Objective To describe neurological outcomes after sudden cardiac arrests (SCAs) in road and long-distance races using a rapid mobile automated external defibrillator system (RMAEDS) intervention.
Methods A total of 42 SCAs from 3 214 701 runners in 334 road and long-distance races from 1 February 2007 to 29 February 2020 were examined. Demographics, SCA interventions, EMS-related data and SCA-related outcomes were measured. Primary endpoints were favourable neurological outcomes (Cerebral Performance Categories 1–2) at 1-month and 1-year post-SCA. Secondary endpoints were factors related to the field return of spontaneous circulation (ROSC) and resuscitation characteristics, including the initial ECG waveform classification and resuscitation sequence times according to the initial ECG rhythm.
Results The SCA incidence rate was 1.31 per 100 000 runners (age: median (IQR), 51 (36.5, 58.3) years). Field ROSC and full neurological recovery at 1-month post-SCA was achieved 90.4% and 92.9% of cases, respectively. In 22 cases in which bystander cardiopulmonary resuscitation was initiated within 1 min and defibrillation performed within 3 min, full neurological recovery was achieved at 1-month and 1-year post-SCA in 95.5.% and 95.5% of cases, respectively.
Conclusions The RMAEDS successfully treated patients with SCA during road and long-distance races yielding a high survival rate and favourable neurological outcomes. These findings support rapid intervention and the proper placement of healthcare teams along the race course to initiate chest compressions within 1 min and perform defibrillation within 3 min.
- cardiovascular diseases
- sports medicine
Data availability statement
No data are available.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
WHAT IS ALREADY KNOWN ON THIS TOPIC
The incidence of sudden cardiac arrest (SCA) associated with marathons is 0.54 per 100 000 runners, and 71% of these runners have died. In our previous study, immediate cardiopulmonary resuscitation (CPR) and defibrillation resulted in 100% survival in 28 witnessed SCA cases, but details of the emergency response and long-term neurological outcomes were not reported. Creating a clear emergency action plan and strategically positioning healthcare providers are vital to early recognition and early intervention, which leads to improved survival in marathon-related SCA.
WHAT THIS STUDY ADDS
A combination of prompt CPR initiation and automated external defibrillator (AED) defibrillation within 3 min will save 95% of SCA cases in marathons. Achieving the return of spontaneous circulation early in cardiac arrest provides favourable long-term neurological outcomes in 95% of cases. A rapid mobile automated external defibrillator system via a bicycle-equipped medical support team allows active course surveillance and rapid emergency response.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE AND/OR POLICY
Sports event organisers should provide a sufficient number of AEDs and first responders to improve survival and neurological outcomes from SCA. To achieve high survival rates at sporting events, it is important to develop detailed event-day protocols for first responders. Participating staff, including physicians, nurses, paramedics and volunteers, should be deployed according to the protocol and positioned to perform early CPR and defibrillation within 3 min.
Sudden cardiac arrest (SCA) is a leading cause of mortality; reports indicate 1 in 7.4 people die due to sudden cardiac death (SCD).1 In 2019, 126 271 out-of-hospital cardiac arrests (OHCAs) occurred in Japan.2 In adults, 39% of SCAs were categorised as sports-related.1 In marathons, research indicates the SCD rate ranges from 0.24 to 0.39 per 100 000 runners.3 4 A study reported the incidence of SCA as 0.54 per 100 000 runners and found 42 cases of SCD among 59 total SCAs in 10.9 million runners.4
In a Japanese study, the incidence of SCA was 0.46 per 100 000 runners.5 The London marathon had the same incidence,6 while marathons in the Tokyo area had a higher rate of 0.65 per 100 000 runners.7 Death occurred in 1 in 80 000 runners in 26 London marathons. To increase the survival rate of SCA during marathons, the early initiation of cardiopulmonary resuscitation (CPR) and defibrillation is vital.7
The cerebral performance category (CPC) score, which ranges from 1 to 5 (1 indicates normal consciousness and 5 indicates brain death), is generally used to evaluate neurological outcomes post-cardiac arrest.8 Favourable neurological outcomes (CPC 1–2) were found in 51.8% of public access defibrillation (PAD) cases and 25.5% of non-PAD cases in public locations.9 It is rare to provide defibrillation immediately after bystanders witness patients collapse by using automated external defibrillators (AEDs) on-scene. In PAD cases in public locations, in which it took 6 min to administer the first shock, 59.9% of patients experienced a pre-hospital return of spontaneous circulation (ROSC) and 58.8% achieved 1-month survival.9 Providing defibrillation as soon as possible achieves high survival rates; defibrillation within 3–5 min reaches survival rates of 50%–70%.10–12 In our previous study, a 100% survival rate was achieved by the initiation of immediate chest compressions and defibrillation for witnessed SCA cases during road races.7 However, details of the resuscitation and long-term neurological outcomes were not reported.
We created the rapid mobile AED system (RMAEDS) to treat runners as a part of an emergency action plan in which various types of healthcare providers initiate resuscitation in cases of SCA. Through the RMAEDS, healthcare professionals previously provided chest compressions within 0.8 min and defibrillation within 2.2 min.7 The purpose of this study was to examine the emergency response, survival rates and neurological outcomes following SCA in road and long-distance races using the RMAEDS intervention.
Study design and setting
This was a prospective, observational and descriptive study based on records of SCAs occurring during 10 km to full marathons in which the RMAEDS was implemented between 1 February 2007 and 29 February 2020. The Institutional Review Board at Kokushikan University approved the study (#17041). All patients with SCA or their families provided informed consent (written or oral) for the use of their AED ECG data and medical records.
The inclusion criteria were as follows: (1) runners were over the age of 18 and officially registered in the race; (2) runners needed first aid (FA) and basic life support (BLS) on- scene in a marathon or road race; (3) the RMAEDS system was implemented in the race; (4) SCA occurred along the race course or, postrace, near the finish line; (5) SCA was confirmed by the RMAEDS team on-scene; (6) AED-ECG records were available; (7) patients were transported to the hospital by ambulance; (8) all SCA on-scene and hospital data were available and (9) all medical records were confirmed by the medical control director. All of our rescued SCA cases met the inclusion criteria.
Rapid mobile AED system
The purpose of the RMAEDS is to perform prompt bystander CPR (BCPR), and defibrillation when SCAs occur during marathon races. Therefore, we placed RMAEDS teams along the roadside, in locations with the goal of initiating CPR in less than 1 min and administering an AED shock in less than 3 min. They also perform field triage and emergency care for heatstroke, shock, trauma and medical emergencies. As illustrated in figure 1, the RMAEDS consists of (1) a mobile bicycle AED team, (2) a dispatch and medical command control system, (3) an first-aid (FA) station with an AED and healthcare staff, (4) roadside volunteers, (5) a running physician and (6) a mobile on-foot AED team.
Mobile bicycle AED team
As depicted in figure 2, a mobile bicycle AED team, carrying first responder kits and AEDs, covers every 1.5–2.0 km of the course and consists of nationally registered licensed paramedics and trainees. They have trained in high-performance CPR as paramedics at a large private 4-year paramedic university in Japan. The teams can travel 425 m per min; thus, locating a team every 1.5 km makes AED shock delivery possible in less than 3 min.13
Dispatch and medical command control
At the command control centre, two or three paramedics act as dispatchers. One works as the command control officer, and marathon medical support is organised by a course coordinator. A physician acts as the medical director (MD), who is responsible for medical oversight via phone or on-site. In cases of cardiac arrest, the command control officer collects information from bystanders who witness the collapse. Real-time locations are tracked by dispatchers through mobile phones using the global positioning system (GPS), basic telecommunication and advanced video-based communication. The MD provides advice directly or via phone in case of an emergency. If an incident occurs near the finish line, the paramedics and the MD at the command control centre can provide BLS and advanced life support (ALS) by responding on foot. Through bystander interviews, medical records and event audits, they create timeline data and medical records.
First aid station with healthcare providers
Throughout the course, FA stations are positioned every 5–10 km. At the FA stations, there are physicians, nurses and athletic trainers. The stations have BLS and ALS kits, containing medications, intubation kits and AEDs. Healthcare professionals initiate FA and ALS for patients who enter on their own or who are brought in by other RMAEDS teams or race personnel.
Race volunteers act as first responders and they were placed every 200–300 m along the roadside. They are non-medical personnel trained to perform FA and BCPR. In the case of an emergency, they call the headquarters to relay the location and situation, which activates an RMAEDS team.
Equipped with GPS phones, running physicians registered as sports medicine doctors run side-by-side with participants during the race. They are certified in FA and BLS. They carry pocket CPR masks to start CPR but not AEDs. They evaluate a patient’s condition and notify the dispatch centre to activate an RMAEDS team. Although the number of running physicians varied by race, running physicians are responsible for covering 10 km distances along the course, and they start at 2 min intervals.
Mobile on-foot AED team
Pairs of paramedic trainees who are certified Heartsaver-Japan CPR providers can perform FA and high-performance CPR and apply an AED. They can travel 232 m per min. On the first half of the race course, they are placed at 1000 m intervals.13 On the last half of the race course, they are placed every 800 m.13 The dispatchers or roadside volunteers activate this team. They run to the scene, and their placement allows them to initiate CPR or FA within 3 min.
Cardiac arrest response
In the case of cardiac arrest, race volunteers who witness a collapse call dispatchers. Then, the dispatch unit activates an RMAEDS team. A paramedic in the dispatch unit contacts at least two of the nearest mobile bicycle AED teams. Mobile on-foot AED teams initiate a smartphone and SNS application. To identify the patient’s location, the mobile bicycle AED teams, running physicians and mobile on-foot AED teams use GPS tracking telecommunication. The bicycle and mobile on-foot AED teams that receive the call dispatch to the scene. The teams are positioned to quickly initiate FA, BLS and defibrillation. However, multiple individuals with different roles, such as paramedics, paramedic students, running physicians, physicians from FA stations, volunteers and runners, can be first responders. Therefore, in this study, the person who reached the patient first and initiated chest compressions is considered the first responder. The mobile AED teams cannot transport patients, so they call an ambulance if the patient needs further assistance. Until the patient recovers or a fire-based ambulance crew arrives, the team remains at the location (figure 3).
We measured the following variables to evaluate the clinical effectiveness of the RMAEDS in marathons and road races: (1) patient and marathon data, (2) SCA interventions, (3) EMS-related data and 4) SCA-related outcomes.
Marathon event variables
These variables include the race distance and incidence ratio.
The age and sex of patients were obtained retrospectively from race organisers.
SCA intervention measurements
Cardiac arrest variables
These variables include witnessed or unwitnessed, time of cardiac arrest, cardiogenic or non-cardiogenic arrest and recognition of gasping.
The type of initial contactor, BCPR provider details, and the time between the collapse and BCPR initiation were examined.
These factors are the resuscitation sequence times, nature of the first responder, AED operator, interval between the collapse and the AED powering on, duration between the collapse and shock delivery, first narrow QRS waveform appearance and the initial ECG waveform (pulseless ventricular tachycardia (VT), ventricular fibrillation (VF), pulseless electrically active (PEA), asystole and other) by AED ECG record.
These variables include the EMS arrival time, initial ECG waveform on EMS arrival, documented level of consciousness (ie, alert, responsive to verbal stimulus, responsive to pain, unresponsive) in the field and the interval between the collapse and hospital arrival.
SCA-related outcome measurements
Data on achieving field ROSC, in-hospital ROSC, long-term neurological outcomes (cerebral performance category (CPC) score at 1 month and 1 year) and hospital admission length were assessed. The proportion of favourable neurological outcomes (CPC1-2 scores), as defined in table 1, was assessed.8
Primary endpoints were favourable neurological outcomes (CPC1-2) at 1-month and 1-year post-SCA.
Factors related to the field ROSC and resuscitation sequence times were assessed.
These characteristics include the initial ECG waveform (AED) classification, resuscitation sequence times according to the initial ECG rhythm, AED performance and differences in the initial ECG rhythm.
Cause of the SCA
The aetiology of the SCA was assessed.
Effects of early BCPR and AED treatment
Neurological outcomes following BCPR performed within and after 1 min and AED treatment within and after 3 min were compared: (1) 22 cases in which BCPR was initiated within 1 min and AED shock was delivered within 3 min (BCPR ≤1min +AED<3 min), (2) 7 cases in which BCPR was initiated within 1 min and AED shock was delivered after 3 min (BCPR ≤1min +AED≥3 min), (3) 3 cases in which BCPR was initiated between 1 min and 3 min and AED was delivered after 3 min (1 min＜BCPR ≤3min +AED≥3 min), (4) 7 cases in which BCPR was initiated within 1 min without AED use (BCPR ≤1 min without AED) and (5) 3 cases in which BCPR was initiated between 1 min and 3 min without AED use (1 min＜BCPR ≤3 min without AED).
Continuous variables are expressed as mean (SD) or median (IQR) in the background characteristics. Categorical variables are presented as the number of cases (%). All analyses were performed using JMP Pro V.13.0.0 (SAS Institute, Cary, North Carolina, USA).
Gasping: Immediately after an SCA, there may be seizure-like activity and difficulty breathing. There is more respiratory movement immediately after an SCA; breathing becomes more agonal as time elapses. We defined gasping as ongoing chest wall movement and respiratory-like movements.
First Responder: Due to the nature of the resuscitation sequence, multiple individuals with different roles delivered help. In this paper, we defined a first responder as a person who performed over 50% of the chest compressions in the first 2 min.
A total of 42 SCA cases among 3 214 701 runners who participated in 334 marathons and road races met inclusion criteria and were eligible for study. The SCA incidence rate was 1.31 per 100 000 runners; 25 SCAs (60%) occurred in long-distance marathons, 14 SCAs (33%) in half marathons and 3 SCAs (7%) in races less than 10 km (table 2). The course was divided into four segments between the start line and finish line: (i) the first half of the race, (ii) the second half of the race up until 1 mile before the finish line, (iii) the last 1 mile before the finish line and (iv) at or past the finish line; 23.8% of SCAs occurred in the last 1 mile before the finish line (figure 4).
In marathons, the majority of SCAs occurred in men (90%; table 2). Regarding the age distribution, SCAs occurred most frequently in runners in their 30s.
Primary outcomes: neurological outcomes
A CPC1-2 at 1 month and 1 year was achieved in 93% of the 42 cases, and all survivors with CPC1-2 at 1 month maintained their CPC1-2 at 1 year (table 2). All patients who achieved ROSC had CPC1-2 at both 1 month and 1 year. Moreover, no survivors had a status worse than CPC1-2. The median hospitalisation was 5 days. The findings indicate that early BCPR and AED treatment was beneficial. A CPC1-2 at 1 month and 1 year was achieved in 95.5% of the BCPR ≤1min+AED<3 min group and 100% of the 1 min＜BCPR ≤3min+AED<3 min group (table 3).
While the field ROSC was achieved 90% of the time, there were two unwitnessed cases and one refractory VF with severe heat stroke case that did not achieve ROSC. The field ROSC was achieved in 96.9% of runners who presented VF and 88% of runners who presented PEA on the initial ECG rhythm (table 4). In cases in which AED was delivered within 3 min, the field ROSC rate was 90.9% (20/22 cases) for the BCPR ≤1min +AED<3 min group, whereas the field ROSC rate was 100% (3/3 cases) for the 1 min＜BCPR ≤3min +AED<3 min group (table 3).
The initial CPR and AEDs were applied to collapsed patients by paramedics or paramedic students (n=21), running physicians or physicians from first aid stations (n=14), volunteers (n=6) and a non-RMAEDS healthcare provider who participated in the race as a runner (n=1). The paramedics and paramedic students arrived at the scene and applied the AED earlier than other responders, and ROSC was achieved earlier, but there was no difference in outcomes. BCPR was delivered within 1 min in 85.7% of all cases. AED was delivered within 3 min in 68.8% of all cases. In one case in the BCPR ≤1min +AED<3 min group, it took 608 s to apply the AED, so the median of the interval between collapse to first narrow QRS appearance showed longer than it is in the 1 min＜BCPR ≤3min +AED<3 min group (table 3).
Initial ECG waveform (AED) classification
In the initial ECG rhythm on AED application as shown in table 4, VF was found in 33 patients (79%). An AED shock was delivered in 32 of 33 shockable cases. One refractory VF case was administered a shock within 1 min, which become PEA. Seven witnessed PEA cases recovered sinus rhythm by CPR only, but one asystole and one unwitnessed PEA case were not resuscitated. In total, 95% of all SCA cases had a witnessed collapse. Recognition of gasping was found in 90.5% of the cases (38/42 cases). Gasping was recognised in 97% of the VF/VT cases (32/33 cases) and 72% of the PEA cases (6/8 cases; table 4).
Resuscitation sequence times according to the initial ECG rhythm
As presented in table 4, the resuscitation sequence times and outcomes sorted by the initial ECG rhythm were compared. The RMAEDS teams successfully contacted runners within 1 min of collapse and initiated CPR promptly in the VF/VT and PEA cases (VF/VT: 1 (0.55, 2.5) min, PEA: 0.9 (0.42, 2.5) min). BCPR was performed in all 42 cases, and BCPR was initiated within 1 min of the witnessed SCA in both the VF/VT and PEA cases (VF/VT: 0.5 (0.5, 1) min, PEA: 0.75 (0.5, 1.38) min). According to AED records, the interval between turning on the AED to delivering the shock was 26.5 (21, 36) s, which resulted in the appearance of the first narrow QRS in 166 (95, 267) s. Sinus rhythm returned in 93.9% of VF/VT and 87.5% of PEA cases (table 4).
Cause of the SCA
Of the 42 cases, three patients died. In the first case, an autopsy diagnosed a left ventricular outflow tract obstruction associated with hypertrophic cardiomyopathy (HCM). In the other two cases, there was no response to progressive ALS in the field, and no autopsy was performed. In the second case, CPR was initiated early but failed to obtain ROSC due to refractory VF, and the patient fell into pulseless electrical activity (PEA) after defibrillation. We consider this case likely due to exertional heat stroke because it was extremely hot on that day, and the patient’s body temperature was high. Percutaneous cardiopulmonary support was also performed but was unsuccessful. The third case was PEA, and the patient was behind the restroom located along the course and difficult for the team to find. Therefore, the ROSC was delayed and not achieved until hospitalisation; a definitive diagnosis was not found.
Coronary angiography was performed in 21 of the 39 surviving patients to investigate the cause of SCA. In 17 of these cases, a responsible coronary lesion, such as vessel occlusion, stenosis, or coronary spasm, was found. In the remaining four cases, cardiac catheterisation did not reveal a responsible lesion, and cardiomyopathy could not be ruled out because echocardiography was not performed. Follow-up ECGs and specific elevations in myocardial enzymes were also absent. Thus, the final diagnosis was unknown.
In the remaining 18 cases, cardiac catheterisation were not performed. No abnormalities other than a specific elevation of myocardial enzymes, possibly due to chest compressions and defibrillation, were found, and follow-up ECGs indicated no abnormalities. The absence of cardiac conditions known to cause SCA in athletes such as HCM, arrhythmogenic right ventricular cardiomyopathy, long-QT syndrome and Wolff-Parkinson-White syndrome, were excluded based on cardiac imaging and non-specific clinical findings. No genetic testing was conducted, and the final diagnosis was unknown.
We found a 90.4% field ROSC in the 334 races and 42 SCA cases in which the RMAEDS was implemented. The rate of BCPR initiated in SCAs occurring during marathons in this study was 100%, which is approximately two times higher than standard OHCA rates. Based on our results, about 90% of cardiac arrests in marathons are caused by myocardial infarctions, arrhythmias and cardiogenic cardiac arrests due to acute cardiac failure. Cardiogenic cardiac arrests constitute the majority of cardiac arrests that occur during marathons.14–19 More than 80% of the AED-applied cases were shockable; hence, we assume that the majority of the initial ECG rhythms of cardiac arrests in marathons was due to VF.17 Considering that 23.8% of SCA cases occurred in the last mile of the course, the risk of cardiac arrest increases at the end of the race.4 7 20 21 This indicates that FA kits and medical resources should be placed in the last quarter of the course. After the finish line or near the finish line, there were eight cases of collapse; thus, AEDs also must be placed at the finish.
We speculate that a rapid change in the sympathetic nervous system near the finish line dramatically affects myocardial and vital organ perfusion, which triggers a lethal arrhythmia. In a report published in the USA, HCM was the most commonly reported cause of SCA during marathon running.4 A report published in Japan found that hypertension during sports causes the innermost layer of arteries to dilate, which leads to plaque rupture and subsequent platelet aggregation.22 In addition to dehydration caused by exercise, blood clotting and coronary artery occlusion occur, which leads to myocardial ischaemia and VF.22 However, Kuroki et al concluded that coronary plaque rupture and thrombus formation, which are typical acute coronary syndrome mechanisms, were rather less common and that coronary artery stenosis and relative myocardial ischaemia led to the development of lethal arrhythmias (so-called ‘demand ischaemia’).23
Cases in which a defibrillation shock is delivered more than 10 min after a collapse result in about a 10% resuscitation rate.24 Even if a pulse returns, there is a risk of deteriorated cerebral function and a higher possibility of unconsciousness. Roadside medical support that can provide effective on-scene defibrillation is needed for large-scale sports events, including international tournaments. We propose five elements of medical support for sports events: (1) adequate composition and arrangement of rescue and emergency response personnel; (2) efficient arrangement of medical and rescue equipment, including sufficient AEDs; (3) preparation of protocols of communication and command control systems; (4) education of emergency medical staff before events and (5) preparation of event-day protocols for rescuers and EMS activity manuals. On-duty medical personnel should aim to defibrillate within 3 min. There should be clear descriptions of the roles of staff members with activity protocols, and clear communication is necessary for all members.
The interval between turning on the AED and applying the pads is usually nearly 1 min,25 but we found an interval of 26.5 s in our study. We believe this may be due to regular BLS training. In cases in which an AED was used, narrow QRS appeared quickly after the collapse, in 166 s. We believe that maintaining cardiac perfusion through the continuation of chest compressions after early defibrillation by minimising interruptions in CPR may also have contributed to the relatively quick onset of a narrow QRS. This effect contributed to CPC1-2 at 1-month and 1-year postresuscitation in 93% of runners. The field ROSC was achieved in 96.9% of the VF cases, as indicated in the initial ECG by AEDs. Gasping is important to recognise as an initial sign of cardiac arrest, which is significant in alerting bystanders to initiate CPR. In our study, many of the cardiac arrests were witnessed. Gasping was identified in 90.5% of cases, indicating the need to begin CPR quickly. Early defibrillation is the most effective tool to save lives in cases of SCA in sports. We highly recommend placing of AEDs to allow the delivery of AED shock within 3 min of collapse. In our experience, this was accomplished by aiming for BCPR within 1 min and defibrillation within 3 min. If CPR is delayed, cerebral hypoxia can impair the brain function prognosis. Organisers of marathon races must understand the SCA risks during marathons; creating marathon-specific medical support is important for delivering BCPR and defibrillation immediately and achieving favourable survival and neurological outcomes.
This study has four limitations. First, this is a descriptive study that does not indicate the specific cause of the effectiveness of the RMAEDS. Second, long-term outcomes may also be affected by EMS treatment and in-hospital care. Third, since this study used the RMAEDS developed by Kokushikan University, the results may vary if different methods or teams of paramedics and paramedic students are used. Fourth, the size of the event, the number of volunteers, and the number of non-RMAEDS-related medical personnel who participate as runners may change the results if the preparation, such as the number of AEDs placed and volunteer education, differs from event to event.
In this study, we analysed 42 cases of SCA occurring during marathons and distance running events that used the RMAEDS over the past 14 years and evaluated its effectiveness. We found that quick responses after SCA occurrence in marathons led to favourable survival and neurological outcomes. Locating medical personnel and resources appropriately with the goal to initiate BCPR within 1 min and deliver defibrillation within 3 min when SCA occurs will achieve favourable field ROSC rates and neurological outcomes. Healthcare providers covering marathons must be continuously trained in high-quality CPR, how to recognise SCA and initiate BLS at an early stage.
Data availability statement
No data are available.
Patient consent for publication
We would like to thank the members of the Graduate School of the Emergency Medical System and Research Institute of Disaster Management and EMS at Kokushikan University. We would also like to thank all the Kokushikan University alumni paramedics for their consistent efforts at marathon races.
HT and ST contributed equally.
Contributors Conceptualisation: HTanaka. Formal analysis: RS, HTakyu. Investigation: HTanaka, TK, ST, RS, HTakahashi, ES, THara, YT. Methodology: HTanaka, TK. Project administration: HTanaka. Supervision: HTakyu. Visualisation: HTanaka, ST. Writing—Original draft: HTanaka, ST. Writing—Review and editing: HTanaka, ST. Guarantor: HTanaka.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.