Decisions at the limit of viability

This is the authors' abstract. We don't add key points for this paper.

Since the first prematurely born infants were mechanically ventilated in the 1960s, medical advances have encouraged life support to ever more immature infants. This development has repeatedly caused concern that risks of severe health consequences might outweigh the benefits to the children and their families. The advocates of a more active approach have pointed out that avoidance would generate a self-fulfilling prophesy of poor outcome. They were proven right for a long time, but maybe not with the latest move towards a gestational age (GA) of 22 weeks? In this issue of Acta Paediatrica, Lundgren et al. publish the Swedish experience on neonatal outcome after adopting proactive obstetric and neonatal interventions to salvage infants born as early as GA 22 and 23 weeks.1 During 2007–2018, the stillbirth rate decreased for infants with GA 22 weeks, and the survival rates of live-born infants increased from 19% to 38% for GA 22 weeks and from 45% to 67% for GA 23 weeks. The largest increase occurred from 2016 when a new national guideline stated that neonatal resuscitation should be considered from GA 22 weeks and recommended from 23 weeks.2 From 2016, 91% of the mothers of 22-week-old infants received antenatal corticosteroids and 22% gave birth by caesarean section suggesting that initiating life support from 22 weeks has become the norm in Sweden. During these years, the rates of severe neonatal morbidities among the survivors did not change, except that necrotizing enterocolitis became more common for those born at GA 22 weeks. At follow-up at ages 2–13 years, 4% of the 22-week GA children had died after discharge, 78% had neurodevelopmental and 87% somatic disorders – only one of 93 surviving children did not have a diagnosis.3 The outcome was moderately better for infants born at 23 weeks, but they had a higher rate of severe cerebral morbidity as newborns than the survivors of GA 22 weeks. It is therefore likely that the difference in neurodevelopmental outcome would have been larger if the criteria for discontinuing life support were the same for infants born at 22 and 23 weeks. The GA threshold for recommending proactive care varies between comparable countries. In a meta-analysis of studies on outcome after extremely preterm births from the late 1990s through 2015 in high-income countries, the mean chance of survival decreased progressively from GA 27 to 22 weeks when stillbirths were included, and from GA 25 weeks for infants born alive.4 Judged from the data, most of the institutions usually practiced comfort care at GA 22 weeks and proactive care from GA 25 weeks while more individualised decisions were made in the intermediate ‘grey zone’. The rate of severe and moderate-to-severe neurodevelopmental impairments increased with each declining GA below 26 weeks, and the rates did not appear to change with year of birth for infants born between 2000 and 2015, as in the Swedish study. The children were assessed with Bayley scales of infant development between 18 and 36 months of age. For the survivors of 22 weeks, the calculated risk was 36% for severe and 41% for moderate-to-severe impairments. The risks were probably higher because the Bayley scales generally overestimate development at this age and thereby under-identify children with developmental delay, as do other assessment tools during young age.5 The trauma of treatment increases exponentially with declining GA below 26 weeks.6 Being born at GA 22 and 23 weeks means months of intensive care, many weeks intubated on mechanical ventilation and countless painful procedures. Although we do not know the life course consequences, most of the survivors will probably experience significant health, social and quality of life challenges, as discussed by Neil Marlow in relation to the follow-up paper on the Swedish cohort.7 At the limit of viability, it is ethically challenging to weigh the burden of treatment and risks of death and severe impairments against the chance of experiencing an independent life without major health problems. Decisions should reflect accepted moral norms and priorities, which may differ between societies, and guidelines should ideally reflect the people's perception of what is morally acceptable and what is in the best interest of the child and family. The crucial question is whether obstetricians and neonatologists act according to the moral compass of our societies. In Norway, we explored the views on proactive care among paediatric residents who attended a compulsory one-week national training courses in neonatology. At the end of each of three courses during 2017–20, the participants were asked to anonymously state whether they would want life support to be started if their own infant was born at GA 22, 23, 24, 25 or 26 weeks. None of 80 respondents wanted life support at 22 or 23 weeks, but 30% expressed uncertainty at 23 weeks. Even at GA 25 weeks, only 59% answered in favour of life support and 34% were uncertain although Norway practice proactive care from GA 24 weeks and provide life support for most infants born alive at GA 23 weeks.8 The discrepancy between their own view and how they act challenges the universal ‘Golden Rule’ in ethics which can be expressed as ‘you should not do to others what you wouldn't want others to do to you’. The residents’ opinion may not, however, reflect that of potential parents since the question was hypothetical and because physicians may acquire attitudes that are different from the population in general. On the other hand, they had the advantage of updated knowledge and time to reflect, as opposed to parents who face the dilemma unexpectedly and unprepared. In this vulnerable situation decisions will tend to be determined at the mercy of the attitude and information conveyed verbally and non-verbally by the physician, even if a guideline states that the parents’ wish should be taken into consideration. The Swedish guideline of 2016 is written by dedicated obstetricians and neonatologists.2 It emphasises the importance of informing parents, but the parents are not described as partners when making decisions. The right to abortion is heatedly debated in most countries while decisions related to limits of viability are rarely discussed, even superficially, in public. It may be defined as the earliest GA when it is possible to save the life of the infant with current technology. Alternative definitions may be the time when consciousness is beginning to emerge, which is believed to be at around GA 24 weeks,9 or within the intermediate ‘grey zone’ when the benefits to the infant and the family is considered to outweigh the burden of treatment and risks of death and impairments. For the individual family, the decision is naturally made in the meeting between the parents and physicians, but hopefully within borders that is accepted in the society. However, both healthcare personnel and parents need guidance from the society. In my view, the development of guidelines concerning decisions at the limit of viability should not only be left to the involved physicians, but also include other relevant health care professionals, experts on ethics and law, and other relevant stakeholders including representatives of the public, and it should be anchored in a governmental institution like a national board of health. I thank Janicke M Syltern, MD, for providing data. None.