Dear Editor,
The recently published study by Nightingale et al(1) on housing, neighbourhood and sociodemographic associations with adult levels of physical activity and adiposity found that those from lower socioeconomic background undertook less physical activity.(1) Insufficient physical activity is a modifiable risk factor for noncommunicable diseases such as heart disease, stroke, and diabetes.(2) Active travel such as walking and the use of bicycles increases the levels of physical activity and reduces emission from vehicle exhausts of the greenhouse gas carbon dioxide, as well as nitrous oxide and methane.(3)
There is widespread agreement that climate change is caused by anthropogenic increases in greenhouse gas emission and can be considered “the biggest global health threat of the 21st century”.(4) Extreme heat events due to climate change are increasing in frequency and intensity, increasing the risk of heat-related mortality especially in vulnerable populations(5) such as those of lower socioeconomic backgrounds or with noncommunicable diseases. Reduction of greenhouse gas emission is an important mitigation strategy to reduce climate change. Encouraging modes of transport not reliant on fossil fuels will reduce carbon dioxide emission, and would see a significant reduction in global warming in the future,(3) as well as provide health co-benefits from physical activity. Woodcock et al(6) produced a model of carbon dioxide emission for the city of...
Dear Editor,
The recently published study by Nightingale et al(1) on housing, neighbourhood and sociodemographic associations with adult levels of physical activity and adiposity found that those from lower socioeconomic background undertook less physical activity.(1) Insufficient physical activity is a modifiable risk factor for noncommunicable diseases such as heart disease, stroke, and diabetes.(2) Active travel such as walking and the use of bicycles increases the levels of physical activity and reduces emission from vehicle exhausts of the greenhouse gas carbon dioxide, as well as nitrous oxide and methane.(3)
There is widespread agreement that climate change is caused by anthropogenic increases in greenhouse gas emission and can be considered “the biggest global health threat of the 21st century”.(4) Extreme heat events due to climate change are increasing in frequency and intensity, increasing the risk of heat-related mortality especially in vulnerable populations(5) such as those of lower socioeconomic backgrounds or with noncommunicable diseases. Reduction of greenhouse gas emission is an important mitigation strategy to reduce climate change. Encouraging modes of transport not reliant on fossil fuels will reduce carbon dioxide emission, and would see a significant reduction in global warming in the future,(3) as well as provide health co-benefits from physical activity. Woodcock et al(6) produced a model of carbon dioxide emission for the city of London. They found that “an increase in active travel and less use of motor vehicles had larger health benefits per million population (7332 disability-adjusted life-years [DALYs] in 1 year) than from the increased use of lower-emission motor vehicles (160 DALYs).”(6) Combining active travel and lower-emission motor vehicles gave the largest benefits (7439 DALYs).(6)
In trying to establish an understanding of the causes of physical activity behaviour Bauman et al(7) identified a strong correlation between physical activity and recreation facilities, location, transportation environments and aesthetics. Nightingale similarly found that “perception of neighbourhood quality were associated with physical activity and adiposity”, especially among those disadvantaged.(1) Urban planning needs to facilitate incidental physical activity by ensuring bike paths and walkways are accessible and safe and lead to desired destinations, and are within an appropriate distance to public transport.(8) Urban planning should also ensure sufficient and accessible green spaces to enable recreational physical activity, especially in population dense urban areas.
Greening of cities also helps reduce the Urban Heat Island (UHI) phenomenon, which is exacerbated by more frequent heat waves caused by global warming. Higher temperatures felt in cities are due to high density building, impervious footpaths and roads and reduced green areas leading to heat retention.(9) These increased temperatures compound extreme heat events and have also been associated with detrimental health effects and higher mortality.(9) Thus urban planning which includes green areas of trees and other vegetation, wider streets allowing for more air flow and natural cooling at night will help reduce the UHI effect making the environment more conducive to physical activity.
To utilize the synergy of active travel on health and climate change, those practicing in public health, urban planning and transportation need to collaborate as a matter of urgency to reduce carbon emission and slow global warming.
References:
1. Nightingale CM, Rudnicka AR, Ram B, et al Housing, neighbourhood and sociodemographic associations with adult levels of physical activity and adiposity: baseline findings from the ENABLE London study. BMJ Open 2018;8:e02127. [cited 2018 Sep 17]. Available from: Doi:10.1136/bmjopen¬-¬2017¬-021257 https://bmjopen.bmj.com/content/8/8/e021257#DC1
2. World Health Organisation. Global Health Observatory data. Noncommunicable diseases. WHO [Internet] 2018 [cited 2018 Sep 21]. Available from: http://www.who.int/news-room/fact-sheets/detail/physical-activity
3. Australian Academy of Science. The science of climate change: Questions and answers. Australian Academy of Science [Internet] 2015 Feb [cited 2018 Sep 20]. Available from: https://www.science.org.au/learning/general-audience/science-booklets-0/...
4. Costello A, Abbas M, Allen A, Ball S, et al Managing the health effects of climate change. Lancet 2009 May 16 [cited 2018 Sep 21];373(9676):1693-733. Available from: https://doi.org/10.1016/S0140-6736(09)60935-1
5. Kovats RS, Hajat S. Heat stress and public health: A critical review. Annu Rev Public Health. 2008 Apr [cited 2018 Sep 20];29:41-55 Available from: https://doi.org/10.1146/annurev.publhealth.29.020907.090843
6. Woodcock J, Edwards P, Tonne C, et al Public health benefits of strategies to reduce greenhouse-gas emissions urban land transport. Lancet 2009 Nov 25 [cited 2018 Sep 18];374(9705):1930-43. Available from: DOI: https://doi.org/10.1016/S0140-6736(09)61714-1https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(09)61714-1/fulltext
7. Bauman AE, Reis RS, Sallis JF et al. Correlates of physical activity: why are some people active and others not? Lancet 2012 Jul 21 [cited 2018 Sep 19];380(9838):258-71. Available from: DOI: https://doi.org/10.1016/S0140-6736(12)60735-1 https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(12)60735-1/fulltext
8. Gerike R, de Nazelle A, Nieuwehuijsen M et al. Physical activity through sustainable transport approaches (PASTA): a study protocol for a multicenter project. BMJ Open 2016 [cited 2018 Sep 20];6:e009924. Available from: doi: 10.1136/bmjopen-2015-009924 https://bmjopen.bmj.com/content/6/1/e009924
9. Filho WL, Icaza LE, Emanche VO et al. An evidence-based review of impacts, strategies and tools to mitigate Urban Heat Islands. Int J Environ Res Public Health 2017 Dec [cited 2018 Sep 22];14(12):1600. Available from: doi:10.3390/ijerph14121600
This is an interesting article which provides useful insights for regulators. Unfortunately the authors state: 'DXP toxicity is mainly
due to its long half-life (15–37 hours), and
it can be increased by concomitant use of
alcohol or sedative drugs', quoting 2 rather old references.
We showed that deaths occurred before hospital admission, early in the course of poisoning, and postulated that this was likely due to the arrythmogenic effect of DXP, its main metabolite combined with the respiratory depressant effect the opioid action. This was one of the main reasons this drug was withdrawn in the UK, saving many hundreds of lives since withdrawal here. (Sandilands EA, Bateman DN. Co-proxamol withdrawal has reduced suicide from drugs in Scotland. British Journal of Clinical Pharmacology 2008; 66: 290-293.)
Data for this article is now available in the Dryad data repository
(doi:10.5061/dryad.2hr40) and can be viewed here
http://datadryad.org/resource/doi:10.5061/dryad.2hr40
In “Screening for carbon monoxide poisoning…in England: a prospective observational study,”(1) Clarke et al mischaracterize Masimo's Rad57 pulse CO-oximeter as a measure of venous carboxyhaemoglobin. As noted in one of their references(2) and the Rad57 Operator’s Manual,(3) Masimo’s trademarked “SpCO” measures arterial carboxyhaemoglobin. Based on this misunderstanding, the authors checked their [arterial] Rad57 ([a]Rad57) results against [venous] carboxyhaemoglobin ([v]COHb) measured by unspecified “point-of-care blood analyzers.”(4)
Instead of publishing these results separately, however, the authors simply combined them—by which 76 of 1758 patients were “positive” for CO poisoning (COp).(1) Only in an unpublished report to their funder did they disclose the R2 correlation among 608 paired [a]Rad57 and [v]COHb measurements was just 0.03.(4) Without any other testing, they assumed [v]COHb was more accurate and used this whenever available, even when [a]Rad57 was higher. By this method, they classified 293 with high [a]Rad57 but normal [v]COHb as false positives—and discharged them without the COp treatment and home inspection given 60 cases confirmed by high [v]COHb.(4)
Also disclosed only to the funder: the authors’ original protocol was “non-invasive” with a nested case-control design.(4) Blood CO-oximetry was only added after the controls—three per case—were dropped because “a number [unspecified] had high COHb readings [unspecified] on the...
In “Screening for carbon monoxide poisoning…in England: a prospective observational study,”(1) Clarke et al mischaracterize Masimo's Rad57 pulse CO-oximeter as a measure of venous carboxyhaemoglobin. As noted in one of their references(2) and the Rad57 Operator’s Manual,(3) Masimo’s trademarked “SpCO” measures arterial carboxyhaemoglobin. Based on this misunderstanding, the authors checked their [arterial] Rad57 ([a]Rad57) results against [venous] carboxyhaemoglobin ([v]COHb) measured by unspecified “point-of-care blood analyzers.”(4)
Instead of publishing these results separately, however, the authors simply combined them—by which 76 of 1758 patients were “positive” for CO poisoning (COp).(1) Only in an unpublished report to their funder did they disclose the R2 correlation among 608 paired [a]Rad57 and [v]COHb measurements was just 0.03.(4) Without any other testing, they assumed [v]COHb was more accurate and used this whenever available, even when [a]Rad57 was higher. By this method, they classified 293 with high [a]Rad57 but normal [v]COHb as false positives—and discharged them without the COp treatment and home inspection given 60 cases confirmed by high [v]COHb.(4)
Also disclosed only to the funder: the authors’ original protocol was “non-invasive” with a nested case-control design.(4) Blood CO-oximetry was only added after the controls—three per case—were dropped because “a number [unspecified] had high COHb readings [unspecified] on the pulse CO-oximeter without displaying symptoms of CO poisoning.”(4)
Unreported even to the funder was a protocol amendment approved in April 2010—four months after testing started—by the London-Surrey Research Ethics Committee (personal communication).
Had the authors realized Rad57 results were arterial, they might have published their controls’ results and some comparison with venous COHb, as this asthma study did.(5) Their unpublished report(4) shows the absolute [a]Rad57-[v]COHb difference was:
1) abnormal in 60% of 608 pairs, with 28% differing by 1-3%, 18% by 3-5%, and 14% by 5-31.7%, while only 40% differed by the 0-1% considered healthy;(5)
2) positive only in the seizure group, which had highest rate of COp detection by [a]Rad57 (5.0%), although the authors only published that it had the lowest rate by [v]COHb (2.1%).
Clearly, the [a]Rad57-[v]COHb difference is not due to instrument error. It varies by disorder in both direction and magnitude, with [a]Rad57>[v]COHb whenever free CO is diffusing from lungs and blood into other tissues, and vice-versa.
Unfortunately, the authors’ withholding of all this information from their paper meets BMJ’s definition of research misconduct and warrants retraction.(6)
References
1. Clarke S, Keshishian C, Murray V, et al. for the Carbon Monoxide in Emergency Departments (COED) Working Group. Screening for carbon monoxide exposure in selected patient groups attending rural and urban emergency departments in England: a prospective observational study. BMJ Open 2012;2:e000877
2. Barker SJ, Curry J, Redford D, et al. Measurement of carboxyhaemoglobin and methemoglobin by pulse oximetry. Anesthesiol 2006;105:892–7
4. Clarke S, Coultrip E, Oetterli S, et al. Earle D, Keshishian C, Murray V, Ruggles R, Ward P, Bush S. Non-invasive screening for carbon monoxide exposure in selected patient groups attending rural and urban emergency departments in England. Report to the Department of Health, PRP:002/0030. Final Report, Second Revision, August 2011. Released 17 August 2018 by the UK NIHR Central Commissioning Facility in WORD format and archived by AD in PDF format at https://www.dropbox.com/s/ww11hzuv9gsngqi/Clarke%202011%202nd%20revision... [accessed 21 August 2018]
[AD note: A later 2012 version of this Final Report is referenced in an unrelated 2017 report entitled “Carbon Monoxide Pre-Hospital Screening Study,” by Foster T, Scott T, and Prothero L. https://www.coportal.org/wp-content/uploads/gst_attachments/b85a312fdeba... [accessed 20 August 2018].
The 2012 version has the same title and nine authors as Clarke et al 2011(4), but eight are re-ordered. Clarke is still first but Keshishian is moved from fifth to second, Murray from sixth to third, Coultrip from second to fourth, Oetterli from third to fifth, Earle from fourth to sixth, Ward from eighth to seventh, Bush from ninth to eighth and Ruggles from seventh to ninth. The list of authors was changed again in the manuscript submitted to BMJ Open on 17 February 2012, when two authors were added for the first time--Kafatos third and Porter eleventh—and Ruggles was moved from last to fourth(1). ]
5. Naples R, Laskowski D, McCarthy K, et al. Carboxyhemoglobin and methemoglobin in asthma. Lung. 2015 Apr;193(2):183-7
6. Anonymous. Analysis: A consensus statement on research misconduct in the UK. BMJ 2012;344:e1111
This cross-sectional survey was a very good starting point for further detailed study between alcohol uses and other clinical implications e.g. behavioral disorder, mood disorder or psychiatric disorder besides emotional effect in different countries and sociodemographic backgrounds.
Overall the survey showed clear-cut connection between various positive and negative feelings with level of risk of alcohol use classified using AUDIT scores. However, selection bias existing in this study, either the under-coverage of the representative populations as the survey was only confined to people who had internet connection and also the small sample of people who did not attend high school. Response bias could probably be involved and must be also taken into consideration in setting up the survey. And individual co-morbidities must be included to further reduce the biases.
According to Constitution-Dependent, Inherited Real Risks (1), Hashimoto’s thyroiditis depends on Autoimmune Constitution (2). On the contrary, exclusively individuals involved by Lithiasis Constitution can suffer from Cholelithiasis (3, 4). Importantly, as all other Inherited Real Risks, also those mentioned above, bedside diagnosed since birth with a common stethoscope, are removed by inexpensive Reconstructing Mitochondrial Quantum Therapy (5).
References.
1) Stagnaro Sergio. Reale Rischio Semeiotico Biofisico. I Dispositivi Endoarteriolari di Blocco neoformati, patologici, tipo I, sottotipo a) oncologico, e b) aspecifico. Ediz. Travel Factory, www.travelfactory.it, Roma, 2009.
2) Stagnaro S., Sindrome percusso-ascoltatoria autoimmune. Gazz. Med. It. 142, 555, 1983.
3) Simone Caramel and Sergio Stagnaro (2012). Vascular calcification and Inherited Real Risk of lithiasis. Front. In Encocrin. 3:119. doi: 10.3389/fendo.2012.00119 http://www.frontiersin.org/Bone_Research/10.3389/fendo.2012.00119/full [MEDLINE]
4) Stagnaro S., Stagnaro-Neri M., Diagnosi percusso-ascoltatoria dei calcoli biliari silenti. 6° Incontro Segusino di Medicina e Chirurgia. Susa 19 Maggio, 1990. Atti, pg. 79. Ed. Minerva Medica
5) Caramel S., Marchionni M., Stagnaro S. Morinda citrifolia Plays a Central Role in the Primary Prevention...
According to Constitution-Dependent, Inherited Real Risks (1), Hashimoto’s thyroiditis depends on Autoimmune Constitution (2). On the contrary, exclusively individuals involved by Lithiasis Constitution can suffer from Cholelithiasis (3, 4). Importantly, as all other Inherited Real Risks, also those mentioned above, bedside diagnosed since birth with a common stethoscope, are removed by inexpensive Reconstructing Mitochondrial Quantum Therapy (5).
References.
1) Stagnaro Sergio. Reale Rischio Semeiotico Biofisico. I Dispositivi Endoarteriolari di Blocco neoformati, patologici, tipo I, sottotipo a) oncologico, e b) aspecifico. Ediz. Travel Factory, www.travelfactory.it, Roma, 2009.
2) Stagnaro S., Sindrome percusso-ascoltatoria autoimmune. Gazz. Med. It. 142, 555, 1983.
3) Simone Caramel and Sergio Stagnaro (2012). Vascular calcification and Inherited Real Risk of lithiasis. Front. In Encocrin. 3:119. doi: 10.3389/fendo.2012.00119 http://www.frontiersin.org/Bone_Research/10.3389/fendo.2012.00119/full [MEDLINE]
4) Stagnaro S., Stagnaro-Neri M., Diagnosi percusso-ascoltatoria dei calcoli biliari silenti. 6° Incontro Segusino di Medicina e Chirurgia. Susa 19 Maggio, 1990. Atti, pg. 79. Ed. Minerva Medica
5) Caramel S., Marchionni M., Stagnaro S. Morinda citrifolia Plays a Central Role in the Primary Prevention of Mitochondrial-dependent Degenerative Disorders. Asian Pac J Cancer Prev. 2015;16(4):1675. http://www.ncbi.nlm.nih.gov/pubmed/25743850[MEDLINE]
One retrospective cohort study conducted by Singh and colleagues reported that gout is associated with a 1.4-fold increased risk of hearing impairment in older adults (hazard ratio 1.44, 95% CI 1.40-1.49).1 Some caveats should be discussed. Gout is a form of arthritis due to the deposition of monosodium urate crystals within joints, which is associated with persistently high levels of uric acid in the blood.2 Clinically, it is not feasible to check the uric acid levels every day. So we cannot be sure the onset date of hyperuricemia. Because hearing impairment is an insidious condition without a well-defined onset date, we can only approximate the onset date by applying the claims-based definition of ICD-9-CM 389. When a cohort study examines the relationship between hyperuricemia and hearing impairment, we cannot definitely determine which condition comes first because of the onset date not clear. Gout is characterized by recurrent acute attacks of joint inflammation.3 When an acute attack subsides, the joint inflammation is also relieved. Thus, how can we make a reasonable link between the remission state of gout and hearing impairment? That is, how does the remission state of gout have an impact on the insidious course of hearing impairment? Similarly Parkinson’s disease is an insidious condition, but no association is detected between gout and Parkinson’s disease in older adults in Taiwan.4
Perhaps there could be an association between gout and hearing impairment....
One retrospective cohort study conducted by Singh and colleagues reported that gout is associated with a 1.4-fold increased risk of hearing impairment in older adults (hazard ratio 1.44, 95% CI 1.40-1.49).1 Some caveats should be discussed. Gout is a form of arthritis due to the deposition of monosodium urate crystals within joints, which is associated with persistently high levels of uric acid in the blood.2 Clinically, it is not feasible to check the uric acid levels every day. So we cannot be sure the onset date of hyperuricemia. Because hearing impairment is an insidious condition without a well-defined onset date, we can only approximate the onset date by applying the claims-based definition of ICD-9-CM 389. When a cohort study examines the relationship between hyperuricemia and hearing impairment, we cannot definitely determine which condition comes first because of the onset date not clear. Gout is characterized by recurrent acute attacks of joint inflammation.3 When an acute attack subsides, the joint inflammation is also relieved. Thus, how can we make a reasonable link between the remission state of gout and hearing impairment? That is, how does the remission state of gout have an impact on the insidious course of hearing impairment? Similarly Parkinson’s disease is an insidious condition, but no association is detected between gout and Parkinson’s disease in older adults in Taiwan.4
Perhaps there could be an association between gout and hearing impairment. Currently, there is no definite evidence to prove a causal relationship between hyperuricemia and hearing impairment.
Contributors
Shih-Wei Lai contributed to the conception of the article and initiated the draft of the article.
Competing interests
The author reports no conflicts of interest.
REFERENCES
1. Singh JA, Cleveland JD. Gout and hearing impairment in the elderly: a retrospective cohort study using the US Medicare claims data. BMJ Open 2018;8:2018-022854.
2. Richette P, Bardin T. Gout. Lancet 2010;375:318-28.
3. Hainer BL, Matheson E, Wilkes RT. Diagnosis, treatment, and prevention of gout. Am Fam Physician 2014;90:831-6.
4. Lai SW, Lin CH, Lin CL, Liao KF. Gout and Parkinson's Disease in Older People: An Observation in Taiwan. International Journal of Gerontology 2014;8:166-7.
In their manuscript, Wong et al. discuss an important aspect of randomized control trial (RCT) design: sample size determination. Although their focus is on RCTs for sepsis with the primary outcome of mortality, their review is generalizable to RCTs with binary primary outcomes (i.e. the presence or absence of an event of interest) and thus, may influence the design of RCTs in many patient populations. Therefore, it is important to address one of the primary conclusions of the manuscript.
Wong et al. reviewed the sample size calculations for 13 RCTs for sepsis where the average treatment effect was defined as the absolute difference in the mortality rate comparing the control arm to the intervention arm. We will subsequently refer to this average treatment effect as AD. To determine the required sample size needed per arm for the RCT, it is required to specify both the anticipated mortality rate in the control arm and the AD. For the 13 RCTs, Wong et al. extracted and compared, via meta-analysis, the anticipated mortality rate in the control arm and AD used in the sample size calculation to the respective values obtained from the completed RCTs. They found that for both the control arm mortality rate and the AD, the anticipated values were, on average, greater than the values from the completed RCTs. The third paragraph of their discussion states “The consistent overestimation of control arm event rate (or lower than anticipated actual control arm event rate...
In their manuscript, Wong et al. discuss an important aspect of randomized control trial (RCT) design: sample size determination. Although their focus is on RCTs for sepsis with the primary outcome of mortality, their review is generalizable to RCTs with binary primary outcomes (i.e. the presence or absence of an event of interest) and thus, may influence the design of RCTs in many patient populations. Therefore, it is important to address one of the primary conclusions of the manuscript.
Wong et al. reviewed the sample size calculations for 13 RCTs for sepsis where the average treatment effect was defined as the absolute difference in the mortality rate comparing the control arm to the intervention arm. We will subsequently refer to this average treatment effect as AD. To determine the required sample size needed per arm for the RCT, it is required to specify both the anticipated mortality rate in the control arm and the AD. For the 13 RCTs, Wong et al. extracted and compared, via meta-analysis, the anticipated mortality rate in the control arm and AD used in the sample size calculation to the respective values obtained from the completed RCTs. They found that for both the control arm mortality rate and the AD, the anticipated values were, on average, greater than the values from the completed RCTs. The third paragraph of their discussion states “The consistent overestimation of control arm event rate (or lower than anticipated actual control arm event rate) may have systematically led to undersized trials from the outset; i.e. given the actual control arm mortality the trials would have been designed to include more patients. This has likely meant that there has been an increased risk of type 2 errors in many sepsis trials that could have potentially resulted in the disregarding of potentially useful treatments.” We believe that this statement requires clarification.
Take, as a hypothetical example, a typical 2-arm, parallel-group sepsis RCT designed to detect an AD of 10% at 28 days post-randomization, assuming the standard 5% Type I error rate and 80% power. Using historical data, the control arm mortality rate is anticipated to be 40%, implying an intervention arm mortality rate of 30%, yielding a required sample size of 356 patients per arm (1,2). The findings of Wong et al. suggest that the anticipated control arm mortality rate may be too high. So as an alternative, we consider computing the required sample size using a reduced control arm mortality rate of 30% (i.e. the sample size will be based on comparing a 30% vs. 20% mortality rate in the control and intervention arms, respectively) which yields a required sample size of only 294 patients per arm. Hence, it is not true that the required sample size of the RCT is underestimated when using the greater of the two control arm morality rates (i.e. using 40% rather than 30%). In fact, for a given absolute difference in the mortality rates (e.g. AD of 10%), assuming a higher mortality rate in the control arm of 50% will maximize the required sample size (e.g., 388 patients per arm in this example) and is a conservative approach for calculating the sample size for a sepsis RCT. This result relies on properties of the Binomial distribution, with some explanation provided below.
For a fixed Type I error rate and power, the sample size calculation is driven by the anticipated AD, as well as, the variation in the number of deaths expected in each arm, which depends on the anticipated control and intervention arm mortality rates. The latter is attributable to the process of sampling; for instance, two samples of 356 patients drawn, at random, from a large population of sepsis patients will yield different numbers of deaths despite the underlying mortality rate being the same for each patient in the population. So why is the variation in the number of deaths greater when the underlying mortality rate is 50% compared to say 20%? Note: here you can replace 20% with any number less than 50% and the same logic applies.
To answer this question while keeping the numeric presentation simple, let’s consider an unrealistic sample size of 10 patients from the population. Assuming an underlying mortality rate of 20% which is the same for each patient, then we will observe 0, 1, 2, 3, 4, 5, and > 5 deaths in roughly 10%, 27%, 30%, 20%, 9%, 3%, and < 1% respectively, of 10-patient samples drawn from the population. However, if the true mortality rate was 50% (rather than 20%), then we would observe <2, 2, 3, 4, 5, 6, 7, 8, and > 8 deaths in roughly 1, 4, 12, 20, 25, 20, 12, 4 and 1 percent of samples of size 10 drawn from the population, respectively. Thus, there is greater variation in the number of deaths from a sample of size 10 when the mortality rate is 50% rather than 20%. Applying the properties from the Binomial distribution, the variance of the number of deaths in samples of size 10 is 2.5 under the mortality rate of 50% vs. 1.6 under the mortality rate of 20%. Intuitively, the sample size required to detect a fixed AD will increase as the variance of the outcome of interest increases, holding Type I error and power fixed; and if the projected mortality rate used for a sample size calculation is greater (i.e. closer to 50% rate) than the actual mortality rate for the control group, then the sample size is larger than required and can be considered conservative.
It is also important to clarify Figure 5 from the Discussion, which the authors use to support their conclusions on the potential harm of overestimating the control arm mortality rate. Figure 5 presents the sample size required per arm to detect a 10% relative reduction in mortality rate (i.e. -0.10 = 1 – intervention arm mortality rate / control arm mortality rate) as a function of the control arm mortality rate, for fixed 5% Type I error rate and 80% power. The figure demonstrates that as the control arm mortality rate decreases, the required sample size increases. This finding may appear to contradict the arguments above; however, it does not. By redefining the average treatment effect as the relative reduction in mortality and fixing this to 10%, the pattern observed in Figure 5 is attributable to changes in the AD not changes in the control arm mortality rate. To demonstrate this, we will use the example presented by Wong et al. of the sample size required for a sepsis RCT with an 18.4% control arm mortality rate and 10% relative reduction in mortality (i.e. AD of 1.84% with 16.6% intervention arm mortality rate) requiring roughly 6700 patients per arm. Changing the control arm mortality rate even modestly to 20% with the fixed 10% relative reduction (i.e. AD of 2%) reduces the required sample size to roughly 6000 patients per arm. The feature driving this reduction (6700 vs. 6000) is the increase in the AD (1.84% vs. 2%) not the control arm mortality rate. If the AD was fixed at 1.84% but the control arm mortality rate was increased to 20%, then the required sample size per arm is roughly 7100 patients per arm (larger than the required sample size of 6700 under control arm mortality rate of 18.4%) which is consistent with the properties of the sample size calculations for tests of two proportions.
We believe that the critical finding from Wong et al. leading to underestimation of sample sizes, is that the AD (i.e. the average treatment effect) observed in completed sepsis RCTs is systematically smaller than the anticipated AD used in sample size calculations – i.e. investigators are commonly over-estimating the projected mortality benefit of their intervention yielding a smaller required sample size. For instance, from the 13 sepsis RCTs reported in their paper, the projected mortality benefits ranged from 6.5% to 12.5% with a median (inter-quartile range) of 9% (7.5% – 10%) whereas the observed mortality benefits ranged from -8% to 2% with a median (inter-quartile range) of -2% (-2.6% to 0.3%). Combining these results from Wong et al. with results from a recent paper by Shankar-Hari et al.3, which estimated that the fraction of deaths attributable to sepsis for ICU patients with sepsis was only 15%, should remind sepsis investigators to carefully consider their estimates of the anticipated average treatment effect when conducting sample size calculations.
References:
1) Agresti, A. 2013. Categorical Data Analysis. 3rd ed. Hoboken, NJ: Wiley.
2) Computed using “power twoproportions 0.4 0.3” in Stata version 15.
3) Shankar-Hari M, Harrison DA, Rowan KM, Rubenfeld GD. Estimating attributable fraction of mortality from sepsis to inform clinical trials. J Crit Care. 2018; 45: 33-39. doi: 10.1016/j.jrcrc.2018.01.018
The Moncada's Manoeuvre, among a lot of other clinical manoeuvres and signs, allows doctors to bedside differentiate easily and quickly benigne from, even initally, malignant disorders (1-3).
For instance, the differential diagnosis between urinary tract hemorrhage benign or malignant in nature is often difficult ayt the bedside. In addition, Moncada 'Manoeuvre proved to be very useful in the differential diagnosis between prostate adenoma and prostate cancer, even initial, as Oncological Terrain-Dependent, Inherited Real Risk..
Reference
1)Sergio Stagnaro (2018).Manovra di Moncada*: Diagnosi Differenziale tra Lesione Benigna e Maligna in 15 secondi. in pdf http://www.sisbq.org/uploads/5/6/8/7/5687930/manovradimoncada.pdf
2) Sergio Stagnaro. Massucco’s Sign in the war against to Prostate Cancer. Letter to FDA; www.melatonina.it ; 2 May, 2010, http://www.melatonina.it/articoli/247-2010-05-02.html
3) Sergio Stagnaro. Bedside Detecting Inherited Real Risk of Prostate Cancer, and overt Cancer: Massucco’s Sign. European Urology. 27 April, 2011, http://www.europeanurology.com/article/S0302-2838%2810%2900944-9/fulltex...
The effect of Thiazolidinedione therapy on the risk of Parkinson’s disease is controversial. One article by Wu and colleagues in Taiwan reported that pioglitazone use was not associated with the risk of Parkinson’s disease in people with diabetes mellitus (hazard ratio 0.90, 95% CI 0.68-1.18).1 To the contrary, another article by Lin and colleagues in Taiwan reported that thiazolidinedione use was associated with a 60% reduced risk of Parkinson’s disease in people with diabetes mellitus.2 Similarly, conflicting results were also found in Western countries.3,4 There was no measure of the hemoglobin A1c in the above studies. We cannot determine whether the risk of Parkinson’s disease is associated with good glycemic control or poor control among people on thiazolidinedione therapy. Therefore, any study exploring the association between anti-diabetic medications and Parkinson’s disease should estimate the hemoglobin A1c levels.
Theoretically, a study should be designed to compare people on thiazolidinedione therapy only with those people not taking any medication. However, according to the recommendation of the American Diabetes Association, metformin is the first-line therapy for type 2 diabetes mellitus.5 People with diabetes mellitus usually take metformin alone or use combined therapy with other anti-diabetic medications. Thiazolidinedione is usually recommended as combined therapy with metformin for type 2 diabetes mellitus. So it is difficult to identify peo...
The effect of Thiazolidinedione therapy on the risk of Parkinson’s disease is controversial. One article by Wu and colleagues in Taiwan reported that pioglitazone use was not associated with the risk of Parkinson’s disease in people with diabetes mellitus (hazard ratio 0.90, 95% CI 0.68-1.18).1 To the contrary, another article by Lin and colleagues in Taiwan reported that thiazolidinedione use was associated with a 60% reduced risk of Parkinson’s disease in people with diabetes mellitus.2 Similarly, conflicting results were also found in Western countries.3,4 There was no measure of the hemoglobin A1c in the above studies. We cannot determine whether the risk of Parkinson’s disease is associated with good glycemic control or poor control among people on thiazolidinedione therapy. Therefore, any study exploring the association between anti-diabetic medications and Parkinson’s disease should estimate the hemoglobin A1c levels.
Theoretically, a study should be designed to compare people on thiazolidinedione therapy only with those people not taking any medication. However, according to the recommendation of the American Diabetes Association, metformin is the first-line therapy for type 2 diabetes mellitus.5 People with diabetes mellitus usually take metformin alone or use combined therapy with other anti-diabetic medications. Thiazolidinedione is usually recommended as combined therapy with metformin for type 2 diabetes mellitus. So it is difficult to identify people on thiazolidinedione therapy alone without using metformin in a study using claim data. Whether the medication effect on the risk of Parkinson’s disease is associated with thiazolidinedione alone or metformin-thiazolidinedione combination should be addressed in future studies. Currently, there is no definite evidence to prove thiazolidinedione having a neuroprotective effect against the development of Parkinson’s disease.
Contributors
Shih-Wei Lai contributed to the conception of the article and initiated the draft of the article.
Competing interests
The author report no conflicts of interest.
REFERENCES
1. Wu HF, Kao LT, Shih JH, et al. Pioglitazone use and Parkinson's disease: a retrospective cohort study in Taiwan. BMJ Open 2018;8:2018-023302.
2. Lin HL, Lin HC, Tseng YF, Chao JC, Hsu CY. Association of thiazolidinedione with a lower risk of Parkinson's disease in a population with newly-diagnosed diabetes mellitus. Ann Med 2018;50:430-6.
3. Connolly JG, Bykov K, Gagne JJ. Thiazolidinediones and Parkinson Disease: A Cohort Study. Am J Epidemiol 2015;182:936-44.
4. Brakedal B, Flones I, Reiter SF, et al. Glitazone use associated with reduced risk of Parkinson's disease. Mov Disord 2017;32:1594-9.
5. Pharmacologic Approaches to Glycemic Treatment. Diabetes Care 2017;40:S64-S74.
Dear Editor,
Show MoreThe recently published study by Nightingale et al(1) on housing, neighbourhood and sociodemographic associations with adult levels of physical activity and adiposity found that those from lower socioeconomic background undertook less physical activity.(1) Insufficient physical activity is a modifiable risk factor for noncommunicable diseases such as heart disease, stroke, and diabetes.(2) Active travel such as walking and the use of bicycles increases the levels of physical activity and reduces emission from vehicle exhausts of the greenhouse gas carbon dioxide, as well as nitrous oxide and methane.(3)
There is widespread agreement that climate change is caused by anthropogenic increases in greenhouse gas emission and can be considered “the biggest global health threat of the 21st century”.(4) Extreme heat events due to climate change are increasing in frequency and intensity, increasing the risk of heat-related mortality especially in vulnerable populations(5) such as those of lower socioeconomic backgrounds or with noncommunicable diseases. Reduction of greenhouse gas emission is an important mitigation strategy to reduce climate change. Encouraging modes of transport not reliant on fossil fuels will reduce carbon dioxide emission, and would see a significant reduction in global warming in the future,(3) as well as provide health co-benefits from physical activity. Woodcock et al(6) produced a model of carbon dioxide emission for the city of...
This is an interesting article which provides useful insights for regulators. Unfortunately the authors state: 'DXP toxicity is mainly
due to its long half-life (15–37 hours), and
it can be increased by concomitant use of
alcohol or sedative drugs', quoting 2 rather old references.
We showed that deaths occurred before hospital admission, early in the course of poisoning, and postulated that this was likely due to the arrythmogenic effect of DXP, its main metabolite combined with the respiratory depressant effect the opioid action. This was one of the main reasons this drug was withdrawn in the UK, saving many hundreds of lives since withdrawal here. (Sandilands EA, Bateman DN. Co-proxamol withdrawal has reduced suicide from drugs in Scotland. British Journal of Clinical Pharmacology 2008; 66: 290-293.)
Data for this article is now available in the Dryad data repository (doi:10.5061/dryad.2hr40) and can be viewed here http://datadryad.org/resource/doi:10.5061/dryad.2hr40
Conflict of Interest:
BMJ Open member of staff
In “Screening for carbon monoxide poisoning…in England: a prospective observational study,”(1) Clarke et al mischaracterize Masimo's Rad57 pulse CO-oximeter as a measure of venous carboxyhaemoglobin. As noted in one of their references(2) and the Rad57 Operator’s Manual,(3) Masimo’s trademarked “SpCO” measures arterial carboxyhaemoglobin. Based on this misunderstanding, the authors checked their [arterial] Rad57 ([a]Rad57) results against [venous] carboxyhaemoglobin ([v]COHb) measured by unspecified “point-of-care blood analyzers.”(4)
Instead of publishing these results separately, however, the authors simply combined them—by which 76 of 1758 patients were “positive” for CO poisoning (COp).(1) Only in an unpublished report to their funder did they disclose the R2 correlation among 608 paired [a]Rad57 and [v]COHb measurements was just 0.03.(4) Without any other testing, they assumed [v]COHb was more accurate and used this whenever available, even when [a]Rad57 was higher. By this method, they classified 293 with high [a]Rad57 but normal [v]COHb as false positives—and discharged them without the COp treatment and home inspection given 60 cases confirmed by high [v]COHb.(4)
Also disclosed only to the funder: the authors’ original protocol was “non-invasive” with a nested case-control design.(4) Blood CO-oximetry was only added after the controls—three per case—were dropped because “a number [unspecified] had high COHb readings [unspecified] on the...
Show MoreThis cross-sectional survey was a very good starting point for further detailed study between alcohol uses and other clinical implications e.g. behavioral disorder, mood disorder or psychiatric disorder besides emotional effect in different countries and sociodemographic backgrounds.
Overall the survey showed clear-cut connection between various positive and negative feelings with level of risk of alcohol use classified using AUDIT scores. However, selection bias existing in this study, either the under-coverage of the representative populations as the survey was only confined to people who had internet connection and also the small sample of people who did not attend high school. Response bias could probably be involved and must be also taken into consideration in setting up the survey. And individual co-morbidities must be included to further reduce the biases.
According to Constitution-Dependent, Inherited Real Risks (1), Hashimoto’s thyroiditis depends on Autoimmune Constitution (2). On the contrary, exclusively individuals involved by Lithiasis Constitution can suffer from Cholelithiasis (3, 4). Importantly, as all other Inherited Real Risks, also those mentioned above, bedside diagnosed since birth with a common stethoscope, are removed by inexpensive Reconstructing Mitochondrial Quantum Therapy (5).
References.
1) Stagnaro Sergio. Reale Rischio Semeiotico Biofisico. I Dispositivi Endoarteriolari di Blocco neoformati, patologici, tipo I, sottotipo a) oncologico, e b) aspecifico. Ediz. Travel Factory, www.travelfactory.it, Roma, 2009.
Show More2) Stagnaro S., Sindrome percusso-ascoltatoria autoimmune. Gazz. Med. It. 142, 555, 1983.
3) Simone Caramel and Sergio Stagnaro (2012). Vascular calcification and Inherited Real Risk of lithiasis. Front. In Encocrin. 3:119. doi: 10.3389/fendo.2012.00119 http://www.frontiersin.org/Bone_Research/10.3389/fendo.2012.00119/full [MEDLINE]
4) Stagnaro S., Stagnaro-Neri M., Diagnosi percusso-ascoltatoria dei calcoli biliari silenti. 6° Incontro Segusino di Medicina e Chirurgia. Susa 19 Maggio, 1990. Atti, pg. 79. Ed. Minerva Medica
5) Caramel S., Marchionni M., Stagnaro S. Morinda citrifolia Plays a Central Role in the Primary Prevention...
One retrospective cohort study conducted by Singh and colleagues reported that gout is associated with a 1.4-fold increased risk of hearing impairment in older adults (hazard ratio 1.44, 95% CI 1.40-1.49).1 Some caveats should be discussed. Gout is a form of arthritis due to the deposition of monosodium urate crystals within joints, which is associated with persistently high levels of uric acid in the blood.2 Clinically, it is not feasible to check the uric acid levels every day. So we cannot be sure the onset date of hyperuricemia. Because hearing impairment is an insidious condition without a well-defined onset date, we can only approximate the onset date by applying the claims-based definition of ICD-9-CM 389. When a cohort study examines the relationship between hyperuricemia and hearing impairment, we cannot definitely determine which condition comes first because of the onset date not clear. Gout is characterized by recurrent acute attacks of joint inflammation.3 When an acute attack subsides, the joint inflammation is also relieved. Thus, how can we make a reasonable link between the remission state of gout and hearing impairment? That is, how does the remission state of gout have an impact on the insidious course of hearing impairment? Similarly Parkinson’s disease is an insidious condition, but no association is detected between gout and Parkinson’s disease in older adults in Taiwan.4
Show MorePerhaps there could be an association between gout and hearing impairment....
In their manuscript, Wong et al. discuss an important aspect of randomized control trial (RCT) design: sample size determination. Although their focus is on RCTs for sepsis with the primary outcome of mortality, their review is generalizable to RCTs with binary primary outcomes (i.e. the presence or absence of an event of interest) and thus, may influence the design of RCTs in many patient populations. Therefore, it is important to address one of the primary conclusions of the manuscript.
Wong et al. reviewed the sample size calculations for 13 RCTs for sepsis where the average treatment effect was defined as the absolute difference in the mortality rate comparing the control arm to the intervention arm. We will subsequently refer to this average treatment effect as AD. To determine the required sample size needed per arm for the RCT, it is required to specify both the anticipated mortality rate in the control arm and the AD. For the 13 RCTs, Wong et al. extracted and compared, via meta-analysis, the anticipated mortality rate in the control arm and AD used in the sample size calculation to the respective values obtained from the completed RCTs. They found that for both the control arm mortality rate and the AD, the anticipated values were, on average, greater than the values from the completed RCTs. The third paragraph of their discussion states “The consistent overestimation of control arm event rate (or lower than anticipated actual control arm event rate...
Show MoreThe Moncada's Manoeuvre, among a lot of other clinical manoeuvres and signs, allows doctors to bedside differentiate easily and quickly benigne from, even initally, malignant disorders (1-3).
For instance, the differential diagnosis between urinary tract hemorrhage benign or malignant in nature is often difficult ayt the bedside. In addition, Moncada 'Manoeuvre proved to be very useful in the differential diagnosis between prostate adenoma and prostate cancer, even initial, as Oncological Terrain-Dependent, Inherited Real Risk..
Reference
1)Sergio Stagnaro (2018).Manovra di Moncada*: Diagnosi Differenziale tra Lesione Benigna e Maligna in 15 secondi. in pdf http://www.sisbq.org/uploads/5/6/8/7/5687930/manovradimoncada.pdf
2) Sergio Stagnaro. Massucco’s Sign in the war against to Prostate Cancer. Letter to FDA; www.melatonina.it ; 2 May, 2010, http://www.melatonina.it/articoli/247-2010-05-02.html
3) Sergio Stagnaro. Bedside Detecting Inherited Real Risk of Prostate Cancer, and overt Cancer: Massucco’s Sign. European Urology. 27 April, 2011, http://www.europeanurology.com/article/S0302-2838%2810%2900944-9/fulltex...
The effect of Thiazolidinedione therapy on the risk of Parkinson’s disease is controversial. One article by Wu and colleagues in Taiwan reported that pioglitazone use was not associated with the risk of Parkinson’s disease in people with diabetes mellitus (hazard ratio 0.90, 95% CI 0.68-1.18).1 To the contrary, another article by Lin and colleagues in Taiwan reported that thiazolidinedione use was associated with a 60% reduced risk of Parkinson’s disease in people with diabetes mellitus.2 Similarly, conflicting results were also found in Western countries.3,4 There was no measure of the hemoglobin A1c in the above studies. We cannot determine whether the risk of Parkinson’s disease is associated with good glycemic control or poor control among people on thiazolidinedione therapy. Therefore, any study exploring the association between anti-diabetic medications and Parkinson’s disease should estimate the hemoglobin A1c levels.
Show MoreTheoretically, a study should be designed to compare people on thiazolidinedione therapy only with those people not taking any medication. However, according to the recommendation of the American Diabetes Association, metformin is the first-line therapy for type 2 diabetes mellitus.5 People with diabetes mellitus usually take metformin alone or use combined therapy with other anti-diabetic medications. Thiazolidinedione is usually recommended as combined therapy with metformin for type 2 diabetes mellitus. So it is difficult to identify peo...
Pages