: a national electronic cohort study and case-control comparison.
Demmler J C, Atkinson M D, Reinhold E, Choy E, Lyons R A, Brophy S T
BMJ Open 2019;9:e031365
We write concerning the paper by Demmler et al., published in BMJ Open. We wish to raise the following concerns:
1. With regard to combining the Joint Hypermobility Syndrome (JHS) and Ehlers-Danlos syndromes (EDS) populations for analysis.
If one combines data from a cohort that is found to be ‘common’ (in this case ‘diagnosed JHS’) with one that is found to be ‘rare’ (in this case ‘diagnosed EDS’), the new combined cohort (i.e. diagnosed JHS/EDS) will be common. To then consider the rare cohort common is a fallacy.
Also, although individuals in a population with a previous diagnosis of JHS (i.e. prior to the 2017 international classification (1,2)) might have Hypermobile EDS (hEDS) by the current classification, it is not known how JHS segregates into Hypermobility Spectrum Disorder (HSD) and hEDS. A JHS population would need to be reassessed to confirm this, or modelling assumptions of the data would need to be applied.
In addition, it is not known what proportion of the EDS cohort have hEDS versus the rare Mendelian types of EDS. As such, there is no way of knowing whether or by what proportion the two cohorts represent the same or similar or dif...
Re. Diagnosed prevalence of Ehlers-Danlos syndrome and hypermobility spectrum disorder in Wales, UK: a national electronic cohort study and case-control comparison.
Demmler J C, Atkinson M D, Reinhold E, Choy E, Lyons R A, Brophy S T
BMJ Open 2019;9:e031365
We write concerning the paper by Demmler et al., published in BMJ Open. We wish to raise the following concerns:
1. With regard to combining the Joint Hypermobility Syndrome (JHS) and Ehlers-Danlos syndromes (EDS) populations for analysis.
If one combines data from a cohort that is found to be ‘common’ (in this case ‘diagnosed JHS’) with one that is found to be ‘rare’ (in this case ‘diagnosed EDS’), the new combined cohort (i.e. diagnosed JHS/EDS) will be common. To then consider the rare cohort common is a fallacy.
Also, although individuals in a population with a previous diagnosis of JHS (i.e. prior to the 2017 international classification (1,2)) might have Hypermobile EDS (hEDS) by the current classification, it is not known how JHS segregates into Hypermobility Spectrum Disorder (HSD) and hEDS. A JHS population would need to be reassessed to confirm this, or modelling assumptions of the data would need to be applied.
In addition, it is not known what proportion of the EDS cohort have hEDS versus the rare Mendelian types of EDS. As such, there is no way of knowing whether or by what proportion the two cohorts represent the same or similar or different conditions.
Given the distribution of JHS to EDS in this study is approximately 6 to 1, the combined cohort would likely share the characteristics of the JHS cohort in a case-control analysis.
By combining the two populations to explore association with other aspects of health and wellbeing the opportunity to compare differences between the two groups is lost. Do the populations in this study differ? What evidence is there that JHS and EDS are behaving statistically in the same way or differently with regard to the parameters collected?
We surmise that the authors in part combined the two populations to identify a larger absolute value for prevalence that informs the need for education, training and healthcare resources. It is increasingly recognized that complex combinations of health concerns arise in HSD and hEDS alike. We recognise why HSD and hEDS would be considered together in this context.
However, by their nature the rare types of EDS will have different specific concerns that influence different education and training needs, and healthcare resource allocation. Rare disease status has significant impact on the way healthcare services plan care (e.g. direct commissioning of specialised services in the UK); or the way organisations support research ( e.g. The Office of Rare Diseases, The National Center for Advancing Translational Sciences, USA), for example.
Regardless of combining the prevalence, it remains necessary that we understand the size effect for each of the different patient populations, as specific needs may differ.
2. With regard to interchanging the term JHS with HSD and hEDS, and determining the prevalence of EDS:
The authors switch from the term JHS to a combination of HSD and hEDS. We recognize this inference. Nevertheless, at this time there is no evidence to inform how JHS segregates into HSD or hEDS, and this uncertainty was not clarified in the paper. There is no modelling of the data that might support inferences with regard to such segregation.
What proportion of the JHS group might add to the overall prevalence of hEDS? How does that modelling affect the prevalence of EDS over all when added to the EDS cohort data?
For example, by our estimate approximately 15% of the JHS cohort would need to have hEDS for this prevalence figure plus that from the EDS cohort to reach a population size of 50 per 100000 for EDS as a whole in this study. Even then that would be 1 in 2000 of the general population; and given that some of the EDS cohort would not have hEDS this would also infer that hEDS per se is rare as it would be a proportion of that EDS population.
If none of the JHS population had hEDS the prevalence calculation for EDS using only the EDS cohort data demonstrates EDS to be rare.
There are other ways to model the data, but we hope that this demonstrates our point.
Until reasonable evidence is available as to the segregation of JHS into hEDS and HSD, we believe studies such as that of Demmler et al. should present the data as found for JHS, or define the assumptions and models they are using to reach their conclusions. In addition, studies should avoid clustering JHS and EDS. The term EDS is a pleural (i.e. Ehlers-Danlos syndromes), grouping together disorders under the title.
Further studies are required to determine how common HSD, hEDS, and other forms of EDS are. Shared knowledge among members of the international medical and scientific community suggest that HSD is likely common; that more work is required to determine if hEDS is common or not; and, that other Mendelian types of EDS are rare or ultra-rare.
3. Concerning the combining the populations and the paper's Conclusion.
The authors conclude that “EDS/HSD are not rare conditions…”. This is asserted because the authors have combined the JHS and EDS populations. We question the reasoning for combining the populations above, but wish, here, to raise further concern about interpretation.
We are aware that this publication is being cited in public forums as saying ‘EDS is common’. At this time, from the data presented, there is no evidence to support that statement, for either hEDS or the rarer types of EDS. Even were it found that hEDS is not rare, it is that observation that would make most sense to assert given many forms of EDS are between rare and ultra-rare.
While there is no direct statement from the authors that ‘EDS is common’, we suggest that the opinion has arisen because of the way the whole paper, including the conclusion has been presented.
We respectfully request that the editors of BMJ Open and the paper’s authors consider the above.
Yours faithfully,
Dr. Alan J Hakim, Adjunct Associate Professor of Medicine, College of Medicine, PennState University, USA, & Consultant Rheumatologist, London
Dr Clair A. Francomano, Professor of Medical and Molecular Genetics, Indiana University School of Medicine, Indianapolis, USA
Dr. Marco Castori, Consultant and Director, Division of Medical Genetics, IRCCS Casa Sollievo della Sofferenza, Foggia, Italy
Prof. Fransiska Malfait, Associate Professor, Center for Medical Genetics, Ghent University Hospital, Ghent, Belgium.
References:
Malfait F, Francomano C, Byers P, et al. 2017. The 2017 international classification of the Ehlers–Danlos syndromes. Am J Med Genet Part C Semin Med Genet 175C:8–26.
Castori M, Tinkle B, Levy H, Grahame R, Malfait F, Hakim A. 2017. A framework for the classification of joint hypermobility and related conditions. Am J Med Genet C Semin Med Genet 175C:148-157.
We would like to thank Hakim et al for the opportunity to continue the debate around the prevalence and definitions of conditions associated with joint hypermobility. Coming to anything like a true population prevalence for these conditions is still a long way off due to both the huge under-diagnosis (1) and ongoing evolution of the clinically-based criteria (2-4) and terminology for patients who do not have one of the known genetic mutations for classical, vascular or the other so-called rare subtypes of Ehlers-Danlos Syndrome (EDS).
Hakim et al assert in their letter that diagnosed Joint Hypermobility Syndrome (JHS) is known to be ‘common’. We have searched the literature for reported whole population prevalence rates for JHS and have been unable to find any. Therefore our paper is the first to report a diagnosed prevalence for this condition, although this figure can also be derived for the population of Sweden, as we pointed out in our work. (5)
We completely agree with the letter’s authors that at this point in time it is not possible to know what proportion of people who met the Brighton Criteria for JHS also meet the 2017 hEDS criteria. We hope further research may reveal this data in the future. What we can say, however, is that for the decade or so prior to 2017, experts in the field considered JHS and EDS-HT to be clinically indistinguishable (1, 6-10), indeed many clinicians used the term JHS/EDS-HT in correspondence in recognition of this. Indeed,...
Show MoreWe are grateful to Levy et al for their comments on our paper.
Much of the response to the paper has been to the media coverage and potential consequences of this. That MDIs have a large carbon footprint is hardly news within technical and academic literature. (1) It’s fair to say we were taken aback by the media response. Any guilt induced by headlines which focussed on individual change is deeply regrettable. Our paper was focussed on modelling at the NHS level. In-line with other previous reports we included an individual level comparison to provide context to our findings. (1-3) The “180-mile car journey” ascribed to us by Levy et al. originates from a story about NICE asthma inhalers decision aid.(4) Clearly we can’t control the media and have tried to correct errors in media reporting where possible. It is our opinion that it would be unethical and paternalistic to withhold significant information about treatment options from patients. As academic authors we are reflecting on how best to communicate information of the environmental impact of healthcare to the public and media, and considering how we might improve this in the future.
Levy et al. highlight concerns “that all patients can be summarily switched from pMDIs to DPIs” or that patients are “deprived of access to pMDI therapy”. We do not propose this in the paper. In fact we make suggestions on how to reduce the greenhouse gas emissions from MDIs where their continuing use is necessary by prio...
Show MoreMany thanks for bringing this to our attention. Unfortunately that report has been removed from their website and we are unable to add it now as a supplementary file. Here is an alternative reference published more recently which contains the same information:
Janson C, Henderson R, Löfdahl M, et al. Carbon footprint impact of the choice of inhalers for asthma and COPD Thorax doi: 10.1136/thoraxjnl-2019-213744
The findings presented here on the carbon footprint of inhalers are entirely concordant with other previous reports.
The paper says "...specialty and associate specialist (SAS) doctors (doctors who have completed specialist training but do not have a staff position)...".
That's not correct. Firstly while SAS doctors may have completed specialist training, they more often have not. Secondly, a permenent SAS doctor does have a "staff position" in ordinary language. If this is a reference to US-specific terminology then I suspect the problem is that there is no direct equivalent.
Dear Professor Cullum
Thank you for your comments regarding the of Cullum et al (2016) and Madden & Morley (2016). These two publications were identified through our search strategy, however, they did not meet our inclusion criteria. The a priory criteria set for inclusion and exclusion of primary studies are listed below:
Inclusion criteria: All steps from James Lind Alliance, list of Top 10 priorities, adults (aged > 18 years or older)
Exclusion criteria: Unpublished literature, articles not written in English, priority setting partnership without James Lind Alliance, James Lind Alliance without priority setting partnership, protocols, errata, editorial, thesis, comments, review, guidelines, randomized controlled trials.
1] Cullum N, Buckley H, Dumville J, Hall J, Lamb K, Madden M, Morley R, O'Meara S, Goncalves PS, Soares M, Stubbs N. Wounds research for patient benefit: a 5 year programme of research. NIHR Journals Library; 2016. This report does not describe prioritized the Top 10 list.
2] Madden M, Morley R. (2016). Exploring the challenge of health research priority setting in partnership: reflections on the methodology used by the James Lind Alliance Pressure Ulcer Priority Setting Partnership. This article does not describe prioritized the Top 10 list.
In England I have used the output area classification with success for over 10 years to identify social groups with higher rates of admission. I have found deprivation to be a very crude measure. Alas much of this is not published. but several studies regarding admission to the CCU. Also a couple of studies looking at outbreaks of a mystery disease. See below.
Hope this helps.
Beeknoo N, Jones R. Using Social Groups to Locate Areas with High Emergency Department Attendance, Subsequent Inpatient Admission and Need for Critical Care. British Journal of Medicine and Medical Research 2016; 18(6): 1-23. doi: 10.9734/BJMMR/2016/29208
Beeknoo N, Jones R. Using social groups to locate areas of high utilization of critical care. Brit J Healthcare Manage 2016; 22(11): 551-560.
Jones R. Year-to-year variation in deaths in English Output Areas (OA), and the interaction between a presumed infectious agent and influenza in 2015. SMU Medical Journal 2017; 4(2): 37-69.
http://smu.edu.in/content/dam/manipal/smu/smims/Volume4No2July2017/SMU%2...(July%202017)%20-%204.pdf
Jones R. Role of social group and gender in outbreaks of a novel agent leading to increased deaths, with insights into higher international deaths in 2015. Fractal Geometry and Nonlinear Analysis in Medicine and Biology 2017; 3(1): 1-7. doi: 10.15761/FGNAMB.1000146
Jones R. Differ...
Show MoreSaez, Vidiella-Martin, and López Casasnovas1 investigate the impact of the Great Recession on self-assessed health in Spain analyzing data from four waves—2005, 2008, 2011 and 2014—of a Survey of Household Finances by the Bank of Spain. The surveys included repeated observations of self-assessed health and other variables measured in the same individuals.
Show MoreThe statistical model of Saez et al. is a mixed logistic regression in which the log of the odds of poor health (log [P / (1 - P)], where P is the probability of declaring poor health) is computed as a linear combination of a random effect for the year of the survey, random effects for individuals and families, and a large set of control variables at both family level (gross wealth, total debt, family income, savings rate, family size, number of family members who work, and type of family residence—owned or rented) and individual level (sex, age—stratified in six groups—, educational level, occupation, and marital status). Saez et al. computed the model with and without adjustment for the control variables and both for the whole sample and for 12 subsamples stratified by sex and age. The observations were weighted, to compensate for the fact that the original survey oversampled the wealthiest households, and the sample was trimmed to eliminate outliers.
Saez et al. found a downward change in self-perceived health during the third wave of the survey, i.e., that of 2011, which coincides with the most severe per...
We thank Dr Pollock and colleagues for their interest in our recent study, which suggested that among patients with non-valvular atrial fibrillation in the UK, inappropriate underdosing was more than twice as common among patients starting on apixaban than those starting on dabigatran or rivaroxaban.
In our analyses, we assumed that patients with missing data on renal function were likely to have unimpaired renal function. Pollock et al expressed their concern regarding the possibility of significant bias resulting from misclassification of renal function among these patients, which could have resulted in the percentage of patients inappropriately prescribed a reduced dose NOAC being overestimated. Among our study population of 30,467 patients, 3856 (12.7%) had missing data on renal function (eGFR values).
Pollock and colleagues also queried the absence of bodyweight data in our results. We acknowledge that it may have been useful for the reader to see these data, although we presented data on BMI in Table 1 of our article as a proxy measure. Nevertheless, we can confirm that in our dataset very few patients had missing data on bodyweight (2.6% of patients starting on apixaban, 3.0% of those starting on dabigatran and 2.3% of those starting on rivaroxaban). Among patients with a recorded bodyweight, the mean bodyweight was 81.4 kg for patients starting on apixaban, 82.6 kg for those starting on dabigatran, and 82.0 kg for those starting on rivaroxaban. While...
Show MoreThank you for this critical note about vitamin D and B12 testing. The aim of our study was to explore the barriers and facilitators for reducing the number of unnecessary ordered vitamin D and B12 laboratory tests. We found that GPs experienced difficulty to request laborotory tests only for evidenced based indications; often vitamin testing was performed to satisfy patients' requests. We acknowledge the presence of certain medical indications to test vitamin D or B12 bloodlevels and we also performed a training for participating GPs of our study on vitamin D and B12 deficiency and people at risk of such deficiency. The purpose of our study was not to reduce the number of vitamin D and B12 tests to zero, but to explore the barriers and facilitators related to vitamin D and B12 testing in order to improve properly indicated vitamin testing in general practice.
Dear Sir or Madam
Re. Diagnosed prevalence of Ehlers-Danlos syndrome and hypermobility spectrum disorder in Wales, UK: a national electronic cohort study and case-control comparison.
Demmler J C, Atkinson M D, Reinhold E, Choy E, Lyons R A, Brophy S T
BMJ Open 2019;9:e031365
We write concerning the paper by Demmler et al., published in BMJ Open. We wish to raise the following concerns:
1. With regard to combining the Joint Hypermobility Syndrome (JHS) and Ehlers-Danlos syndromes (EDS) populations for analysis.
If one combines data from a cohort that is found to be ‘common’ (in this case ‘diagnosed JHS’) with one that is found to be ‘rare’ (in this case ‘diagnosed EDS’), the new combined cohort (i.e. diagnosed JHS/EDS) will be common. To then consider the rare cohort common is a fallacy.
Also, although individuals in a population with a previous diagnosis of JHS (i.e. prior to the 2017 international classification (1,2)) might have Hypermobile EDS (hEDS) by the current classification, it is not known how JHS segregates into Hypermobility Spectrum Disorder (HSD) and hEDS. A JHS population would need to be reassessed to confirm this, or modelling assumptions of the data would need to be applied.
Show MoreIn addition, it is not known what proportion of the EDS cohort have hEDS versus the rare Mendelian types of EDS. As such, there is no way of knowing whether or by what proportion the two cohorts represent the same or similar or dif...
Pages