BACKGROUND:

It is widely acceptable to involve parents in decision-making about the resuscitation of extremely preterm infants (EPIs) in the gray zone. However, there are different views about where the boundaries of the gray zone should lie. Our aim in this study was to compare the resuscitation thresholds for EPIs between neonatologists in the United Kingdom, Sweden, and the Netherlands.

METHODS:

We distributed an online survey to consultant neonatologists and neonatal registrars and fellows that included clinical scenarios in which parents requested resuscitation or nonresuscitation. Respondents were asked about the lowest gestational age and/or the worst prognosis at which they would provide resuscitation and the highest gestational age and/or the best prognosis at which they would withhold resuscitation. In additional scenarios, influence of the condition at birth or consideration of available health care resources was assessed.

RESULTS:

The survey was completed by 162 neonatologists (30% response rate). There was a significant difference between countries; the gray zone for most UK respondents was 23 + 0/7 to 23 + 6/7 or 24 weeks’ gestation, compared with 22 + 0/7 to 22 + 6/7 or 23 weeks’ gestation in Sweden and 24 + 0/7 to 25 + 6/7 or 26 weeks’ gestation in the Netherlands. Resuscitation thresholds were higher if an infant was born in poor condition. There was wide variation in the prognosis that warranted resuscitation or nonresuscitation. Consideration of resource scarcity did not alter responses.

CONCLUSIONS:

In this survey, we found significant differences in approach to the resuscitation of EPIs, with a spectrum from most proactive (Sweden) to least proactive (Netherlands). Most survey respondents indicated shifts in decision-making that were associated with particular weeks’ gestation. Despite the different approaches to decision-making in the 3 countries, there was relatively little difference between countries in neonatologists’ prognostic thresholds for resuscitation.

Extremely preterm infants (EPIs) have high rates of mortality and morbidity.1 Those who are born at the lowest gestational ages require long periods of respiratory support and intensive care.2 It is widely thought to be ethically acceptable to involve parents in decisions about resuscitation for infants who are at the margin of viability or within the so-called “gray zone.”3 However, there are different views about where the boundaries of the gray zone should lie4 (ie, when infants are too immature for resuscitation, it should not be attempted even if desired [the “lower threshold” of the gray zone], and when infants are sufficiently mature, active resuscitation should be considered mandatory [the “upper threshold” of the gray zone]).3 

There are a large number of national guidelines regarding the resuscitation of EPIs in high-income countries.4,6 Most of these guidelines provide recommendations for the lower and upper thresholds on the basis of estimated gestational age. However, it is not clear how strictly neonatologists adhere to these recommendations or how they take into account infants’ conditions at birth. Gestational age–based guidelines have also been criticized.7,9 Such guidelines are argued to be overly simplistic, to ignore uncertainty in the estimation of gestational age, to neglect other important prognostic factors, and even to represent discrimination against EPIs.7,10 Some authors have argued that decisions should instead be individualized8,11,12 or explicitly based on infants’ expected prognosis.13 Yet, there is little existing guidance on what prognosis would justify a decision to provide or withhold resuscitation.3 It is unclear when neonatologists would judge a prognosis to be too poor to resuscitate or too good to provide comfort care at birth. Furthermore, thresholds and treatment strategies may also be context dependent.

Our aim in this study was to explore and compare the resuscitation thresholds for EPIs between different European countries. We surveyed neonatologists in the United Kingdom, Sweden, and the Netherlands, countries that have previously been shown to have broadly similar approaches to ethical decision-making in neonatal intensive care.14,15 

We sought the thresholds that they would apply on the basis of either gestational age or prognosis. We also aimed to explore the clinicians’ experience of resuscitating outside of prevailing guidance and whether concern for limited health care resources impacts decisions.

Consultant neonatologists and neonatal registrars and fellows in the United Kingdom, Sweden, and the Netherlands were invited to participate in an anonymous online survey between May 2016 and September 2016. Recruitment differed between countries. In the United Kingdom, doctors who were attending a conference on neonatal ethics were invited to participate before the conference. Members of the Swedish Neonatal Society and the Neonatology Section of the Dutch Pediatric Association were invited to participate by e-mail.

The survey was developed by the study authors after a review of the medical literature and after receiving feedback from practicing neonatologists in each country. The survey included basic demographic questions (sex, country, years of experience, and religion) and questions about approaches to the resuscitation of EPIs (Supplemental Information).

Respondents were asked questions relating to 2 main scenarios. In the first scenario, parents expected delivery of an EPI and requested active resuscitation. Respondents were asked about the lowest gestational age at which they would be prepared to provide active resuscitation if the infant was born in good condition (spontaneously breathing and with a heart rate of 100), or in poor condition (poor tone, no respiratory effort, and a heart rate of 40). Available options included a gestational age range from 21 to 25 weeks and “always” or “other” (free text) responses. They were also asked what probability of survival without severe disability would be too low to justify resuscitation for an EPI.

In the second scenario, parents expected delivery of an EPI and requested nonresuscitation because they were concerned that the infant would die despite treatment or survive with severe disability. Respondents were asked about the highest gestational age (between 21 and 27 weeks) at which they would be prepared to withhold active resuscitation if the infant was born in good or poor condition as described above. They were also asked what probability of survival without severe disability would be too high to justify withholding resuscitation.

In additional survey questions, respondents were asked if they had local or national professional guidelines relating to the resuscitation or nonresuscitation of preterm infants. They were asked if, in the preceding 5 years, they had provided or withheld resuscitation outside of those guidelines. Finally, respondents were asked about their willingness to provide resuscitation or nonresuscitation if there was no limit to available health care resources.

A statistical analysis was conducted by using Wizard for Mac version 1.8.24. Differences between countries and responses were explored by using the Pearson’s χ2 test. Participants’ responses to different questions were compared by using the Stuart-Maxwell test for marginal homogeneity in paired comparisons. To represent graphically the total number of neonatologists who were willing to resuscitate at a given gestational age, we combined all those who indicated a resuscitation threshold at or below a given level (providing resuscitation on request) or at or above a given level (withholding resuscitation on request). We performed a multivariate ordered probit logistic regression to determine the independent effect of baseline characteristics, including country, sex (female versus male), years of work in the NICU, and religion, on the thresholds for providing or withholding resuscitation. A P value of ≤.05 was considered to indicate statistical significance (P < .01 for multivariate analysis because of multiple comparisons).

The study was approved by the Medical Sciences Interdivisional Research Ethics Committee at the University of Oxford (R45847/RE001) in May 2016.

We contacted 553 neonatal consultants, fellows, and registrars from the 3 countries (United Kingdom: 103; Sweden: 250; Netherlands: 200), and 162 completed the survey (response rates: 41%, 24%, and 29%, respectively). There was a higher proportion of consultants among Swedish and Dutch respondents, and they were more experienced (Table 1). There was a higher proportion of religious respondents from the United Kingdom and a higher proportion of non-Christian religions.

TABLE 1

Baseline Characteristics of Participants by Location

CharacteristicsOverall, n (%)United Kingdom, n (%)Sweden, n (%)Netherlands, n (%)Pa
Professional role (N = 162)      
 Consultant 125 (77) 22 (52) 61 (85) 42 (88) <.001 
 Registrar or fellow 37 (23) 20 (48) 11 (15) 6 (13) — 
Sex (N = 161)      
 Male 80 (50) 16 (38) 39 (55) 25 (52) .2 
 Female 81 (50) 26 (62) 32 (45) 23 (48) — 
Years working in the NICU (N = 162)      
 1–5 39 (24) 17 (41) 13 (18) 9 (19) .04 
 6–10 25 (16) 9 (21) 9 (13) 7 (15) — 
 11–15 35 (22) 9 (21) 15 (21) 11 (23) — 
 16–20 22 (14) 1 (2) 12 (17) 9 (19) — 
 >20 41 (26) 6 (14) 23 (32) 12 (25) — 
Religious (N = 161)      
 Yes 63 (39) 30 (71) 20 (28) 13 (27) <.001 
 Atheist 69 (43) 9 (21) 35 (49) 25 (52) — 
 Agnostic 29 (18) 3 (7) 16 (23) 10 (21) — 
Religion (N = 81)b      
 Christianity 58 (72) 14 (44) 29 (94) 15 (83) <.001 
 Islam 3 (4) 3 (9) 0 (0) 0 (0) — 
 Judaism 1 (1) 0 (0) 0 (0) 1 (6) — 
 Hinduism 11 (14) 11 (34) 0 (0) 0 (0) — 
 Other 8 (10) 4 (13) 2 (6) 2 (11) — 
CharacteristicsOverall, n (%)United Kingdom, n (%)Sweden, n (%)Netherlands, n (%)Pa
Professional role (N = 162)      
 Consultant 125 (77) 22 (52) 61 (85) 42 (88) <.001 
 Registrar or fellow 37 (23) 20 (48) 11 (15) 6 (13) — 
Sex (N = 161)      
 Male 80 (50) 16 (38) 39 (55) 25 (52) .2 
 Female 81 (50) 26 (62) 32 (45) 23 (48) — 
Years working in the NICU (N = 162)      
 1–5 39 (24) 17 (41) 13 (18) 9 (19) .04 
 6–10 25 (16) 9 (21) 9 (13) 7 (15) — 
 11–15 35 (22) 9 (21) 15 (21) 11 (23) — 
 16–20 22 (14) 1 (2) 12 (17) 9 (19) — 
 >20 41 (26) 6 (14) 23 (32) 12 (25) — 
Religious (N = 161)      
 Yes 63 (39) 30 (71) 20 (28) 13 (27) <.001 
 Atheist 69 (43) 9 (21) 35 (49) 25 (52) — 
 Agnostic 29 (18) 3 (7) 16 (23) 10 (21) — 
Religion (N = 81)b      
 Christianity 58 (72) 14 (44) 29 (94) 15 (83) <.001 
 Islam 3 (4) 3 (9) 0 (0) 0 (0) — 
 Judaism 1 (1) 0 (0) 0 (0) 1 (6) — 
 Hinduism 11 (14) 11 (34) 0 (0) 0 (0) — 
 Other 8 (10) 4 (13) 2 (6) 2 (11) — 

—, not applicable.

a

P values were obtained by using the χ2 test for comparison.

b

Respondents who indicated that they had a religion or were agnostic (not including 11 agnostic respondents who left this question blank).

For an EPI who was born in good condition and whose parents wished for active resuscitation, there was a significant difference between countries in regard to the lowest gestational age at which resuscitation would be provided (P < .001; Fig 1A). Of UK respondents, 60% would only resuscitate beyond 23 weeks’ gestation, whereas 56% of Swedish respondents would resuscitate beyond 22 weeks’ gestation, and 58% of Dutch respondents would only resuscitate beyond 24 weeks’ gestation. Ten Swedish respondents (6%) indicated that they would provide resuscitation at 21 weeks’ gestation or at any gestation if the infant was born alive.

FIGURE 1

Lower threshold: cumulative probability of resuscitation based on the lowest gestational age at which respondents would provide resuscitation. The figure indicates the proportion of respondents who would be prepared to resuscitate at, or above, a given gestational age if parents requested active treatment (n = 158). Four respondents who indicated an “other” free-text response were excluded. A, For an infant born in good condition (spontaneously breathing and with a heart rate of 100). B, For an infant born in poor condition (poor tone, no respiratory effort, and heart rate of 40).

FIGURE 1

Lower threshold: cumulative probability of resuscitation based on the lowest gestational age at which respondents would provide resuscitation. The figure indicates the proportion of respondents who would be prepared to resuscitate at, or above, a given gestational age if parents requested active treatment (n = 158). Four respondents who indicated an “other” free-text response were excluded. A, For an infant born in good condition (spontaneously breathing and with a heart rate of 100). B, For an infant born in poor condition (poor tone, no respiratory effort, and heart rate of 40).

If the infant was born in poor condition, there was a significant increase in the lower gestational age threshold for resuscitation (Fig 1B; P < .001). A smaller proportion of UK respondents would only resuscitate if the infant’s gestational age was 23 + 0/7 weeks (43% poor condition vs 60% good condition), whereas a larger proportion (45% poor vs 5% good) would only resuscitate if the infant’s gestational age was ≥24 weeks. A smaller proportion of Swedish respondents would resuscitate at 22 + 0/7 weeks’ gestation (36% poor vs 56% good), whereas a larger proportion (33% poor vs 10% good) of Swedish respondents would only resuscitate after 23 weeks’ gestation. A larger proportion of Dutch respondents would only resuscitate after 25 weeks’ gestation (19% poor vs 2% good).

Respondents varied in their judgement of the probability of survival without severe disability that would be too low to provide resuscitation (Fig 2, Supplemental Fig 4). Of respondents, 35 (22%) would be prepared to provide resuscitation no matter how low the probability. Of those who provided a probability, the largest group (20%) indicated that they would resuscitate if the infant had a >10% chance of survival without severe disability. There was some difference between countries in the probability thresholds. For example, a larger proportion of Swedish respondents than Dutch respondents would be prepared to resuscitate no matter how low the chance of survival (33% vs 6%; P < .01).

FIGURE 2

Prognosis-based thresholds: cumulative proportion of respondents who were prepared to either provide resuscitation at parental request or withhold resuscitation at parental request as the prognosis improved (n = 144 [provide resuscitation]; n = 137 [withhold resuscitation]). For example, for an infant with a 10% predicted chance of survival without profound disability, 52% of Dutch neonatologists were prepared to resuscitate if parents wished this, whereas 100% of Dutch neonatologists were prepared to withhold resuscitation (if parents did not wish the infant resuscitated). Respondents (18 and 25, respectively) were excluded in the analyses if they gave only a free-text response (ie, did not indicate a numerical threshold).

FIGURE 2

Prognosis-based thresholds: cumulative proportion of respondents who were prepared to either provide resuscitation at parental request or withhold resuscitation at parental request as the prognosis improved (n = 144 [provide resuscitation]; n = 137 [withhold resuscitation]). For example, for an infant with a 10% predicted chance of survival without profound disability, 52% of Dutch neonatologists were prepared to resuscitate if parents wished this, whereas 100% of Dutch neonatologists were prepared to withhold resuscitation (if parents did not wish the infant resuscitated). Respondents (18 and 25, respectively) were excluded in the analyses if they gave only a free-text response (ie, did not indicate a numerical threshold).

When respondents were asked to imagine that there was no limit to available resources, there was no significant change in the probability at which resuscitation would be offered nor was there a change in the proportion of respondents who were prepared to provide resuscitation in a hypothetical case of a preterm delivery at 22 + 4/7 weeks’ gestation (38% vs 40%).

There was a significant difference between countries in regard to the highest gestational age at which resuscitation would be withheld at parental request (P < .001; Fig 3). The majority of UK respondents (24 of 42 [57%]) would withhold resuscitation at a maximum of either 23 + 6/7 or 24 weeks’ gestation. In comparison, the majority of Dutch respondents (30 of 48 [60%]) would be prepared to withhold resuscitation at 25 + 6/7 or 26 weeks’ gestation, whereas half of Swedish respondents (36 of 72 [50%]) would withhold resuscitation at a maximum of 22 + 6/7 or 23 weeks’ gestation. If the infant was born in poor condition, there was a small increase in the gestational age at which respondents in the 3 countries were prepared to withhold resuscitation (Supplemental Fig 5).

FIGURE 3

Upper threshold: cumulative probability of nonresuscitation based on the highest gestational age at which neonatologists would withhold resuscitation at parental request in an infant who was born in good condition. The figure indicates the proportion of respondents who would be prepared to withhold resuscitate at (or below) a given gestational age if parents requested that palliative care be provided.

FIGURE 3

Upper threshold: cumulative probability of nonresuscitation based on the highest gestational age at which neonatologists would withhold resuscitation at parental request in an infant who was born in good condition. The figure indicates the proportion of respondents who would be prepared to withhold resuscitate at (or below) a given gestational age if parents requested that palliative care be provided.

Within countries, there was a wide range of variation in the prognosis that was judged to be too good to withhold resuscitation from an infant who was born in good condition, although there was not a significant difference between countries (Fig 2). The most frequent response was that resuscitation would be provided despite parental wishes if there was a >50% chance of survival without severe disability (44 of 162 [25%]). Eleven respondents (2 of 42 in the United Kingdom, 5 of 72 in Sweden, and 4 of 48 in the Netherlands; 7%) indicated that they would be prepared to withhold resuscitation at parental request at any level of prognosis.

Almost all of the respondents indicated that they had local or national guidelines relating to the resuscitation of preterm infants (157 of 162 [96.9%]). Most UK respondents (79.5%) indicated that according to this guideline, the lower threshold for resuscitation was 23 weeks’ gestation, although 17.9% of UK respondents indicated that it was 24 weeks’ gestation. Most Swedish respondents (78.9%) indicated that the guideline threshold was 22 weeks’ gestation, although 7% of Swedish respondents indicated that the threshold was 23 weeks’ gestation, and 7% of Swedish respondents indicated that according to the guideline, resuscitation should always be provided. All Dutch respondents indicated that their guideline specified a lower gestational age threshold for resuscitation of 24 weeks’ gestation.

Of respondents, 42 (26.6%) indicated that in the last 5 years, they had resuscitated infants below their local or national lower threshold. This was somewhat more common among Dutch and UK respondents compared with Swedish respondents (33.3%, 35.9%, and 16.9%, respectively). In free-text responses, many indicated that this was because infants were close to the gestational age threshold. For example, 1 Dutch respondent replied that the infant was “just a few hours before 24 weeks.”

Regarding guidelines on resuscitation, a smaller proportion of respondents (71.6% vs 96.9%; P < .001) indicated that they had local or national guidelines relating to nonresuscitation of preterm infants. A guideline was cited by a smaller proportion of Swedish respondents than UK and Dutch respondents (58.3%, 81%, and 83.3%, respectively; P < .01). Of those who indicated that there was such a guideline, there was a variation between countries and within countries in regard to where respondents thought this threshold lie (Supplemental Table 4; P < .001). Sixteen respondents (12.9%) indicated that they had withheld resuscitation in the last 5 years for infants more mature than this limit. Seven of these 16 respondents indicated in free-text responses that the infants concerned were extremely growth restricted.

In a multivariate regression analysis, the respondents’ country was significantly associated with the lower gestational age threshold for resuscitation (P < .01) but not the respondents’ professional role, sex, or experience. The importance of religion was associated with the lower gestational age threshold, with those respondents who indicated that religion was very important or fairly important having a higher gestational age threshold than those for whom religion was not very important (P < .01). The addition of other demographic variables to the regression model did not change the strength of association between country and the lower threshold.

The respondents’ country was also associated with the upper threshold for resuscitation (P < .01). In general, demographic factors, including respondents’ country, were not associated with the probability thresholds for resuscitation. Greater respondent experience appeared to be associated with the probability threshold for providing resuscitation, with respondents with >20 years of experience indicating a higher threshold for resuscitating than those with 1 to 5 years of experience (P < .01).

In our survey, we found striking differences in the thresholds for resuscitation of EPIs reported by neonatologists in 3 European countries previously reported to have similar approaches to ethical decision-making in neonatal intensive care.14,15 Those differences in resuscitation practices appeared to be heavily influenced by guidelines in the 3 countries and closely tied to completed weeks’ gestation. However, clinicians varied in their interpretation of the upper threshold (the gestational age beyond which resuscitation was mandatory). In addition, clinicians were influenced by an infant’s condition at birth: they indicated a higher threshold for resuscitating an infant who was born in poor condition. There was relatively little difference between countries in prognosis-based thresholds, although individual respondents varied in the prognosis that they believed would justify providing or withholding resuscitation.

The majority of neonatologists in our survey indicated that they would not resuscitate infants who were more premature than a gestational age of 22 + 0/7 weeks in Sweden, 23 + 0/7 weeks in the United Kingdom, and 24 + 0/7 weeks in the Netherlands. This appears to reflect current national guidance. The British Association of Perinatal Medicine published a framework document in 2009; it notes that it would be standard practice not to resuscitate at <23 + 0/7 weeks’ gestation.16 The Dutch Pediatric Society’s 2010 guideline advises active care measures for neonates from a gestational age of 24 + 0/7 weeks onward and not below that.17 In Sweden, a national guideline that was issued in 2016 recommends an antenatal transfer to a level-3 setting from 22 + 0/7 weeks’ gestation and consideration of antenatal steroids and resuscitation.18 

These national guidelines endorse shifts in management on the basis of gestational age. This might lead to changes in the permissibility of resuscitation from 1 day to the next, a phenomenon that could be compared to a “Cinderella effect,” (referencing the impact of the stroke of midnight in the Cinderella fairytale). In the 3 countries, there were clear increments in willingness to resuscitate at the boundary between 1 gestational age and the next (Figs 1 and 3). Of the surveyed cohort, 70% (113 of 162) elected to provide a whole-week cutoff for resuscitation. A minority of respondents indicated in free-text responses that they would occasionally resuscitate below their guideline limit for infants who were close to the week boundary. Of note, only 1 respondent in the cohort declined to provide a gestational age threshold, giving the following reason: “I’d like to consider more elements than gestational age only.” In a large multicenter study in the United States, there were significant increments in the rates of active resuscitation between the last day and the first day of an estimated gestational age,19 suggesting that completed weeks’ gestation do influence clinical management.

Our survey respondents indicated a gray zone for parental discretion in regard to resuscitation of ∼1 week. There were significant differences between countries in the highest gestational age at which nonresuscitation was regarded as permissible. However, within countries, respondents were also divided on where they would draw the line. Among Dutch respondents, 29% would withhold resuscitation at a maximum of 25 + 6/7 weeks’ gestation, whereas 33% would withhold resuscitation at up to 26 + 0/7 weeks’ gestation; a similar division was evident in Swedish respondents (21% indicating a maximum of up to 22 + 6/7 weeks’ gestation and 29% indicating a maximum of up to 23 + 0/7 weeks’ gestation) and UK respondents (29% indicating a maximum of up to 23 + 6/7 weeks’ gestation and 29% indicating a maximum of up to 24 + 0/7 weeks’ gestation).

In previous studies, authors have compared national guidelines for the care of EPIs and found variation in guidance at 23 and 24 weeks’ gestation but more common ground at 22 and 25 weeks’ gestation.4,6 There have been recent national surveys of approaches to neonatal resuscitation (Table 2). Of relevance, a study in Sweden that was undertaken in 200320 revealed a lower threshold of 23 + 0/7 weeks’ gestation, suggesting that practice in this country has shifted in the last decade.

TABLE 2

National or Regional Surveys of Neonatologists and Pediatricians in High-Income Countries That Were Performed After the Year 2000

LocationSurvey SampleYearLower Threshold (Median), wkUpper Threshold (Median), wk
United Kingdom (Southeast England)21  111 consultants and trainees 2008 23a N/A 
Ireland22  170 health professionals (obstetricians, neonatologists, and neonatal nurses) N/A 24a N/A 
Norway23  62 medical directors (obstetric and pediatric units) 2005 23a N/A 
Canada24,25  121 neonatologists 2004 N/A 24a 
Sweden20  88 neonatologists 2003 23 + 0/7 23 + 1/7–23 + 6/7 
United States     
 United States (national)26  637 members of the American Academy of Pediatrics perinatal section 2012 23a 25a 
 New Jersey27  20 NICU directors and associate directors 2006 23a 24a 
 Illinois28  85 neonatologists 2002 22a 25a 
 New England29  149 neonatologists 2002 23+0 24 + 0/7 
 Connecticut and Rhode Island30  48 neonatologists 2001 23a 23a 
LocationSurvey SampleYearLower Threshold (Median), wkUpper Threshold (Median), wk
United Kingdom (Southeast England)21  111 consultants and trainees 2008 23a N/A 
Ireland22  170 health professionals (obstetricians, neonatologists, and neonatal nurses) N/A 24a N/A 
Norway23  62 medical directors (obstetric and pediatric units) 2005 23a N/A 
Canada24,25  121 neonatologists 2004 N/A 24a 
Sweden20  88 neonatologists 2003 23 + 0/7 23 + 1/7–23 + 6/7 
United States     
 United States (national)26  637 members of the American Academy of Pediatrics perinatal section 2012 23a 25a 
 New Jersey27  20 NICU directors and associate directors 2006 23a 24a 
 Illinois28  85 neonatologists 2002 22a 25a 
 New England29  149 neonatologists 2002 23+0 24 + 0/7 
 Connecticut and Rhode Island30  48 neonatologists 2001 23a 23a 

Only surveys that included a range of gestational ages and asked about thresholds for resuscitation and nonresuscitation were included. From surveys, the median lower threshold was defined as the lowest gestational age at which 50% of neonatologists would actively resuscitate. The median upper threshold was defined as the highest gestational age at which 50% of neonatologists would withhold resuscitation. N/A, data not available.

a

Only whole-week options were included in the survey.

In an earlier international survey of neonatologists that was performed in 1999 and 2000, researchers compared 6 Pacific Rim countries.31 The lower threshold varied from 22 weeks’ gestation in Japan to 25 weeks’ gestation in Malaysia.31 EURONIC (European Project on Parents’ Information and Ethical Decision Making in Neonatal Intensive Care Units), a large European study that was conducted in 1996–1997, revealed significant differences in physicians’ overall approaches to end-of-life decision-making between different countries.14 In that study, physicians from the United Kingdom, the Netherlands, and Sweden appeared to apply similar values to decision-making14 and have a similar overall experience of making specific end-of-life decisions.15 However, the EURONIC study group did not specifically gauge thresholds for the resuscitation of EPIs.

A proportion of respondents in our survey indicated that they would adjust their threshold for resuscitation if an infant was born in poor condition (Fig 1). This was less evident among Dutch respondents and was less apparent for the upper threshold. Although previous surveys have also indicated that neonatologists consider condition at birth in decisions about providing intensive care,29,32 this has been criticized.32,33 Physician assessment of infant condition at the time of birth is variable34 and is a poor predictor of outcome for EPIs.32,34,35 

Gestational age–based thresholds have been criticized. One alternative would be to focus on the most ethically salient factor for decisions: an infants’ prognosis.13 To the best of our knowledge, our study is the first in which neonatal physicians’ views about prognosis-based thresholds for resuscitation are assessed. Half of our respondents would not be prepared to resuscitate at parental request if an infant had <5% to 10% chance of survival without severe morbidity. Half of our respondents indicated that they would not be prepared to withhold resuscitation if an infant had a >20% to 30% chance of survival without severe morbidity. However, 10% to 15% of respondents declined to provide a prognostic threshold for decisions. In free-text responses, several respondents mentioned that they did not consider decisions in this way.

One potential reason for the differences between countries in gestational age thresholds is a reflection of variation in the value placed on saving the life of a newborn or in the burden of feeling responsible for a newborn’s disabilities.36 There is some indication of this from our survey because among Swedish respondents (who as a group had the lowest thresholds for resuscitating EPIs), 33% indicated that they would consider providing resuscitation no matter how low the chance of an infant surviving was (compared with 21% and 6% among UK and Dutch respondents, respectively). In contrast, in the earlier EURONIC study, only 2% of Swedish neonatologists indicated a belief that every neonate should receive maximal treatment regardless of outcome.14 Differences in ethical values between the United Kingdom, Sweden, and the Netherlands also do not appear to fully explain the differences in gestational age thresholds between countries because the prognosis thresholds were similar overall (Fig 2). Physicians’ religious beliefs were mostly not associated with prognosis or gestational age thresholds (although more religious respondents appeared to indicate a higher gestational age threshold for resuscitation). Still, contextual differences in the 3 countries on how life and death is perceived could contribute to the differences found herein.

An alternative possible explanation for the difference in gestational age thresholds that was observed might be differences in mortality and morbidity for EPIs in the 3 countries. In Table 3, recent national cohort studies that include survival and severe disability (assessed at 2 years of age) are summarized. These studies reveal important differences in mortality rates, although not necessarily in rates of severe disability. For example, on the basis of these data, EPIs in Sweden at 23 weeks’ gestation appear to have a higher probability of survival without severe disability if they are resuscitated and admitted to the NICU than EPIs in the United Kingdom (52% compared with 22%). However, some of the differences in outcome could reflect the differences between countries in rates of provision of active perinatal management and active resuscitation and in subsequent limitation of treatment.37,38 This is evident in the differences in survival rates among live births and in the absence of survivors at 22 and 23 weeks’ gestation in the Netherlands. This makes it more challenging to evaluate the prognosis if full active management was provided.39 Also, of note, survival rates for infants at 24 and 25 weeks’ gestation appear slightly better in the Dutch cohort than in the UK EPICure study. Therefore, this cannot explain the greater willingness of Dutch neonatologists to withhold treatment at these gestational ages. Another possibility is that neonatologists in the 3 countries differ in their perception of the risk of severe impairment among survivors at low gestational ages. For example, in the Extremely Preterm Infants in Sweden Study, more active resuscitation was not associated with an increase in the rate of neurodevelopmental impairment.37 This result may have encouraged a more proactive approach in Sweden. In contrast, researchers in the UK EPICure study have reported a clear association between early gestational age and increased rates of severe impairment.40 This may discourage UK neonatologists from offering resuscitation to some EPIs.

TABLE 3

Outcome for EPIs in the United Kingdom, Sweden, and the Netherlands From Recent National Cohort Studies

LocationStudyOutcome22 wk, %23 wk, %24 wk, %25 wk, %
United Kingdom EPICure 2 study (England and Wales, 2006)40  Survival (live birth)a 19 40 65 
Survival (NICU)b 16 29 46 68 
Severe disabilityc 26d 26d 15 15 
Sweden EXPRESS (Swedish Perinatal Quality Register 2004–2007)41  Survival (live birth)a 10 53 67 82 
Survival (NICU)b 26 65 73 84 
Severe disabilityc 40 21 13 
Netherlands Zegers et al (Netherlands Perinatal Registry, 2007–2011)42  Survival (live birth)a n/a n/a 31 71 
Survival (NICU)b n/a n/a 56 73 
Severe disabilityc n/a n/a n/a n/a 
LocationStudyOutcome22 wk, %23 wk, %24 wk, %25 wk, %
United Kingdom EPICure 2 study (England and Wales, 2006)40  Survival (live birth)a 19 40 65 
Survival (NICU)b 16 29 46 68 
Severe disabilityc 26d 26d 15 15 
Sweden EXPRESS (Swedish Perinatal Quality Register 2004–2007)41  Survival (live birth)a 10 53 67 82 
Survival (NICU)b 26 65 73 84 
Severe disabilityc 40 21 13 
Netherlands Zegers et al (Netherlands Perinatal Registry, 2007–2011)42  Survival (live birth)a n/a n/a 31 71 
Survival (NICU)b n/a n/a 56 73 
Severe disabilityc n/a n/a n/a n/a 

EXPRESS, Extremely Preterm Infants in Sweden Study; n/a, not applicable.

a

Survival: proportion of live births.

b

Survival: proportion of NICU admissions.

c

Severe disability: proportion of those who were seen at follow-up with nonambulant cerebral palsy (Gross Motor Function Classification System levels 3–5), blindness, profound sensorineural hearing loss not improved by aids, or a developmental quotient <3 SDs below the mean for age.

d

Results were combined for 22 and 23 wk.

A third possible explanation for the differences in resuscitation decisions between countries might be on the basis of limited resources within a public health care system. However, the United Kingdom, Sweden, and the Netherlands all have universal public health care systems with similar spending on health care as a proportion of the gross domestic product.43 Lower rates of resuscitation at the lowest gestational ages could reflect a lower priority for treatment of this group of infants (or higher costs),44 although there was no evidence that this was influencing neonatal physicians’ decision-making in our survey. When asked to imagine a scenario with unlimited NICU resources, there was no change in the proportion of respondents who were prepared to provide resuscitation at 22 + 4/7 weeks’ gestation nor in the prognosis that was considered sufficient to offer resuscitation.

There are some limitations to the conclusions that can be drawn from our survey. The survey was piloted but not formally validated. We had modest response rates, albeit our overall results are consistent with average response rates for electronic surveys (34%).45 The results are not necessarily generalizable to all neonatologists in the 3 countries. Although respondents indicated particular gestational ages at which resuscitation would or would not be provided in a hypothetical case, actual clinical decisions may incorporate a wider range of factors into decision-making. Although we provided clinicians with the option to enter free-text responses, a future qualitative study might provide richer insight into the reasons why neonatologists hold particular views about the treatment of EPIs. We only surveyed neonatal physicians; the views of neonatal nurses, obstetricians, and parents are also of critical importance for decision-making and should be included in future research.

Our survey provides novel comparative data on the views of neonatologists in 3 European countries on the resuscitation of EPIs. Although limited by the response rate, the survey indicates significant differences in approach to resuscitation, with Sweden being the most proactive and the Netherlands being the least proactive. Most survey respondents indicated significant shifts in decision-making in regard to particular weeks’ gestation, implying that changes in the permissibility of resuscitation or nonresuscitation occur at the stroke of midnight, a phenomenon we have called the Cinderella effect. One alternative to gestational age thresholds would be to focus on an infant’s prognosis. This survey provides the first data on neonatologists’ prognostic thresholds for resuscitation. Of interest, despite the different approaches to decision-making in the 3 countries, there was relatively little difference between countries in the prognosis that was judged as sufficient to justify resuscitation or nonresuscitation. Between neonatologists, there was a wide range of different thresholds that were applied. Future work is needed to explore whether agreement is possible or desirable on prognostic thresholds for the resuscitation of EPIs.

     
  • EPI

    extremely preterm infant

  •  
  • EURONIC

    European Project on Parents’ Information and Ethical Decision Making in Neonatal Intensive Care Units

Dr Wilkinson conceptualized and designed the study, conducted the initial analyses, drafted the initial manuscript, and reviewed and revised the manuscript; Drs Verhagen and Johansson designed the survey, collected and analyzed data, and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: Dr Wilkinson was supported for this work by a grant from the Wellcome trust (WT106587/Z/14/Z).

1
Saigal
S
,
Doyle
LW
.
An overview of mortality and sequelae of preterm birth from infancy to adulthood.
Lancet
.
2008
;
371
(
9608
):
261
269
[PubMed]
2
Griffin
IJ
,
Lee
HC
,
Profit
J
,
Tancedi
DJ
.
The smallest of the small: short-term outcomes of profoundly growth restricted and profoundly low birth weight preterm infants.
J Perinatol
.
2015
;
35
(
7
):
503
510
[PubMed]
3
Wilkinson
D
. Who should decide for critically ill neonates and how? The grey zone in neonatal treatment decisions. In:
McDougall
R
,
Delany
C
,
Gillam
L
, eds.
When Doctors and Parents Disagree: Ethics, Paediatrics and the Zone of Parental Discretion
.
Sydney, Australia
:
The Federation Press
;
2016
4
Guillén
Ú
,
Weiss
EM
,
Munson
D
, et al
.
Guidelines for the management of extremely premature deliveries: a systematic review.
Pediatrics
.
2015
;
136
(
2
):
343
350
[PubMed]
5
Pignotti
MS
,
Donzelli
G
.
Perinatal care at the threshold of viability: an international comparison of practical guidelines for the treatment of extremely preterm births.
Pediatrics
.
2008
;
121
(
1
). Available at: www.pediatrics.org/cgi/content/full/121/1/e193
[PubMed]
6
Gallagher
K
,
Martin
J
,
Keller
M
,
Marlow
N
.
European variation in decision-making and parental involvement during preterm birth.
Arch Dis Child Fetal Neonatal Ed
.
2014
;
99
(
3
):
F245
F249
[PubMed]
7
Meadow
W
,
Lantos
J
.
Moral reflections on neonatal intensive care.
Pediatrics
.
2009
;
123
(
2
):
595
597
[PubMed]
8
Janvier
A
,
Barrington
KJ
,
Aziz
K
,
Lantos
J
.
Ethics ain’t easy: do we need simple rules for complicated ethical decisions?
Acta Paediatr
.
2008
;
97
(
4
):
402
406
[PubMed]
9
Janvier
A
,
Barrington
KJ
,
Aziz
K
, et al
.
CPS position statement for prenatal counselling before a premature birth: simple rules for complicated decisions.
Paediatr Child Health
.
2014
;
19
(
1
):
22
24
[PubMed]
10
Turillazzi
E
,
Fineschi
V
.
How old are you? Newborn gestational age discriminates neonatal resuscitation practices in the Italian debate.
BMC Med Ethics
.
2009
;
10
:
19
[PubMed]
11
Batton
D
.
Resuscitation of extremely low gestational age infants: an advisory committee’s dilemmas.
Acta Paediatr
.
2010
;
99
(
6
):
810
811
[PubMed]
12
Batton
DG
;
Committee on Fetus and Newborn
.
Clinical report–antenatal counseling regarding resuscitation at an extremely low gestational age.
Pediatrics
.
2009
;
124
(
1
):
422
427
[PubMed]
13
Gillam
L
,
Wilkinson
D
,
Xafis
V
,
Isaacs
D
.
Decision-making at the borderline of viability: who should decide and on what basis?
J Paediatr Child Health
.
2017
;
53
(
2
):
105
111
[PubMed]
14
Rebagliato
M
,
Cuttini
M
,
Broggin
L
, et al;
EURONIC Study Group (European Project on Parents’ Information and Ethical Decision Making in Neonatal Intensive Care Units)
.
Neonatal end-of-life decision making: physicians’ attitudes and relationship with self-reported practices in 10 European countries.
JAMA
.
2000
;
284
(
19
):
2451
2459
[PubMed]
15
Cuttini
M
,
Nadai
M
,
Kaminski
M
, et al;
EURONIC Study Group
.
End-of-life decisions in neonatal intensive care: physicians’ self-reported practices in seven European countries.
Lancet
.
2000
;
355
(
9221
):
2112
2118
[PubMed]
16
Wilkinson
AR
,
Ahluwalia
J
,
Cole
A
, et al
.
Management of babies born extremely preterm at less than 26 weeks of gestation: a framework for clinical practice at the time of birth.
Arch Dis Child Fetal Neonatal Ed
.
2009
;
94
(
1
):
F2
F5
[PubMed]
17
de Laat
MW
,
Wiegerinck
MM
,
Walther
FJ
, et al;
Nederlandse Vereniging voor Kindergeneeskunde
;
Nederlandse Vereniging voor Obstetrie en Gynaecologie
.
Practice guideline ‘perinatal management of extremely preterm delivery’[in Dutch].
Ned Tijdschr Geneeskd
.
2010
;
154
:
A2701
[PubMed]
18
Domellöf
M
,
Pettersson
K
.
Guidelines for threatening premature birth will provide better and more equal care [in Swedish].
Lakartidningen
.
2017
;
114
:
EEYI
19
Rysavy
MA
,
Li
L
,
Bell
EF
, et al;
Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network
.
Between-hospital variation in treatment and outcomes in extremely preterm infants.
N Engl J Med
.
2015
;
372
(
19
):
1801
1811
[PubMed]
20
Peerzada
JM
,
Schollin
J
,
Håkansson
S
.
Delivery room decision-making for extremely preterm infants in Sweden.
Pediatrics
.
2006
;
117
(
6
):
1988
1995
[PubMed]
21
Duffy
D
,
Reynolds
P
.
Babies born at the threshold of viability: attitudes of paediatric consultants and trainees in South East England.
Acta Paediatr
.
2011
;
100
(
1
):
42
46
[PubMed]
22
Khan
RA
,
Burgoyne
L
,
O’Connell
MP
,
Dempsey
EM
.
Resuscitation at the limits of viability–an Irish perspective.
Acta Paediatr
.
2009
;
98
(
9
):
1456
1460
[PubMed]
23
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
[PubMed]
24
Lavoie
PM
,
Keidar
Y
,
Albersheim
S
.
Attitudes of Canadian neonatologists in delivery room resuscitation of newborns at threshold of viability.
J Obstet Gynaecol Can
.
2007
;
29
(
9
):
719
725
[PubMed]
25
Albersheim
SG
,
Lavoie
PM
,
Keidar
YD
.
Do neonatologists limit parental decision-making authority? A Canadian perspective.
Early Hum Dev
.
2010
;
86
(
12
):
801
805
[PubMed]
26
Arzuaga
BH
,
Meadow
W
.
National variability in neonatal resuscitation practices at the limit of viability.
Am J Perinatol
.
2014
;
31
(
6
):
521
528
[PubMed]
27
Bhat
SB
,
Weinberger
B
,
Hanna
NN
.
Resuscitation at the limit of viability: trapped between a rock and a hard place.
J Neonatal Perinatal Med
.
2013
;
6
(
1
):
31
36
[PubMed]
28
Weiss
AR
,
Binns
HJ
,
Collins
JW
 Jr
,
deRegnier
RA
.
Decision-making in the delivery room: a survey of neonatologists.
J Perinatol
.
2007
;
27
(
12
):
754
760
[PubMed]
29
Peerzada
JM
,
Richardson
DK
,
Burns
JP
.
Delivery room decision-making at the threshold of viability.
J Pediatr
.
2004
;
145
(
4
):
492
498
[PubMed]
30
Cummings
B
,
McKay
K
,
Hussain
N
.
Neonatologists’ opinions regarding resuscitation of extremely premature infants in Connecticut and Rhode Island.
Conn Med
.
2002
;
66
(
12
):
733
738
[PubMed]
31
Martinez
AM
,
Partridge
JC
,
Yu
V
, et al
.
Physician counselling practices and decision-making for extremely preterm infants in the Pacific Rim.
J Paediatr Child Health
.
2005
;
41
(
4
):
209
214
[PubMed]
32
Singh
J
,
Fanaroff
J
,
Andrews
B
, et al
.
Resuscitation in the “gray zone” of viability: determining physician preferences and predicting infant outcomes.
Pediatrics
.
2007
;
120
(
3
):
519
526
[PubMed]
33
Meadow
W
,
Meadow
X
,
Tanz
RR
,
Lagatta
J
,
Lantos
J
.
The value of a trial of therapy - football as a ‘proof-of-concept’.
Acta Paediatr
.
2011
;
100
(
2
):
167
169
[PubMed]
34
Manley
BJ
,
Dawson
JA
,
Kamlin
CO
,
Donath
SM
,
Morley
CJ
,
Davis
PG
.
Clinical assessment of extremely premature infants in the delivery room is a poor predictor of survival.
Pediatrics
.
2010
;
125
(
3
). Available at: www.pediatrics.org/cgi/content/full/125/3/e559
[PubMed]
35
Lagatta
J
,
Yan
K
,
Hoffmann
R
.
The association between 5-min Apgar score and mortality disappears after 24 h at the borderline of viability.
Acta Paediatr
.
2012
;
101
(
6
):
e243
e247
[PubMed]
36
Willems
DL
,
Verhagen
AA
,
van Wijlick
E
;
Committee End-of-Life Decisions in Severely Ill Newborns of Royal Dutch Medical Association
.
Infants’ best interests in end-of-life care for newborns.
Pediatrics
.
2014
;
134
(
4
). Available at: www.pediatrics.org/cgi/content/full/134/4/e1163
[PubMed]
37
Serenius
F
,
Blennow
M
,
Maršál
K
,
Sjörs
G
,
Källen
K
;
EXPRESS Study Group
.
Intensity of perinatal care for extremely preterm infants: outcomes at 2.5 years.
Pediatrics
.
2015
;
135
(
5
). Available at: www.pediatrics.org/cgi/content/full/135/5/e1163
[PubMed]
38
Marlow
N
.
The elephant in the delivery room.
N Engl J Med
.
2015
;
372
(
19
):
1856
1857
[PubMed]
39
Wilkinson
D
.
The self-fulfilling prophecy in intensive care.
Theor Med Bioeth
.
2009
;
30
(
6
):
401
410
[PubMed]
40
Moore
T
,
Hennessy
EM
,
Myles
J
, et al
.
Neurological and developmental outcome in extremely preterm children born in England in 1995 and 2006: the EPICure studies.
BMJ
.
2012
;
345
:
e7961
[PubMed]
41
Serenius
F
,
Källén
K
,
Blennow
M
, et al;
EXPRESS Group
.
Neurodevelopmental outcome in extremely preterm infants at 2.5 years after active perinatal care in Sweden.
JAMA
.
2013
;
309
(
17
):
1810
1820
[PubMed]
42
Zegers
MJ
,
Hukkelhoven
CW
,
Uiterwaal
CS
,
Kollée
LA
,
Groenendaal
F
.
Changing Dutch approach and trends in short-term outcome of periviable preterms.
Arch Dis Child Fetal Neonatal Ed
.
2016
;
101
(
5
):
F391
F396
[PubMed]
43
Eurostat
. Healthcare expenditure statistics. 2017. Available at: http://ec.europa.eu/eurostat/statistics-explained/index.php/Healthcare_expenditure_statistics. Accessed May 23, 2017
44
Hayden
D
,
Wilkinson
D
.
Asymmetrical reasons, newborn infants, and resource allocation.
Am J Bioeth
.
2017
;
17
(
8
):
13
15
45
Shih
TH
,
Fan
X
.
Comparing response rates from web and mail surveys: a meta-analysis.
Field Methods
.
2008
;
20
(
3
):
249
271

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.

Supplementary data