In Norway, a national consensus-based guideline used to address thresholds for offering life support at extreme preterm birth was issued in 1998. Since then, therapeutic advances may have changed attitudes and expectations to treatment, both within the medical community and the public, and there are concerns that systematic variations in treatment practices may exist. With this article, we describe current practices and relate them to other ethical and legal comparable areas in health care. We conclude that a revision of the 1998 guideline is warranted to obtain a common understanding of prognoses and appropriate decision processes at the limit of viability.
Norway adopted a national policy on life-sustaining treatment of infants born at the limits of viability in 1998. That year, the Research Council of Norway organized a national consensus conference with the intention to “promote good medical practice and fair prioritization in an area where professionals disagree or benefit is dubious.”1 The conference was organized by a multidisciplinary group of neonatologists from different Norwegian NICUs, psychologists, nurses, ethicists, and representatives from the Council. During the conference, the challenges of predicting sequelae and the burdens of treatment of both the infant and the family were discussed. Legal and ethical issues were analyzed, and international practice variation was considered. Finally, a separate multidisciplinary consensus panel concluded on the consensus questions. They recommended that treatment before 23 weeks’ gestation was futile and should be considered experimental. Treatment at 23 and 24 weeks’ gestation should be optional and based on the infant’s vitality and the individual physician’s judgement. From week 25, the prognosis was considered good enough that life-supportive treatment should be considered mandatory unless there were other major negative prognostic markers (Table 1).1 The definition of the gray zone between 23 and 24 weeks’ gestation is still in line with guidelines in other high-income countries.2,3
The Conclusions of the Norwegian Consensus Conference of 1998
| Infants born after 25 completed wk should usually receive life-saving treatment. |
| Treating infants with GA <23 wk is considered experimental and should only be initiated according to a research protocol approved by a research ethics committee and after obtaining informed consent from parents. |
| Treatment of infants born at 23–25 completed wk should be assessed on an individual basis, taking into account the child’s vitality and the individual doctor’s clinical judgement. |
| Decisions to withhold or withdraw treatment should be based on defined diagnostic and prognostic criteria, not on assessments of future quality of life. Methods for assessing quality of life lack sensitivity, and judgment of the quality of life of others is influenced by subjective opinions. |
| On the basis of discussions, each unit should establish procedures and guidelines that ensure that all relevant considerations, including the parents’ attitudes, are included in the decision-making process. |
| Infants born after 25 completed wk should usually receive life-saving treatment. |
| Treating infants with GA <23 wk is considered experimental and should only be initiated according to a research protocol approved by a research ethics committee and after obtaining informed consent from parents. |
| Treatment of infants born at 23–25 completed wk should be assessed on an individual basis, taking into account the child’s vitality and the individual doctor’s clinical judgement. |
| Decisions to withhold or withdraw treatment should be based on defined diagnostic and prognostic criteria, not on assessments of future quality of life. Methods for assessing quality of life lack sensitivity, and judgment of the quality of life of others is influenced by subjective opinions. |
| On the basis of discussions, each unit should establish procedures and guidelines that ensure that all relevant considerations, including the parents’ attitudes, are included in the decision-making process. |
Adapted from Miljeteig I, Markestad T, Norheim OF. Physicians’ use of guidelines and attitudes to withholding and withdrawing treatment for extremely premature neonates in Norway. Acta Paediatr. 2007;96(6):825–829.
The Norwegian guideline states specifically that the decision should be made by the responsible physician and not by the parents. However, the parents’ opinion should be sought and included in the physician’s decision-making. This maternalistic approach is based on an assumption that it is humane not to burden parents with the responsibility of making decisions about withholding life-sustaining treatment.
In recent years, the use of gestational age (GA) as the sole factor guiding decisions to limit life support has been rightfully criticized.4,5 The basis for decision-making in the gray zone, as defined in the Norwegian consensus document, is therefore problematic.
In this article, we will discuss if there is a need for a new periviability guideline in Norway. We will also explore if there are legitimate arguments for considering the smallest premature infants as different from other critically sick patients.6,7
Background
The first neonatal ward in Norway opened in Oslo in 1957. Since then, NICUs were established in all university hospitals. During the 1980s, the GA below which infants were not considered viable has shifted downward from 28 weeks. By 1998, it was 23 or 24 weeks, and this corresponds with a birth weight between 400 and 750 g.8
Neonatal intensive care has contributed to dramatically improved outcome for sick and premature neonates. In Norway, neonatal mortality decreased 60% during the 1980s and 1990s, and infant mortality is now among the world’s lowest at 2.2 per 1000 live births.9
In parallel with the technological and medical advances, neonatal care in Norway has given rise to ethical controversies about whether there should be a lower GA limit for offering life support and how and by whom such decisions should be made.10,11 The question of whether to provide neonatal resuscitation is preceded by a number of obstetric decisions, such as the timing of transfer to a tertiary center, the use of antenatal steroids, monitoring of the fetus, and the consideration of cesarean delivery.12 These decisions involve parents and siblings as well as obstetricians, midwives, neonatologists, neonatal nurses, ethicists, and society as a whole.
Is the 1998 Consensus Guideline Obsolete? Current Practices
Preterm birth usually occurs unexpectedly and with consequences that most people know little about. In that situation, an individualized approach leaves room for each individual physician’s knowledge, opinions, and values to be the main determinants for the decisions being made. A guideline that allows gray-zone controversies to be resolved by the individual physician’s judgement allows flexibility but will also inevitably create controversies that are more complex than those seen in many other medical fields. Unlike similar decisions for other patient populations, decisions about providing life support to infants born at the borderline of viability are closely related to questions about when the life of a human being starts (conception, birth, or somewhere in between or even after), ideas about the creation of the human species (by God or by nature or evolution), and ideas about abortion (murder or the legitimate right for a woman to decide).
Guidelines must be based on effects of new treatment options and current prognosis, capture present attitudes and opinions among those working in the field, and reflect common moral norms in the society. Indeed, there are indications that Norwegian health care personnel have become more willing to provide life support at lower GA since the 1998 guidelines. Such evidence comes from a questionnaire-based study on changes in attitudes and practices over time in Norwegian obstetric and pediatric units in 2005. The study revealed that the mean reported GA threshold for resuscitating infants had decreased by almost a week from 23 + 6/7 to 23 + 0/7 weeks since 1998.13 Physicians did not perceive this change in threshold, but 30% felt that decisions regarding provision of life support had become more difficult. Almost half of the responding physicians agreed with the statement that Norway was too liberal in providing life support to the most immature infants.
In 2009, the 2 university hospitals in Oslo merged into 1 organization. The 2 institutions appeared to have a somewhat different approach to providing life support when faced with preterm labor between 22 and 25 weeks’ GA. In addition, treatment and practice variations for infants born at the limit of viability had received increased public attention.14 Therefore, in 2014, the Perinatal Committee of the South-Eastern Norway Regional Health Authority asked the National Council for Priority Setting in Health Care to consider a revision of the 1998 guidelines.15 The Council conducted a questionnaire study on current practices and local guidelines, which was supplemented by data from the Norwegian Neonatal Network (NNN). On request from the Council, the Norwegian Knowledge Center for the Health Services summarized international literature on recent GA-specific prognosis on survival and neurodevelopmental outcome.16
The Council collected information from all 21 neonatal units and 40 obstetric units in Norway during January to April 2015. It seemed that the units could be divided into the following 2 categories15:
Units mainly providing life support to all infants from 23 + 0/7 weeks’ GA; and
Units providing life support to all infants from 24 + 0/7 weeks’ GA, making an individual judgment of infants born at 23 weeks’ GA.
These findings are reflected in data from the NNN revealing that in the period from 2009 to 2014, the proportion of live-born infants transferred to a NICU was nearly 97% at 24 weeks, 74% at 23 weeks, and 19% at 22 weeks. During those years, only 5 infants born at 22 weeks had been treated. These all occurred in just 3 of Norway’s 21 NICUs.
The gray zone in Norway appears to have moved from 24 to 23 weeks; at 22 weeks treatment is rarely offered, and at 24 weeks treatment is normally started. It is difficult to delineate the driving force behind this change in attitudes and apparently also in practice. One plausible explanation would be if prognosis has changed to the better. International studies revealing increased survival among the tiniest may have had an influence, especially when coming from our neighbor Sweden.17 Both internationally and nationally, intensive care is now sometimes provided for infants born at 22 weeks.17,–20
Authors of a recent study from the NNN examined 1-year survival and rates of major neonatal morbidities among infants born at 22 to 26 weeks’ GA between 2013 and 2014 (Norwegian Extreme Prematurity Study 2) and compared the results to those of the first Norwegian Extreme Prematurity Study born between 1999 and 2000 and other contemporary studies.20 Although there is considerable variation between different studies, and results clearly reflect practice variations (if you do not offer treatment to a 23-weeker, survival is 0), the best-estimated survival rates have not changed substantially since 1998 (Table 2). In the Norwegian Extreme Prematurity Study 2, 55% of the survivors experienced major neonatal morbidity, and major morbidity was moderately associated with decreasing GA (odds ratio: 1.6; 95% confidence interval [CI], 1.2–2.2). The disability rate also correlated with GA in the systematic review performed by the Norwegian Knowledge Center (Table 3).16
Survival Rates of Periviable Infants Admitted to the NICU, 1995–2014
| GA, wk . | Epicure21 United Kingdom and Ireland 1995, n of N (%) (95% CI) . | NEPS 122 Norway 1999–2000, n of N (%) (95% CI) . | NEPS 220 Norway 2013–2014, n of N (%) (95% CI) . | EXPRESS23 Sweden 2004–2007, n of N (%) . | NICHD24 2012, n of N (%) . | NKC16 Systematic Review 2000–2016, n of N (%) (95% CI) . |
|---|---|---|---|---|---|---|
| 22 | 2 of 22 (9) (0–21) | 0 of 2 (0) (0) | 3 of 5 (60) (15–95) | 5 of 51 (10) | 7 of 75 (9) | 98 of 285 (33) (27–40) |
| 23 | 26 of 131 (20) (13–27) | 9 of 23 (39) (19–59) | 12 of 34 (35) (20–54) | 53 of 101 (53) | 50 of 150 (33) | 516 of 1052 (48) (31–66) |
| 24 | 100 of 298 (34) (28–39) | 35 of 68 (60) (49–74) | 35 of 60 (58) (45–71) | 96 of 144 (67) | 174 of 269 (65) | 1255 of 1923 (68) (51–81) |
| 25 | 186 of 357 (52) (47–57) | 55 of 69 (80) (70–89) | 59 of 69 (86) (75–93) | 167 of 205 (82) | 249 of 308 (81) | 73% (65–81) |
| GA, wk . | Epicure21 United Kingdom and Ireland 1995, n of N (%) (95% CI) . | NEPS 122 Norway 1999–2000, n of N (%) (95% CI) . | NEPS 220 Norway 2013–2014, n of N (%) (95% CI) . | EXPRESS23 Sweden 2004–2007, n of N (%) . | NICHD24 2012, n of N (%) . | NKC16 Systematic Review 2000–2016, n of N (%) (95% CI) . |
|---|---|---|---|---|---|---|
| 22 | 2 of 22 (9) (0–21) | 0 of 2 (0) (0) | 3 of 5 (60) (15–95) | 5 of 51 (10) | 7 of 75 (9) | 98 of 285 (33) (27–40) |
| 23 | 26 of 131 (20) (13–27) | 9 of 23 (39) (19–59) | 12 of 34 (35) (20–54) | 53 of 101 (53) | 50 of 150 (33) | 516 of 1052 (48) (31–66) |
| 24 | 100 of 298 (34) (28–39) | 35 of 68 (60) (49–74) | 35 of 60 (58) (45–71) | 96 of 144 (67) | 174 of 269 (65) | 1255 of 1923 (68) (51–81) |
| 25 | 186 of 357 (52) (47–57) | 55 of 69 (80) (70–89) | 59 of 69 (86) (75–93) | 167 of 205 (82) | 249 of 308 (81) | 73% (65–81) |
Survival to discharge: Epicure, NEPS, NICHD. Survival to 1 y: EXPRESS. EXPRESS, Extremely Preterm in Sweden Study; NEPS, Norwegian Extreme Prematurity Study; NICHD, Eunice Kennedy Shriver National Institute of Child Health and Human Development; NKC, Norwegian Knowledge Center.
Meta-analysis: Risk of Severe Impairment at 1.5–3 Years in Surviving Periviable Infants
| GA, wk . | Risk of Severe Disability: n Severely Disabled of N Total (No. Studies) . | Estimate % (95% CI) . |
|---|---|---|
| 22 | 9 of 24 (4) | 37 (21–58) |
| 23 | 34 of 125 (6) | 20 (5–52) |
| 24 | 62 of 299 (9) | 18 (11–29) |
| 25 | 62 of 485 (8) | 11 (7–18) |
| GA, wk . | Risk of Severe Disability: n Severely Disabled of N Total (No. Studies) . | Estimate % (95% CI) . |
|---|---|---|
| 22 | 9 of 24 (4) | 37 (21–58) |
| 23 | 34 of 125 (6) | 20 (5–52) |
| 24 | 62 of 299 (9) | 18 (11–29) |
| 25 | 62 of 485 (8) | 11 (7–18) |
Data are from the Norwegian Knowledge Center.
All of these data and developments raise questions about what policy should guide obstetricians and neonatologists in Norway today.
Are Premature Infants Different? The Continuum of Human Development
In some countries, national, regional, and/or institutional guidelines have been developed. Others eschew guidelines and claim that infants born at the border of viability should be treated like any other critically ill patient.2,7 Numerous surveys have disclosed that attitudes toward these tiniest premature infants differ from those toward other critically sick patients. Strikingly, health care professionals are more reluctant to treat their own hypothetical periviable infant than others.25,26
It is difficult to disentangle questions about the treatment of extremely premature infants from questions on the ethics and legality of abortion. Many abortion regulations, including those of Norway, consider the GA of the fetus when deciding whether to permit abortion. In Norway, the fetus has a graded protection; week 12 + 0/7 is the threshold for abortion-on-demand, until week 18 + 0/7 an abortion can be performed after individual consideration by a committee, and until week 21 + 6/7 abortion can be allowed in extreme circumstances. After week 21 + 6/7, pregnancy cannot be terminated unless the fetus suffers from a condition incompatible with life, or the pregnant woman’s life is endangered by continuing the pregnancy.27
The transition from late abortion, when we consider the fetus to be nonviable, to the extreme premature birth of a potentially viable infant is not straightforward. Because of the interaction between multiple factors (GA, sex, plurality [twins or triplets], provision of antenatal steroids, growth restriction, infection) and the infant’s unknown genetic potential, we cannot predict with certainty the individual infant’s prognosis. Combined with the heavy burden of treatment and lack of established social relationships, this may justify the existence of a broader gray zone than for other patient groups.
During pregnancy, the woman has a wide decision-making range. Up to 21 + 6/7 weeks’ GA, she can choose to abort a fetus presenting congenital anomalies with potentially better prognosis than a periviable infant. We acknowledge that the woman’s bodily right is no longer at stake once the fetus or infant is outside her body. At birth, the ethical focus passes from beneficence and autonomy on part of the pregnant woman to the child’s best interest. However, in terms of conceiving the family as a unit and the coming child as a part of their family project, the 2 situations are not far from each other.
Parental Counseling and Shared Decision-making
As neonatologists, we have an ethical and legal obligation to act in the child’s best interest. This is established in the United Nations Convention on the Rights of the Child.28 Infants have the right to individual clinical assessment, the right to life, and the right to be protected from futile treatment and undue suffering.29 These considerations sometimes conflict, as when potentially life-saving treatment causes suffering and may turn out to be futile. But who should decide?
There is a general trend toward increased patient autonomy in our society. This is reflected in the Norwegian Patients’ Rights Act of 1999. Parents or other persons with parental responsibility are entitled to consent to health care for patients <16 years of age (section 4-4) but with clear limitations: Parents cannot deny treatment that is clearly in the child’s best interest nor can they demand futile treatment. This is also specified in a national guideline for decision-making when considering limitations of life-sustaining treatment, which was issued by the Norwegian Directorate of Health in 2013. The guideline states that parental values and opinions should play a larger role the more uncertain the medical professionals are on prognosis and on what is in the child’s best interest.30 Still, it is unclear who has the final say in life and death decisions. A close reading of the law suggests that parents, not doctors, are the legal decision-makers in cases in which it is unclear whether life support or palliative care are in the best interest of the infant.31
The parents are the ones who will have to live with the consequences of a decision to try to save a tiny premature infant. Such care has uncertain outcomes. If the infant survives, it leads to months of intensive care and can be associated with ongoing suffering for both the child and the family. A trial of treatment, which by many is considered ethically superior to primary palliative care (“at least she was given a chance”32), should be treated with caution. Although withdrawal and withholding are ethically equivalent, many report that it is more difficult for parents to allow redirection of care toward palliation once they have bonded with their infant than it is for them to authorize palliative care at birth. Authors of a recent prospective multicenter study from the United Kingdom found that a significant proportion of parents chose to continue life-sustaining treatment after discussions on limitation of treatment of their newborn infants.33 Data suggest that even physicians find it more difficult to withdraw than to withhold treatment.34
It may be difficult for many parents to accept withdrawal of life support. Some doctors even consider it to be inhumane to ask parents to accept such a decision. A Norwegian study revealed that many parents of preterm infants expressed an unwillingness to be decision-makers.35 However, authors of a similar study in Scotland indicated that parents saw the decision-making as a part of their parental responsibility.36 The desire to protect Norwegian parents against the burden of decision-making is reiterated in a recent article, in which authors stated that many parents wish to be exempt from decisional responsibility in end-of-life decisions for their children.32 We believe that an improved information-sharing process can positively impact the parents’ ability to accept redirection of care and to perceive this as caring decisions for their infant.
Do We Need Consensus? National Guideline as a Medical and Ethical Tool
Life and death decisions at the border of viability are high stake, time sensitive, and emotionally charged. Ethically, we can choose the “safe option” and start life-saving treatment in all periviable infants on the basis of the principle of the child’s right to life. This approach was rejected by the Consensus Panel in 1998 on the basis of the belief that the costs for the infant, family, and society were too high. The chosen strategy was to strive to reduce uncertainty by taking into account all prognostic factors individually and to work toward consensus between all involved parties. The final decision was ultimately left to the discretion of the responsible neonatologist. This approach may now seem outdated because it may not reflect national values in an age of patient-centered, nonpaternalistic shared decision-making.
But the other approach may not be better. Many parents prefer paternalistic doctors.
One argument against parental decision-making in this particular setting is the frequent urgency of the situation. The parents are in shock, there is no time for sensitive and thorough counseling, and, thereby, there is no possibility for the parents to make a well-informed decision on the basis of their personal values.
To overcome these challenges and to facilitate a well-informed, value-based shared decision-making process, we suggest working toward an increased awareness in society of the dilemmas at the border of viability. This, in turn, will allow a more robust process of shared decision-making. Both parents and professionals need accurate and up-to-date information about prognosis. The information should be provided in a neutral and nondirective style and supported by national written information. Both the medical and ethical dilemmas of periviability should be clarified. This knowledge, in addition to compassionate and respectful communication with a coordinated health care team will hopefully facilitate a better shared decision-making process, finding individualized goals of care with each family according to parental values and the child’s best interest.37
Conclusions
We should strive to reduce medically and ethically unjustified variation in practice at the border of viability in Norway. This does not mean that all infants must be treated identically. Instead, it requires that variations in practice should not be based on physician preferences or the policies of individual hospitals and NICUs. Instead, they should be based on the facts that parents are permitted to hold and live by different values.
Individualized decision-making allows an important role for national guidelines. Guidelines define an acceptable prognostic gray zone. Parents should receive sensitive counseling based on up-to-date national information material, which should also include a description of their veto right against intensive care at birth in that guideline-defined gray zone. At the same time, all expecting parents should have equal access to life-saving treatment of their periviable infants, regardless of which hospital they are admitted to or the neonatologist on call.
Future guidelines should be based on a thorough multidisciplinary process including parent organizations. We do not want to anticipate the conclusion of this work, and we support the National Council for Priority Setting in Health Care’s call for a broad national guideline.
Dr Syltern wrote the first draft, coordinated the manuscript, and is the corresponding author; Dr Støen supervised the development of work, revised the manuscript critically for important intellectual content, and helped to evaluate and edit the manuscript; Drs Markestad and Saugstad revised the manuscript critically for important intellectual content and helped to evaluate and edit the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: Supported by St Olav’s University Hospital and Norwegian University of Science and Technology grant.
Acknowledgments
We thank John Lantos for his useful suggestions for improving the article. Funding from St Olavs Hospital and Norwegian University of Science and Technology making the seminar on Tautra possible is gratefully acknowledged.
References
Competing Interests
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
Comments