Article Text

Research paper
Neck weakness is a potent prognostic factor in sporadic amyotrophic lateral sclerosis patients
  1. Ryoichi Nakamura1,
  2. Naoki Atsuta1,
  3. Hazuki Watanabe1,
  4. Akihiro Hirakawa2,
  5. Hirohisa Watanabe1,
  6. Mizuki Ito1,
  7. Jo Senda1,
  8. Masahisa Katsuno1,
  9. Fumiaki Tanaka3,
  10. Yuishin Izumi4,
  11. Mitsuya Morita5,
  12. Kotaro Ogaki6,
  13. Akira Taniguchi7,
  14. Ikuko Aiba8,
  15. Koichi Mizoguchi9,
  16. Koichi Okamoto10,
  17. Kazuko Hasegawa11,
  18. Masashi Aoki12,
  19. Akihiro Kawata13,
  20. Koji Abe14,
  21. Masaya Oda15,
  22. Masaaki Konagaya16,
  23. Takashi Imai17,
  24. Masanori Nakagawa18,
  25. Shoji Tsuji19,
  26. Ryuji Kaji4,
  27. Imaharu Nakano5,
  28. Gen Sobue1
  1. 1Department of Neurology, Nagoya University Graduate School of Medicine, Nagoya, Japan
  2. 2Biostatistics Section, Center for Advanced Medicine and Clinical Research, Nagoya University Graduate School of Medicine, Nagoya, Japan
  3. 3Department of Neurology, Yokohama City University Graduate School of Medicine, Yokohama, Japan
  4. 4Department of Clinical Neuroscience, Institute of Health Biosciences, University of Tokushima Graduate School, Tokushima, Japan
  5. 5Division of Neurology, Department of Internal Medicine, Jichi Medical University, Tochigi, Japan
  6. 6Department of Neurology, Juntendo University School of Medicine, Tokyo, Japan
  7. 7Department of Neurology, Mie University Graduate School of Medicine, Tsu, Japan
  8. 8Department of Neurology, National Hospital Organization Higashinagoya National Hospital, Nagoya, Japan
  9. 9Department of Neurology, Shizuoka Institute of Epilepsy and Neurological Disorders, Shizuoka, Japan
  10. 10Department of Neurology, Gunma University Graduate School of Medicine, Gunma, Japan
  11. 11Division of Neurology, National Hospital Organization, Sagamihara National Hospital, Sagamihara, Japan
  12. 12Department of Neurology, Tohoku University School of Medicine, Sendai, Japan
  13. 13Department of Neurology, Tokyo Metropolitan Neurological Hospital, Tokyo, Japan
  14. 14Department of Neurology, Okayama University Graduate School of Medicine, Okayama, Japan
  15. 15Department of Neurology, Vihara Hananosato Hospital, Miyoshi, Japan
  16. 16Department of Neurology, National Hospital Organization, Suzuka National Hospital, Suzuka, Japan
  17. 17Division of Neurology, National Hospital Organization, Miyagi National Hospital, Miyagi, Japan
  18. 18Director of North Medical Center, Kyoto Prefectural University of Medicine, Kyoto, Japan
  19. 19Department of Neurology, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan
  1. Correspondence to Professor Gen Sobue, Department of Neurology, Nagoya University Graduate School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466–8550 Japan; sobueg{at}med.nagoya-u.ac.jp

Abstract

Objective To clarify the emergence of muscle weakness in regions of the body that affect survival, and deterioration in activities of daily living (ADL) in amyotrophic lateral sclerosis (ALS) patients.

Methods We conducted a multicentre-based prospective cohort study of patients with ALS. We enrolled 401 sporadic patients with ALS. Death or the introduction of invasive ventilation was defined as the primary endpoint, and the time to five clinical markers of ADL deterioration associated with bulbar paralysis or limb weakness were defined as ADL milestones. Muscle weakness was assessed in the neck flexor muscles; the bilateral abductors of the shoulders; the bilateral wrist extensor muscles; the bilateral flexor muscles of the hips; and the bilateral ankle dorsiflexion muscles. We performed Cox proportional hazards regression analyses for the primary endpoint and the five ADL milestones, adjusting for known covariate prognostic factors for ALS.

Results The Medical Research Council (MRC) score for the neck flexors was the most significant prognostic factor for the primary endpoint (HR 0.74, p<0.001), loss of speech (HR 0.66, p<0.001), and loss of swallowing function (HR 0.73, p<0.001), and was one of the significant prognostic factors for loss of upper limb function, difficulty turning in bed, and loss of walking ability (p=0.001, 0.002, and 0.008, respectively). The MRC score for the neck flexors was also a significant prognostic factor for covariates of the previously reported prognostic factors.

Conclusions Neck weakness is an independent prognostic factor for survival and deterioration in ADL in Patients with ALS.

  • ALS
  • Neuroepidemiology
  • Clinical Neurology
View Full Text

Statistics from Altmetric.com

Introduction

Amyotrophic lateral sclerosis (ALS) is an adult-onset neurodegenerative disease characterised by progressive upper and lower motor neuron loss, which leads to limb and bulbar paralysis and respiratory failure.1 Symptoms develop at a progressive rate, and the median survival time from disease onset is 2–4 years.2–4 However, patients with ALS show extensive variability in clinical courses, with durations ranging from a few months to more than 10 years. Furthermore, major symptoms that differentially affect activities of daily living (ADL) and prognosis also show variability among patients with different disease forms.5 A better understanding of the factors influencing deterioration in ADL and prognosis would help physicians and patients determine whether and when to introduce non-invasive positive pressure ventilation, tube feeding, tracheostomy and artificial ventilation, and would lead to effective stratification strategies in clinical trials. Several studies have shown that age,6–10 bulbar symptom onset,6 ,7 respiratory function,3 ,8 ,11 ,12 time from symptom onset to diagnosis,2 ,6 ,10 ,13 ,14 functional score2 ,14 and rate of disease progression2 ,15–17 are predictors of survival. Muscle weakness in particular regions of the body affect the prognosis of ALS, although it has not been sufficiently determined which regions are most predictive.18 To investigate the longitudinal course of patients with ALS and clarify the emergence of muscle weakness, which affects deterioration in ADL and ALS prognosis, we conducted a prospective, multicentre study.

Methods

Patient registry and follow-up system

We constructed a multicentre registration and follow-up system called the Japanese Consortium for Amyotrophic Lateral Sclerosis research (JaCALS), which consists of 21 neurology facilities in Japan. Patients with ALS diagnosed in these facilities were consecutively registered with written informed consent. The ethics committees of all the participating institutions approved the study. Full clinical examinations were conducted at registration by neurologists in each of the respective institutions. Muscle strength was manually tested and scored with the scale of the Medical Research Council (six points, range: 0–5)19 in nine muscle groups as follows: neck flexors; bilateral abductors of shoulders as representatives of proximal upper extremity muscles; wrist extensors muscles as representatives of distal upper extremity muscles; bilateral flexors of hips as representatives of proximal lower extremity muscles; and ankle dorsiflexion muscles as representatives of distal lower extremity muscles. All manual muscle testing was performed with standard positioning and procedures by certified neurologists.20 The MRC score of the neck flexors was determined with the patient in the supine position. We confirmed the inter-rater reliability of the manual muscle testing method employed in this study using 23 patients with neuromuscular disease. The values of the kappa statistics of each muscle ranged from 0.65 to 0.93. To standardise the procedures and the examinations, the three organising doctors (NA, RN, HaW) visited each participating facility and ascertained the evaluation methods for this study.

Disease onset was defined as when the patients became initially aware of muscle weakness or impairment of swallowing, speech, or respiration. We enrolled patients who fulfilled the revised El Escorial criteria.21 The diagnostic accuracy of the enrolled patients was then assessed by members of the steering committee of the JaCALS. Included patients were then followed-up with telephone surveys conducted by clinical research coordinators (CRC) or with examinations by neurologists every 3 months, and the degree of deterioration in ADL was determined at each time point. We employed the Japanese version of the ALSFRS-R as a scale for ADL, which was validated by Ohashi et al, using a telephone survey system.22 We confirmed the reliability of the telephone survey system for the Japanese version of the ALSFRS-R previously,23 and the English version of the telephone survey system has been confirmed in several previous studies.24–26 Prior to the study, we informed and trained the CRCs of the study plan, procedures for the telephone survey, ethical issues relevant to the study, and requisite considerations for patients with ALS and caregivers, and then provided them with a general knowledge of ALS.

We defined a primary endpoint and ADL milestones in the disease course of the patients with ALS and determined their occurrence by telephone survey or examinations by neurologists. The introduction of tracheostomy positive pressure ventilation (TPPV) or death of the patient was defined as the primary endpoint, and TPPV-free survival was defined as survival. Loss of speech function, loss of swallowing function, loss of upper limb function, difficulty turning in bed, and loss of walking ability were set as ADL milestones. The time at which each ADL milestone occurred was defined as follows: loss of speech function was determined to have occurred when non-vocal communication became needed; loss of swallowing function was determined to have occurred when parenteral or enteral feeding became needed exclusively; loss of upper limb function occurred when the patient needed to be fed and became unable to grip a pen; difficulty turning in bed occurred when the patient became unable to turn in bed alone; loss of walking ability occurred when the patient became unable to walk without assistance.

Patients

A total of 520 patients with ALS were initially registered in the JaCALS from January 2006 to June 2011. We excluded 26 patients with known gene mutations: 17 patients with SOD-1 mutations, two patients with TDP-43 mutations, two patients with FUS/TLS mutations, three patients with angiogenin mutations, and two patients with C9ORF72 repeat expansions. We also excluded 13 patients with family histories of ALS and 40 patients who were categorised as clinically possible or suspected according to the revised El Escorial criteria. An additional 20 patients for whom we could not obtain follow-up information to their refusal to participate in the telephone survey were also excluded. Twenty patients were excluded due to invalid data. The study finally included 401 sporadic patients with ALS diagnosed as clinically definite, probable, or probable laboratory-supported. Of these, 382 patients were followed for more than a year or died within a year of registration, and 19 patients were censored within a year from registration. Eleven patients declined the telephone survey during the course of the study, and we lost contact with eight patients during the survey.

Statistical analysis

The clinical data of the registered patients were anonymised in each participating facility of the JaCALS and assigned unique patient numbers. The data were then sent to the clinical data centre located at the Nagoya University Graduate School of Medicine and inputted into the JaCALS database.

We performed Cox proportional hazards regression analyses for the time of registration to the primary endpoint or onset of each ADL milestone to evaluate the impact of muscle weakness on the time to the primary endpoint and each decline in ADL. Specifically, for the primary endpoint and each ADL, we evaluated the HR for the MRC scores in nine muscle groups (ie, neck flexors, left and right abductors of shoulders, wrist extensor muscles, flexors of hips and ankle dorsiflexion muscles) at registration, identifying the muscles groups associated with the primary endpoint and five common ADL milestones. Additionally, we examined the HR for each muscle group after adjusting for known prognostic factors as follows: age at registration,6–10 gender (male vs female),6 ,27 disease duration,2 ,6 ,10 ,13 ,14 percent vital capacity (%VC),3 ,8 ,11 ,12 ALSFRS-R score,14 riluzole use (yes vs no),28 bulbar symptom,6 ,7 and classification according to the revised El Escorial criteria (definite vs probable or probable laboratory-supported).7 ,8 ,10 ,14 We compared the time from registration to the primary endpoint or each of the previously defined ADL milestones in the patients divided by their degree of muscle weakness using the Kaplan–Meier method. The log-rank test was used to test the null hypothesis that all the Kaplan–Meier curves were equal. A two-sided p<0.05 was considered statistically significant. All statistical analyses were conducted using the PASW V.18.0 program (SPSS Inc, Chicago, Illinois, USA).

Results

Demographic characteristics of the registered patients

The patient sample comprised 244 men and 157 women. The median age at disease onset was 62.2 years (IQR: 53.5–68.5 years), and the mean follow-up period was 2.1±1.5 years. The follow-up rate at 1 year after registration was 95.3%. As initial symptoms, 47.4% of the patients showed upper limb weakness, 31.4% lower limb weakness, 22.9% dysarthria, 5.5% dysphagia and 2.0% cervical weakness. At registration, the median score on the ALSFRS-R was 38 (IQR: 32–42). (see online supplementary table S1).

Identification of weakened muscle groups that affect survival and progression to the ADL milestone

Cox proportional hazard regression analyses for the primary endpoint and the ADL milestones

Table 1 shows the results of Cox proportional hazard regression analyses for the primary endpoint and the five ADL milestones, including the MRC scores of the nine muscle groups. The MRC score for the neck flexors was the most significant negative prognostic factor for the primary endpoint, loss of speech, and loss of swallowing function (HR 0.74, p<0.001, HR 0.66, p<0.001, HR 0.73, p<0.001, respectively). For the loss of upper limb function, difficulty turning in bed and loss of walking ability, the MRC score for the neck flexors was a significant negative prognostic factor (HR 0.77, p=0.001, HR 0.77, p=0.002, and HR 0.80, p=0.008, respectively). Whereas, the MRC score for the left wrist extensors was a significant positive prognostic factor for the primary endpoint and each ADL milestone except for difficulty turning in bed.

Table 1

Multivariate Cox regression analyses for the primary endpoint and each activity of daily living milestone using the MRC score of each muscle group at registration

According to table 1, the MRC score for the neck flexors was commonly identified as a possible prognostic factor for the primary endpoint and the five ADL milestones. We further examined its impact after adjusting for the other established or potential risk factors, that is, age at registration, gender, disease duration from onset to registration, percent vital capacity (%VC) at registration, ALSFRS-R score at registration, classification according to revised El Escorial criteria, riluzole use and bulbar symptom at registration (table 2). As seen in table 2, the MRC score for the neck flexors was an independent and significant prognostic factor for the primary endpoint, loss of speech, loss of swallowing, loss of upper-limb function and difficulty turning in bed in patients with ALS except for loss of walking ability. (p<0.001, p=0.001, p=0.003, p<0.001, p=0.027, respectively). At registration, there were moderate and significant correlations between the MRC score for the neck flexors and the %VC or the ALSFRS-R score. Pearson's correlation coefficients were 0.367 (p<0.001) and 0.496 (p<0.001), respectively.

Table 2

Multivariate Cox regression analyses with the adjustments of the covariates that we selected for the primary endpoint and each activity of daily living milestone using known predictive factors

Differences in survival time and time to ADL milestones in patients in terms of the MRC score grade for the neck flexors

We divided the registered patients into four categories according to their MRC score for the neck flexors (ie, 5, 4, 3 and ≤2). Figure 1 shows the Kaplan–Meier curves for the four categories for the primary endpoint and each ADL milestone. All the differences between the curves were significant according to a log-rank test (p<0.001).

Figure 1

Kaplan–Meier curves according to the MRC score for the neck flexors. The Kaplan–Meier curves for the primary endpoint and each activity of daily living milestone among four categories divided according to the MRC score for the neck flexors were compared by the log-rank test. Curves are shown for the MRC score 5 (blue), the MRC score 4 (green), the MRC score 3 (orange), and the MRC score ≤2 (purple) groups. All the differences of the curves were significant (p<0.001).

Discussion

In a prospective and multicentre cohort study, we identified that weakness of the neck flexors is a potent factor for the prediction of survival and for the deterioration of ADL, such as speech, swallowing, upper limb function, turning in bed, and walking, in sporadic patients with ALS.

The neck flexors consist of the sternocleidomastoid muscle (SCM), the platysma muscle, hyoid muscle, longus capitis muscle, longus colli and scalenus. These muscles are innervated by motor neurons in the cervical cord (C1–8) and accessory nerve nuclei,29 ,30 primarily the C2–4 segments. By contrast, respiratory muscles consist of the diaphragm and the internal and external intercostals muscles, which are innervated by motor neurons of the upper cervical cord (C3–5) and thoracic cord (Th1–Th12), respectively.30 Thus, the muscles for neck flexion and those for respiration partially share spinal segments of the motor neuron column for their motor innervations. Furthermore, significant correlations are present between compound muscle action potentials of the SCM and those of the diaphragm in patients with ALS,31 suggesting that neck muscle weakness is correlated with weakness of the diaphragm to some extent in ALS. Because the main cause of death in patients with ALS is respiratory insufficiency, it is reasonable that neck flexor weakness was associated with respiratory impairments and, eventually, survival time. The motor response amplitude of the phrenic nerve motor neurons which are located in the C3–5 segments has been shown to be a significant prognostic factor for survival in patients with ALS.32 This supports our findings.

Why then is weakness of the neck flexors a determinant factor for the deterioration of ADL for speech, swallowing, upper limb function, truncal turning and walking ability? Recently, some studies have suggested that the degeneration of motor neurons is initially a focal process in ALS that later spreads contiguously throughout the three-dimensional anatomy of connected or neighbouring neurons.33–36 Dysfunction of speech and swallowing involves the impairment of motor neurons relayed via the glossopharyngeal, vagus, accessory and hypoglossal nerves to the medulla oblongata.30 The medulla oblongata and cervical cord motor neurons innervating the neck flexion muscles are anatomically different in their three-dimensional layering, while these two groups of neurons are rather contiguously located. Thus, it may be speculated that if the contiguous spreading of motor neuron degeneration occurs according to the local spreading hypothesis, neck flexion impairment may eventually affect speech and swallowing functions. Furthermore, motor neurons for the neck flexion muscles, which are located in the C1–8 segments,29 ,30 are also contiguous or overlapping with those for the upper limb muscles in the C5–Th1 segments, particularly the proximal upper limb muscles.29 ,30 Neck flexion and upper limb function may be correlated with disease progression through the local spreading view of motor neuron degeneration. Truncal turning and walking require not only lower limb muscle activities but also power in a broad area of the chest, abdominal and back muscles, which are innervated by the cervical to lumbar cord.37–39 Therefore, propagation of weakness from the cervical and lumbar areas may affect truncal turning or walking. We need, however, further investigations to demonstrate the underlying mechanisms of the correlation between the neck muscles and other muscles of the body that together determine ADL.

In this study, the MRC score for the left wrist extensors shows a positive prognostic factor for the primary endpoint and some ADL milestones, the reason for which might be that the weakness of the distal muscle in the non-dominant arm was least relevant to survival, or ADL declines so that it was shown to be a positive factor in the multivariate analyses.

A number of studies have demonstrated survival curves for patients with ALS and some factors that influence these survival curves.18 The majority of these studies have found that older age is a strong risk factor for shorter survival in patients with ALS,6–10 and the onset of bulbar symptoms is associated with a worse prognosis than the onset of spinal symptoms.6 ,7 Several studies have found that a longer diagnostic delay correlates with a better prognosis,2 ,6 ,10 ,13 ,14 and that a lower %VC or a percent forced vital capacity (%FVC) is correlated with shorter survival for patients with ALS.3 ,8 ,11 ,12 The progression rate of the ALSFRS-R at the time of diagnosis was also related to ALS prognosis.17 Neck flexor weakness has not been listed as a prognostic factor for patients with ALS, and most of these studies evaluated survival alone as an endpoint, and did not determine the onset of loss of speech, swallowing, limb and truncal function. In this study, we showed that neck flexor weakness was not only a novel prognostic factor for survival but also a significant prognostic marker for non-survival events related to ADL decline for patients with ALS.

In the course of ALS, patients must make some difficult decisions, including the use of gastrostomy for tube feeding, the use of assisted ventilation, and end-of-life planning, which require the support of the attending physician and a multidisciplinary team. All patients with ALS should be provided with sufficient information concerning these interventions and given sufficient opportunity for the careful consideration of their decision. In the medical, nursing and social care of patients with ALS, simple and robust indicators for predicting the status of each patient for several months or a year after diagnosis are necessary for patient management. Medical staff and caregivers need to have a predictor of the patient's status in the near future, including survival prognosis and also estimates for the loss of speech, swallowing, limb and truncal function. Our findings may contribute to such prediction.

The course of ALS is highly variable between patients,5 which is one of the major factors limiting the power of ALS clinical trials.40 ,41 Therefore, robust stratification factors that could divide ALS patient groups depending upon prognosis are needed for designing trials. Compared with known prognostic factors for patients with ALS, such as age, duration from onset to registration, ALSFRS-R at registration, and presence of bulbar symptom, weakness of the neck flexors was a potent and independent prognostic factor. Thus, the MRC score for the neck flexors might be used for stratification factor in a future clinical trial.

Neck extensor muscle weakness with head drop as an early symptom has been reported in a few patients with ALS.42 ,43 However, Katz et al44 wrote that neck flexor weakness is typically observed. We assert that neck flexor weakness is commonly observed in patients with ALS, and is useful for the prediction of prognosis.

The limitations of this study are as follows: registered patients were followed-up by telephone survey, and we did not examine longitudinal changes in the strength of multiple muscles. As we demonstrated, the relationship between the involved muscle groups and survival prognosis and estimates of ADL deterioration would offer insights into the modalities of progression in patients with ALS. However, to examine the pattern of spread more precisely, a cohort study that observes longitudinal changes in the strength of muscle groups and extensions of muscle weakness will be required.

This study analysed a multicentre cohort of patients with ALS in a single nation, Japan. Although the clinical profiles of ALS are broadly similar among countries in previous reports, the outcome of our study would be better confirmed in cohorts of patients with ALS in multiple countries.

In conclusion, we showed that neck weakness is an independent prognostic factor for survival and deterioration in ADL in patients with ALS. We hope that our report will be helpful for clinicians who want to provide medical, social and nursing care to patients with ALS with proper timing, and to researchers as they plan clinical trials for ALS.

Acknowledgments

We thank all Patients with ALS who participated in this study. We also thank all participating doctors and staff of JaCALS.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • RN and NA contributed equally.

  • Collaborators JaCALS members included Drs Tatsuhiko Yuasa (Kamagaya General Hospital); Takahiro Kano (Hokkaido University); Hidenao Sasaki (Hokkaido University); Masaaki Kato (Tohoku University); Tomohiko Ishihara (Niigata University); Masatoyo Nishizawa (Niigata University); Masaki Ikeda (Gunma University); Kazumoto Shibuya (Chiba University); Satoshi Kuwabara (Chiba University); Hideaki Hayashi, MD (Tokyo Metropolitan Neurological Hospital); Yuji Takahashi (The University of Tokyo); Hiroyuki Tomiyama (Juntendo University); Nobutaka Hattori (Juntendo University); Hitoshi Aizawa (Tokyo National Hospital); Osamu Kano (Toho University Omori Medical Center); Yasuo Iwasaki (Toho University Omori Medical Center); Takamura Nagasaka (University of Yamanashi); Yoshihisa Takiyama (University of Yamanashi); Motoko Sakai (Suzuka National Hospital); Kensuke Shiga (Kyoto Prefectural University); Hirofumi Yamashita (Kyoto University); Ryosuke Takahashi (Kyoto University); Takuji Fujita (Takumikai Neurology Clinic); Toru Yamashita (Okayama University); Masanori Hiji (Vihara Hananosato Hospital); Yasuhiro Watanabe (Tottori University); Kenji Nakashima (Tottori University); Shintaro Hayashi (Kyushu University); and Jun-ichi Kira (Kyushu University).

  • Contributors RN: drafting/revising the manuscript, acquisition, analysis and interpretation of the data, statistical analysis, and research project execution; NA: drafting/revising the manuscript, acquisition, analysis and interpretation of the data, statistical analysis, research project execution, and study design and concept; HaW: acquisition and interpretation of the data, and research project execution; AH: manuscript review and critique, analysis and interpretation of the data, and statistical analysis design; HiW and MI: acquisition and interpretation of the data, research project execution, and study design and concept; JS, MKatsu, FT, YuI, KotO, AT, IA, KM, KoiO, KH, AK, KA, MO, MKo, TaI, MNa, and IN: revising the manuscript, and acquisition and interpretation of the data; MM, MA, ST, and RK: revising the manuscript, acquisition and interpretation of the data, and the members of the JaCALS steering committee; GS: research project organisation and execution, drafting/revising the manuscript, interpretation of the data, study design and concept, and the member of the JaCALS steering committee. TY: the member of the JaCALS steering committee; TK, HS, MKato, ToI, MNi, MI, KaS, SK, HH, YuT, HT, NH, HA, OK, YaI, TN, YaT, MS, KeS, HY, RT, TF, TY, MH, YW, KN, SH, and JK: acquisition of the data.

  • Funding This study was supported by Health and Labour Sciences Research grants (Research on intractable diseases H23-015 and H24-012) and grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan (grant number 25461277 ).

  • Competing interests None.

  • Ethics approval The ethics committees of all the participating institutions approved the study.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.