You are here

Article Text

Article menu
Original article
Changes in ventilator strategies and outcomes in preterm infants
  1. Valentina Vendettuoli1,
  2. Roberto Bellù2,
  3. Rinaldo Zanini2,
  4. Fabio Mosca1,
  5. Luigi Gagliardi3,
  6. for the Italian Neonatal Network
  1. 1NICU, Department of Clinical Sciences and Community Health, Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano, Milano, Italy
  2. 2Neonatal Intensive Care Unit, Ospedale ‘A Manzoni’, Lecco, Italy
  3. 3Pediatrics and Neonatology Division, Woman and Child Health Department, Ospedale Versilia, Viareggio, Italy
  1. Correspondence to Dr Valentina Vendettuoli, NICU, Department of Clinical Sciences and Community Health, Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano, Via della Commenda 12, Milano 20122, Italia, valentina.vendettuoli{at}mangiagalli.it

Abstract

Background Although life-saving, intubation and mechanical ventilation can lead to complications including bronchopulmonary dysplasia (BPD). In order to reduce the incidence of BPD, non-invasive ventilation (NIV) is increasingly used.

Objective The aim of our study was to describe changes in ventilator strategies and outcomes between 2006 and 2010 in the Italian Neonatal Network (INN).

Design Multicentre cohort study.

Settings 31 tertiary level neonatal units participating in INN in 2006 and 2010.

Patients 2465 preterm infants 23–30 weeks gestational age (GA) without congenital anomalies.

Main outcomes measures Death, BPD and other variables defined according to Vermont Oxford Network. Logistic regressions, adjusting for confounders and clustering for hospitals, were used.

Results Similar numbers of infants were studied between 2006 and 2010 (1234 in 2006 and 1231 in 2010). The baseline risk of populations studied (GA, birth weight and Vermont Oxford Network Risk-Adjustment score) did not change. After adjusting for confounding variables, infants receiving invasive mechanical ventilation decreased (OR=0.72, 95% CI 0.58 to 0.89) while NIV increased (OR=1.75, 95% CI 1.39 to 2.21); intubation in delivery room decreased (OR=0.64, 95% CI 0.51 to 0.79). Considering outcomes, there was a significant reduction in mortality (OR=0.73, 95% CI 0.55 to 0.96) and in the combined outcome mortality or BPD (OR=0.76, 95% CI 0.62 to 0.94).

Conclusions Despite a stable baseline risk, from 2006 to 2010, we observed a lower level of invasiveness, a reduction of mechanical ventilation and an increase of NIV use, and this was accompanied by a decrease in risk-adjusted mortality and BPD.

  • Epidemiology
  • Ventilator strategies
  • Outcomes
  • Preterm Infants

https://doi.org/10.1136/archdischild-2013-305165

Statistics from Altmetric.com

    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.

    What is already known on this topic

    • Ventilator assistance remains one of the most important and most used procedures in very preterm infants.

    • Different techniques, both invasive (with endotracheal intubation) and non-invasive, are available.

    What this study adds

    • In a large cohort of preterm infants, in last 5 years we observed a reduction of invasive mechanical ventilation and an increase of non-invasive ventilation.

    • Changes in ventilator practices were accompanied by a reduction in both mortality and bronchopulmonary dysplasia.

    Introduction

    Between 2000 and 2009, the rates of mortality and major neonatal morbidities including chronic lung disease1 in survivors decreased for infants with a birth weight (BW) of 501–1500 g. However, pulmonary disorders are responsible of over 50% of neonatal deaths in very preterm infants,2 ,3 and are associated with morbidities that have severe long-term consequences.1 ,2

    Several different methods for providing respiratory support can be applied in the care of preterm infants. In previous decades, it was common to initiate endotracheal intubation and mechanical ventilation in neonates with moderate to severe respiratory distress syndrome (RDS).4 These procedures can be life-saving, but may have adverse effects on the respiratory system.4 Bronchopulmonary dysplasia (BPD) represents one of the most significant pulmonary complications. It is a multifactorial disease, its pathogenesis being linked to immature lung tissue, barotrauma, volutrauma5 and prolonged mechanical ventilation.6

    Nasal continuous positive airway pressure (nCPAP) has been used to wean preterm infants off mechanical ventilation and, more recently, as a primary mode of respiratory support for preterm infants with respiratory insufficiency.7 ,8 Nowadays, improvements in the measurement of volume and flow in modern neonatal ventilators have led to a variety of alternative non-invasive ventilation (NIV) procedures, in addition to the well-known nCPAP.9 NIV refers to any technique that uses constant or variable pressure to provide ventilator support, but without tracheal intubation such as continuous positive airway pressure, nasal intermittent mandatory ventilation (nIMV), nasal intermittent positive pressure ventilation and high flow nasal cannula.

    Currently, NIV is widely used in neonatal intensive care units;9 however, there are few data on how these techniques have been taken up by neonatologists, resulting in actual changes in ventilator strategies in very preterm infants, and if these changes have been associated with changes in respiratory outcomes, such as BPD.5 ,6

    Neonatal networks allow the collection of large multicentre databases and are an ideal tool to answer these scientific questions.10 ,11

    We undertook this study to identify changes in respiratory procedures and outcomes for preterm infants enrolled in the Italian Neonatal Network (INN) 5 years apart (between 2006 and 2010).

    Methods

    A cohort of neonates 23–30 weeks gestational age (GA), born in 2006 and 2010, and assisted in tertiary level neonatal intensive care units adhering to the INN was analysed. INN is a voluntary collaboration of neonatal units in Italy, with the purpose of coordinating national data collection, adhering to the Vermont Oxford Network (VON).

    In 2006, 31 hospitals participated, and in the following years the number of hospitals progressively grew to 86 in 2012; however, for the purpose of this study we only considered infants assisted in the 31 hospitals participating in both 2006 and 2010.

    We excluded from the study infants dying in the delivery room and those with congenital anomalies.

    All the variables were defined according to VON (http://www.vtoxford.org). All data undergo automated data checks for quality, and completeness is verified and certified at the time of submission as part of participation in the VON. Concerning ventilator support, for each infant the database collects data on such procedures as oxygen supplementation; use of nCPAP, and use of nCPAP before intubation; use of high flow nasal cannula; use of nIMV/synchronised intermittent mandatory ventilation; use of conventional ventilation; use of high frequency ventilation; and surfactant administration during initial resuscitation in delivery room or at any time.

    The outcomes analysed for this study are mortality before discharge from the hospital, or BPD, defined according to the VON.12

    Briefly, BPD is defined as being present if the infant received any supplemental oxygen at 36 weeks postmenstrual age; infants discharged or transferred between 34 and 36 weeks were classified based on their oxygen status at discharge (see online supplementary appendix).

    Data analysis

    In order to obtain unbiased estimates, we tried to ensure comparability of populations analysed between years by taking the following steps: first, we restricted the inclusion criteria only to infants admitted to hospitals that participated in the INN in both 2006 and 2010. Second, we compared infants in both years for variables known to influence both ventilation procedures and outcomes, and that could act as confounders such as GA, BW, antenatal steroids, sex, being inborn, mode of delivery, being twins, intubation in delivery room and RDS. We also calculated for each infant the severity of illness by using the Vermont Oxford Network Risk-Adjustment (VON-RA) score, which takes into consideration GA, BW for GA, mode of delivery, multiple pregnancy, 1-min Apgar score, race, being inborn, and sex (coefficients kindly provided by J Horbar and E Edwards, VON), a validated infant severity score.13 ,14

    Finally, we estimated changes in frequencies in procedures (NIV, mechanical ventilation, any respiratory support, etc) and outcomes (death, BPD, death or BPD) between years by using random-effects logistic models which took into account clustering for hospitals. Results are expressed as ORs and 95% CI. Estimates were adjusted for severity of illness of infant and antenatal steroid prophylaxis. Missing data never exceeded 2.4%, except for BPD (6.1%).

    Analyses were performed using STATA (V.11) statistical software.

    Results

    Infants’ characteristics

    The number and the characteristics of infants remained constant between years.

    A total of 2465 infants of 23–30 weeks were cared for at the 31 third level neonatal intensive care units in the 2 years studied (1234 infants in 2006 and 1231 in 2010).

    Table 1 reports the main characteristics of the studied infants in the 2 years considered. The baseline risk, expressed by the VON-RA score, did not change between years (p=0.95), while there was a statistically significant increase in the use of antenatal steroids (p<0.001), in multiple gestation (p=0.01) and caesarean section (p=0.02) from 2006 to 2010. There was also a small but statistically significant increase in the Apgar score at 1 min (p=0.03). We did not find any statistically significant difference in the number of infants who died in the delivery room (10 in 2006 vs 8 in 2010; p=0.64) in the 2 years considered.

    NICU respiratory practices

    While the baseline risk did not change, the ventilator strategies changed considerably over the 5 years examined; the frequency of selected ventilator procedures for 2006 and 2010 is shown in table 2.

    View this table:
    Table 1

    Characteristics of infants in the 2 years considered expressed as means and SD or per cent

    View this table:
    Table 2

    Frequency of selected ventilator procedures and risk difference (RD)

    Our results showed a statistically significant reduction in the rate of delivery room intubation (p<0.001) and a significant increase in the use of nCPAP before intubation (p<0.001). We found a significant increase in surfactant administration during initial resuscitation (p=0.01) while the administration of surfactant at any time did not increase (p=0.92) over the study period. Even after logistic regression the results did not change: surfactant use during initial resuscitation increased (OR=1.34, 95% CI 1.07 to 1.70) whereas surfactant administered at any time did not increase from 2006 to 2010 (OR=1.06 95% CI 0.87 to 1.28).

    From 2006 to 2010, we observed only a very slight reduction in the rate of postnatal steroid (p=0.73), and a significant increase in the number of infants receiving any ventilator support (p=0.003), with a decrease in those receiving mechanical ventilation (p<0.001) and an increase in infants receiving NIV (p<0.001).

    Interestingly, analysing different techniques of invasive mechanical ventilation, we observed a statistically significant increase in high frequency ventilation from 2006 to 2010 (p<0.001), whereas conventional ventilation showed a statistically significant decrease 5 years apart (p<0.001).

    On the other hand, considering different techniques of NIV, we found that the use of all of them was significantly increased from 2006 to 2010: nIMV (p<0.001), nCPAP (p<0.001) and humidified high flow nasal cannula (p<0.001).

    Outcomes

    In table 3, the frequencies of the main outcomes such as mortality, BPD and mortality or BPD are shown. In these 5 years, mortality (p<0.05) and mortality or BPD incidence (p=0.04) had a statistically significant reduction while the incidence of BPD alone was reduced although not reaching statistical significance.

    View this table:
    Table 3

    Frequency of outcomes: BPD, mortality, mortality or BPD

    In table 4 we report the ORs of procedures and outcomes in 2006 versus 2010, adjusting for baseline severity of illness (VON-RA score) and antenatal steroid prophylaxis.

    View this table:
    Table 4

    Multivariate analysis showing changes in several procedures and outcomes in 2010 versus 2006

    After risk-adjustment and multivariate analysis, the results did not change: delivery room intubation (OR=0.64, 95% CI 0.51 to 0.79) and mechanical ventilation (OR=0.72, 95% CI 0.58 to 0.89) decreased, while NIV increased (OR=1.75, 95% CI 1.39 to 2.21). Also, the main outcomes did not change: mortality decreased (OR=0.73, 95% CI 0.55 to 0.96) as well as the composite outcome ‘death or BPD’ (OR=0.76, 95% CI 0.62 to 0.94); also, the incidence of BPD showed a trend towards reduction (OR=0.87, 95% CI 0.68 to 1.12) without reaching statistically significance.

    Discussion

    In the last decades, several technological and clinical advances have been made in neonatal care, which have led to an improvement in neonatal outcomes.

    A recent large cohort study, which collected data from over 350 000 very low BW infants born from 2000 to 2009 and cared for at 669 hospitals in VON, found significant changes in many obstetric, delivery room and neonatal practices.15 The most striking change was the trend towards increased use of non-invasive approaches to respiratory support, such as nasal CPAP, nIMV, nasal intermittent positive pressure ventilation and high flow nasal cannula.15

    Also our cohort study showed from 2006 to 2010 an important shift in obstetric practices and techniques of ventilator support, which paralleled changes in the outcomes of the studied population. In particular, this study shows that in the last 5 years we observed a reduction of mechanical ventilation and an increase of NIV use in infants of 23–30 weeks, and this was accompanied by a decrease in risk-adjusted mortality and combined outcome mortality or BPD.

    Over the 5 years, no changes were noted in infants’ characteristics and in order to eliminate the impact of changes in participating centres, we restricted the analysis to the subset of hospitals that were members of INN throughout the study period.

    As in the VON cohort study,15 we found an increased use in the treatment with antenatal steroids and caesarean section delivery during the study period: the use of antenatal steroids has a strong basis in evidence from randomised controlled trials,16 unlike the increased rate of caesarean deliveries.17

    Few years ago, elective tracheal intubation in the delivery room was considered as the routine approach in managing respiratory failure in very preterm infants. Over recent years, the indications and principles of neonatal resuscitation of these infants have been reviewed.18 As in the VON cohort study,15 we observed a significant reduction in the rate of delivery room intubation, which is probably related to the recently reviewed neonatal resuscitation guidelines.18

    Again, as in the VON cohort study,15 our results show changes in surfactant administration; surfactant treatment during initial resuscitation increased substantially during the study period while the percentage of infants receiving surfactant at any time did not increase. In the past decades, randomised, controlled trials have demonstrated that prophylactic or early surfactant therapy compared with delayed rescue surfactant treatment results in improved outcomes for preterm infants at high risk.19 Whether to treat infants at risk for RDS with CPAP or surfactant therapy immediately after birth is still unclear.20 ,21

    We found that fewer infants are intubated in the delivery room, but those intubated receive surfactant shortly after, and this may be related to the decrease in invasive ventilation. No specific data related to Intubate-SURfactant administration and Extubation (INSURE) or Less Invasive Surfactant Administration (LISA) are reported in VON database, and therefore we cannot assess the frequency of these techniques.

    The use of nCPAP as a primary mode of respiratory support to avoid invasive ventilation and facilitate weaning from the ventilator has become standard practice.7 ,8 Our analysis confirms an increase in the use of nCPAP as a primary mode of respiratory support from 2006 to 2010.

    New methods of non-invasive respiratory support have been developed and the available evidence suggests important physiological advantages of NIV in premature infants.22 A recent survey of chiefs of neonatal units in Italy confirmed that NIV is the first-intention treatment for preterm infants with respiratory diseases.10

    We observed an increase of different modalities of NIV such as heated humidified high flow nasal cannula and nIMV in addition to nCPAP. The use of high flow nasal cannula is an increasingly popular alternative to nCPAP for non-invasive respiratory support of very preterm infants.23 ,24 Data on the efficacy or safety of such cannula in this population are still insufficient;24 our results reinforce the urgent need for research to evaluate their use in newborns.

    The reduction of invasive respiratory support techniques and adverse neonatal outcomes could be due to both an improved risk profile of infants (eg, due to a better obstetrical care), that is, to ‘better babies’, and/or to improved neonatal assistance (‘better care’). GA, BW, location of birth (inborn/outborn), sex and other characteristics of infants remained constant; multiple pregnancies and caesarean sections increased, and a slight improvement of Apgar score at 1 min was observed. We measured the baseline risk of infants by using the VON-RA score that takes into account GA, BW for GA, mode of delivery, multiple pregnancy, 1-min Apgar score, race, being inborn and sex, and it remained practically the same.25 Among other possible confounding variables, antenatal steroid prophylaxis (which is not accounted for in the VON-RA model) increased. However, even after adjusting for baseline characteristics and antenatal steroids, being born in 2010 was associated with a 29% reduction in mortality and 11% reduction in BPD. We can therefore postulate that this highly significant improvement can be interpreted as a sign of ‘better neonatal care’ between the years.25 Although we described an increase of NIV accompanied by a reduction in both mortality and BPD, we cannot claim a causal relationship between the changes in ventilator management and the better outcomes reported in adjusted analyses. A randomised trial would be necessary to support this relationship, but by the very nature of the problem, this is impossible.

    In conclusion, our results confirm that the care of very preterm infants during the studied period has considerably changed and has been helped by a number of additional non-invasive approaches and therapies that can reduce the adverse effects associated with intubation and mechanical ventilation.

    Acknowledgments

    We thank J Horbar and E Edwards, Vermont Oxford Network, for providing us with the coefficients used to calculate Vermont Oxford Network Risk-Adjustment.

    References

    1. ↵
      1. Horbar JD,
      2. Carpenter JH,
      3. Badger GJ,
      4. et al
      . Mortality and neonatal morbidity among infants 501–1500 grams from 2000 to 2009. Pediatrics 2012;129:1019–26.
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Corchia C,
      2. Ferrante P,
      3. Da Frè M,
      4. et al
      . Cause-specific mortality of very preterm infants and antenatal events. J Pediatr 2013;162:1125–32.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Ersch J,
      2. Roth-Kleiner M,
      3. Baeckert P,
      4. et al
      . Increasing incidence of respiratory distress in neonates. Acta Paediatr 2007;96:1577–81.
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Sekar KC,
      2. Corff KE
      . To tube or not to tube babies with respiratory distress syndrome. J Perinatol 2009;29(Suppl 2):S68–72.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Mosca F,
      2. Colnaghi M,
      3. Fumagalli M
      . BPD: old and new problems. J Matern Fetal Neonatal Med 2011;24(Suppl 1):80–2.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Gagliardi L,
      2. Bellù R,
      3. Lista G,
      4. et al
      . Do differences in delivery room intubation explain different rates of bronchopulmonary dysplasia between hospitals? Arch Dis Child Fetal Neonatal Ed 2011;96:F30–5. Epub 2010 Jul 21.
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Morley CJ,
      2. Davis PG,
      3. Doyle LW,
      4. et al
      .; COIN Trial Investigators. Nasal CPAP or intubation at birth for very preterm infants. N Engl J Med 2008;358:700–8.
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      Committee on Fetus and Newborn. Respiratory support in preterm Infants at birth. Pediatrics 2014;133;171.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Donn SM,
      2. Sinha SK
      . Invasive and non-invasive neonatal mechanical ventilation. Respir Care 2003;48:426–39; discussion 439–41.
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Gagliardi L,
      2. Tagliabue P,
      3. Bellù R,
      4. et al
      . Survey of neonatal respiratory support use in very preterm infants in Italy. J Matern Fetal Neonatal Med 2012;25(Suppl 3):1–5.
      OpenUrlPubMed
    11. ↵
      1. Horbar JD,
      2. Rogowski J,
      3. Plsek PE,
      4. et al
      . Collaborative quality improvement for neonatal intensive care. NIC/Q Project Investigators of the Vermont Oxford Network. Pediatrics 2001;107:14–22.
      OpenUrlAbstract/FREE Full Text
    12. ↵
      Vermont Oxford Network Database Manual of Operations for Infants Born in 2013. Release 17.1 Vermont Oxford Network. Burlington, published April 2013.
    13. ↵
      1. Zupancic JA,
      2. Richardson DK,
      3. Horbar JD,
      4. et al
      . Vermont Oxford Network SNAP Pilot Project Partecipants. Revalidation of the Score for Neonatal Acute Physiology in the Vermont Oxford Network. Pediatrics 2007;119:e156–63.
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Gagliardi L,
      2. Bellù R
      . Score for Neonatal Acute Physiology (SNAP) or Vermont Oxford risk-adjustment model for very low birth weight infants? Pediatrics 2007;119:1246–7; author reply 1247.
      OpenUrlFREE Full Text
    15. ↵
      1. Soll RF,
      2. Edwards EM,
      3. Badger GJ,
      4. et al
      . Obstetric and neonatal care practices for infants 501 to 1500 g from 2000 to 2009. Pediatrics 2013;132:222–8.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Roberts D,
      2. Dalziel S
      . Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth. Cochrane Database Syst Rev 2006;(3):CD004454.
    17. ↵
      1. Malloy MH
      . Impact of caesarean section on neonatal mortality rates among very preterm infants in the United States, 2000–2003. Pediatrics 2008;122:285–92.
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Kattwinkel J,
      2. Perlman JM,
      3. Aziz K,
      4. et al
      . Neonatal resuscitation: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Pediatrics 2010;126:e1400–13.
      OpenUrlFREE Full Text
    19. ↵
      1. Engle WA
      , American Academy of Pediatrics Committee on Fetus and Newborn. Surfactant-replacement therapy for respiratory distress in the preterm and term neonate. Pediatrics 2008;121:419–32.
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Sandri F,
      2. Plavka R,
      3. Ancora G,
      4. et al
      . Prophylactic or early selective surfactant combined with nCPAP in very preterm infants. Pediatrics 2010;125:e1402–9.
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Finer NN,
      2. Carlo WA,
      3. Walsh MC,
      4. et al
      .; SUPPORT Study Group of the Eunice Kennedy Shriver NICHD Neonatal Research Network. Early CPAP versus Surfactant in Extremely Preterm Infants. N Engl J Med 2010;362:1970–9.
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Bancalari E,
      2. Claure N
      . The evidence for non-invasive ventilation in the preterm infant. Arch Dis Child Fetal Neonatal Ed 2013;98:F98–F102.
      OpenUrlFREE Full Text
    23. ↵
      1. Manley BJ,
      2. Owen LS,
      3. Doyle LW,
      4. et al
      . High-flow nasal cannulae in very preterm infants after extubation. N Engl J Med 2013;369:1425–33.
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Wilkinson D,
      2. Andersen C,
      3. O'Donnell CP,
      4. et al
      . High flow nasal cannula for respiratory support in preterm infants. Cochrane Database Syst Rev 2011;(5):CD006405.
    25. ↵
      1. Richardson DK,
      2. Gray JE,
      3. Gortmaker SL,
      4. et al
      . Declining severity adjusted mortality: evidence of improving neonatal intensive care. Pediatrics 1998;102(4 Pt 1):893–9.
      OpenUrlAbstract/FREE Full Text

    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

    • Collaborators The Italian Neonatal Network included: D Turoli, G Weissmann, G De Nisi, M Barbarini, E Villa, C Franco, V Fasolato, MR Contiero, L Cattarossi, S Ellero, P Tagliabue, L Abbiati, C Fabris, G Prandi, F Vielmi, M Agosti, F Tandoi, M De Curtis, R Lucchini, F Ferrari, C Gallo, E Bellante, S Boccacci, G Latini, R Giannuzzi, S Martinelli, A Brunelli, ML Di Nunzio, M Filippone, G Carli, G Ancora, P Mammoliti, GP Gancia, C Dalmazzo, F Messina, R Magaldi, M Rinaldi, P Lago, OS Saia, S Visentin, G Presta, D Cella, C Poggiani, D Ferrari, G Barera, M Bove, G Burgio, E Sala, I Barberi, V Tiralongo, G Faldella, S Grandi and AG Cimatti.

    • Funding The Italian Neonatal Network has been supported by a grant by Chiesi.

    • Competing interests None.

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