Resuscitation of Asphyxiated Newborn Infants with Room Air Or Oxygen: an International Controlled Trial: the Resair 2 Study

Resuscitation of Asphyxiated Newborn Infants With Room Air or Oxygen: An International Controlled Trial: The Resair 2 Study

Ola Didrik Saugstad, MD*; Terje Rootwelt, MD*; and Odd Aalen, PhD‡

ABSTRACT. Objective. Birth asphyxia represents a ment in agreement with the new proposal of the US Food serious problem worldwide, resulting in ϳ1 million and Drug Administration’s rules governing investiga- deaths and an equal number of serious sequelae annu- tional drugs and medical devices to permit clinical re- ally. It is therefore important to develop new and better search on emergency care without the consent of sub- ways to treat asphyxia. Resuscitation after birth asphyxia jects. The protocol was approved by the ethical traditionally has been carried out with 100% oxygen, and committees at each participating center. Entry criterion most guidelines and textbooks recommend this; how- was apnea or gasping with heart rate <80 beats per ever, the scientific background for this has never been minute at birth necessitating resuscitation. Exclusion cri- established. On the contrary, theoretic considerations in- teria were birth weight <1000 g, lethal anomalies, hy- dicate that resuscitation with high oxygen concentrations drops, cyanotic congenital heart defects, and stillbirths. could have detrimental effects. We have performed a Primary outcome measures were death within 1 week series of animal studies as well as one pilot study indi- and/or presence of hypoxic–ischemic encephalopathy, cating that resuscitation can be performed with room air grade II or III, according to a modification of Sarnat and just as efficiently as with 100% oxygen. To test this more Sarnat. Secondary outcome measures were Apgar score at thoroughly, we organized a multicenter study and hy- 5 minutes, heart rate at 90 seconds, time to first breath, pothesized that room air is superior to 100% oxygen time to first cry, duration of resuscitation, arterial blood when asphyxiated newborn infants are resuscitated. gases and acid base status at 10 and 30 minutes of age, Methodology. In a prospective, international, con- and abnormal neurologic examination at 4 weeks. trolled multicenter study including 11 centers from six The existing routines for resuscitation in each partici- countries, asphyxiated newborn infants with birth pating unit were followed, and the ventilation tech- weight >999 g were allocated to resuscitation with either niques described by the American Heart Association room air or 100% oxygen. The study was not blinded, and were used as guidelines aiming at a frequency of manual the patients were allocated to one of the two treatment ventilation of 40 to 60 breaths per minute. groups according to date of birth. Those born on even Results. Forms for 703 enrolled infants from 11 cen- dates were resuscitated with room air and those born on ters were received by the steering committee. All 94 pa- odd dates with 100% oxygen. Informed consent was not tients from one of the centers were excluded because of obtained until after the initial resuscitation, an arrange- violation of the inclusion criteria in 86 of these. There- fore, the final number of infants enrolled in the study was 609 (from 10 centers), with 288 in the room air group From the *Department of Pediatric Research, National Hospital, Oslo, Nor- and 321 in the oxygen group. way, and the ‡Section of Medical Statistics, University of Oslo, Oslo, Nor- Median (5 to 95 percentile) gestational ages were 38 way. (32.0 to 42.0) and 38 (31.1 to 41.5) weeks (NS), and birth Resuscitation of Asphyxiated Newborn Infants With Room Air or Oxygen: weights were 2600 (1320 to 4078) g and 2560 (1303 to 3900) The Second Multicenter Study (Resair 2) study group participants: S. Ramji, g (NS) in the room air and oxygen groups, respectively. MD, Department of Pediatrics, Maulana Azad Medical College, New Delhi, There were 46% girls in the room air and 41% in the India; S. F. Irani, MD, Department of Pediatrics, King Edward Memorial Hospital, Bombay, India; S. Jayam, MD, Kasturba Hospital for Women and oxygen group (NS). Mortality in the first 7 days of life Children, Madras, India; S. El-Meneza, MD, Faculty of Medicine for Girls, was 12.2% and 15.0% in the room air and oxygen groups, with 95% 0.82 ؍ (Al-Azhar University, Cairo, Egypt; A. Narang, MD, Department of Pediat- respectively; adjusted odds ratio (OR Neonatal mortality .1.35–0.50 ؍ (rics, Post Graduate Institute of Medical Education and Research, Chandi- confidence intervals (CI ؍ with 95% CI 0.72 ؍ garh, India; M. Khasaba, MD, Department of Pediatrics, Mansoura, Egypt; was 13.9% and 19.0%; adjusted OR S. Sallab, MD, Department of Pediatrics, Mansoura, Egypt; E. A. Hernandez, 0.45–1.15. Death within 7 days of life and/or moderate or MD, Department of Pediatrics, Santo Thomas University Hospital, Manila, severe hypoxic–ischemic encephalopathy (primary out- Philippines; T. Talvik, MD, Children’s Hospital, Tartu University, Tartu, come measure) was seen in 21.2% in the room air group with 95% 0.94 ؍ Estonia; P. Ilves, MD, Children’s Hospital, Tartu University, Tartu, Estonia; and in 23.7% in the oxygen group; OR ؍ ,G. Samy Aly, PhD, Institute of Childhood Studies, Ain Shams University Cairo, Egypt; M. Vento, MD, Department of Pediatrics, Hospital Catolico, CI 0.63–1.40. “Casa de Salud” De Santa Ana, Valencia, Spain; F. Garcia-Sala, Department Heart rates did not differ between the two groups at of Pediatrics, Hospital Catolico, “Casa de Salud” De Santa Ana, Valencia, any time point and were (mean ؎ SD) 90 ؎ 31 versus Spain; R. Solberg, MD, Department of Pediatrics, Vestfold County Central 93 ؎ 33 beats per minute at 1 minute and 110 ؎ 27 versus Hospital, Tønsberg, Norway. 113 ؎ 30 beats per minute at 90 seconds in the room air Resair 2 steering committee members: O. D. Saugstad, MD (Principal In- and oxygen groups, respectively. vestigator), T. Rootwelt, MD, O. Aalen, PhD, Department of Pediatric Apgar scores at 1 minute (median and 5 to 95 percen- Research, The National Hospital and Institute for Medical Statistics, Uni- tiles) were significantly higher in the room air group (5 [1 versity of Oslo, Oslo, Norway. to 6.7]) than in the oxygen group (4 [1 to 7]); however, at Received for publication Jan 20, 1998; accepted Apr 3, 1998. Reprint requests to (O.D.S.) Department of Pediatric Research, Rikshospi- 5 minutes there were no significant differences, with 8 (4 talet, 0027 Oslo, Norway. to 9) versus 7 (3 to 9). There were significantly more PEDIATRICS (ISSN 0031 4005). Copyright © 1998 by the American Acad- infants with very low 1-minute Apgar scores (<4) in the emy of Pediatrics. oxygen group (44.4%) than in the room air group (32.3%). http://www.pediatrics.org/cgi/content/full/102/1/Downloaded from www.aappublications.org/newse1 by guestPEDIATRICS on October 1, Vol. 2021 102 No. 1 July 1998 1of7 There also were significantly more infants with 5-minute We were able to demonstrate that room air was as Apgar score <7 in the oxygen group (31.8%) than in the efficient as 100% O2 for resuscitation with regard to room air group (24.8%). There were no differences in acid normalization of heart rate, time to first breath, and base status or SaO2 during the observation period be- acid base status during the first 30 minutes of life. tween the two groups. Mean (SD) PaO2 was 31 (17) versus Apgar scores at 5 minutes were significantly higher 30 (22) mm Hg in cord blood in the room air and oxygen in the room air group. There were no significant groups, respectively (NS). At 10 minutes PaO2 was 76 (32) versus 87 (49) mm Hg (NS), and at 30 minutes, the values differences between the groups for neurologic symp- were 74 (29) versus 89 (42) mm Hg in the room air and toms or survival during the first week of life. The oxygen groups, respectively. small number of infants in that study precluded Median (95% CI) time to first breath was 1.1 (1.0–1.2) more extensive statistical analyses. Therefore, to test minutes in the room air group versus 1.5 (1.4 to 1.6) our hypothesis further, a larger sample size was minutes in the oxygen group. Time to the first cry also needed. Accordingly, a multicenter study was initi- was in mean 0.4 minute shorter in the room air group ated; the results are reported in this article. compared with the oxygen group. In the room air group, there were 25.7% so-called resuscitation failures (brady- METHODS cardia and/or central cyanosis after 90 seconds) that were The protocol was approved by the ethical committee for human switched to 100% oxygen after 90 seconds. The percent- investigation at each participating center. For practical reasons, age of resuscitation failures in the oxygen group was informed consent could not be obtained before enrollment. After 29.8%. resuscitation, informed consent for continued inclusion in the Conclusions. This study with patients enrolled pri- planned follow-up study was obtained from the parents. This marily from developing countries indicates that asphyx- arrangement was approved by the ethical committees and is in iated newborn infants can be resuscitated with room air agreement with the consensus statement from the coalition con- as efficiently as with pure oxygen. In fact, time to first ference of acute resuscitation and critical care researchers, and the breath and first cry was significantly shorter in room air- new proposal of the US Food and Drug Administration’s rules versus oxygen-resuscitated infants. Resuscitation with governing investigational drugs and medical devices to permit clinical research on emergency care without the consent of the 100% oxygen may depress ventilation and therefore de- subjects.8,9 lay the first breath. More studies are needed confirming these results before resuscitation guidelines are changed. Hypothesis and Study Organization Pediatrics 1998;102(1). URL: http://www.pediatrics.org/ cgi/content/full/102/1/e1; asphyxia, newborn infant, room The hypothesis of the study was that room air is superior to 100% oxygen when asphyxiated newborn infants are resuscitated. air, resuscitation, oxygen. The study was organized as an international multicenter trial with 11 participating centers from six countries. The steering committee was located in Oslo, Norway. There were one or two contact ABBREVIATIONS. HIE, hypoxic–ischemic encephalopathy; CI, persons responsible for the organization at each participating confidence interval; ANOVA, analysis of variance; OR, odds ratio. center. None of the participants had any conflicts of interest. Based on the sample size calculation and the approximate incidences of birth asphyxia at the participating centers, the nec- irth asphyxia continues to represent a major essary period for recruiting patients was estimated to be 18 clinical problem, and worldwide ϳ4 million months. However, it was decided before the study started that the newborn infants are affected annually. It has enrollment period should not exceed 2 years. Patients were en- B rolled from June 1, 1994, to May 31, 1996; some of the centers been estimated that of these, 1 million die, and an approximately equal number develop sequelae such joined the study after it had started. 1 as cerebral palsy, mental retardation, and epilepsy. Criteria for Inclusion and Exclusion Although birth asphyxia is a more serious problem ϳ The entry criterion for enrollment in the study was apnea or in developing countries, still 6/1000 newborn gasping with heart rate Ͻ80 beats per minute at birth necessitating infants in developed countries develop hypoxic– resuscitation. Exclusion criteria were birth weight Ͻ1000 g, lethal ischemic encephalopathy (HIE) after birth asphyxia,2 anomalies, hydrops, cyanotic congenital heart defects, and still- and ϳ25% of the moderate and 75% to 100% of the births. A stillbirth was diagnosed when a heart rate was never severe cases later have major neurodevelopmental established. 3 sequelae. Therefore it is important to develop new Treatment Allocation and Guidelines for Management and more effective ways to prevent and treat For practical reasons, the study was not blinded. In addition, asphyxia. after thorough considerations it was decided that formal random- Traditionally, newborn infants have been resusci- ization was not feasible in this study. Because resuscitation is a tated with 100% oxygen. In fact, most textbooks and medical emergency requiring immediate treatment, we were con- guidelines in the field recommend the use of 80% to cerned that a formal randomization (with, for instance, sealed 4,5 envelopes) could have resulted in a delay in treatment. Such a 100% O2. Although neonatal resuscitation is fre- delay also may have resulted in a reduced number of enrolled quently performed, we have not been able to find infants, especially the most depressed infants, possibly giving a any scientific basis for this recommendation. The nonrepresentative sample. As in the pilot study,7 the infants were only related experimental study took place in the therefore placed in one of the two treatment groups according to 1960s, and this study indicates that the outcome in date of birth. Those born on even dates were resuscitated with room air (room air group), and those born on odd dates were newborn rabbits resuscitated with room air is com- resuscitated with 100% oxygen (oxygen group). 6 parable with the outcome with 100% O2. The existing routines for resuscitation in each participating unit To investigate whether newborn infants can be were followed. A face mask and bag were used, and the infants resuscitated with room air, we performed a series of were intubated endotracheally when necessary. The ventilation techniques described by the American Heart Association were animal experiments before we carried out a clinical used as guidelines4 aiming at a frequency of manual ventilation of pilot study in which we resuscitated 42 asphyxiated 40 to 60 breaths per minute. At each center, at least two trained 7 newborn infants with room air and 42 with 100% O2. people involved in the study took part in the resuscitation of all

2of7 RESUSCITATION OFDownloaded ASPHYXIATED from www.aappublications.org/news NEWBORNS WITH by ROOM guest on AIR October OR 1, OXYGEN 2021 infants in the study. Resuscitation started immediately after de- dates given room air) because of a mistake concern- livery of the infant when a stop watch was started by one of the ing the date in 15 infants, in one case, oxygen was not members of the resuscitation team. To reduce any incremental risk of this study to a minimum as available and the infant was resuscitated with room in the pilot study, if an infant treated with room air did not air instead. These 16 infants were included in the respond adequately to resuscitation within 90 seconds after deliv- data analysis in the group in which they had been ery, the infant was switched to 100% oxygen. This was designated treated. In addition, four enrolled infants who died treatment failure. Criteria for treatment failure were heart rate were included, although they had potentially lethal Ͻ80 beats per minute and/or central cyanosis at 90 seconds after delivery, and they were recorded for both groups. Treatment congenital conditions that were not recognized until failures in the room air group, although switched to 100% oxygen resuscitation had been attempted. Two of these in- after 90 seconds, were analyzed in the room air group. fants had diaphragmatic herniae, one had esopha- The duration of resuscitation was measured from delivery until geal atresia with tracheoesophageal fistula, and one the infant had spontaneous breathing with a heart rate Ͼ100 beats per minute, as suggested by the American Heart Association.4 had congenital syphilis; all belonged to the room air Resuscitation was withdrawn after 30 to 45 minutes if spontane- group. Therefore, a total of 609 infants were eligible ous breathing had not been established. for analysis (Fig 1). Table 1 shows the distribution of these 609 infants Outcome Measures in the 10 remaining centers. A total of 107 infants The primary outcome measures were death within 1 week eligible for enrollment in the recruitment period at and/or presence of HIE, grade II or III, according to a modification each participating center were not included in the 2,10 of Sarnat and Sarnat. According to this classification, HIE grade study, for varying reasons. In some cases, the study I (mild) includes irritability, hyperalertness, mild hypotonia, and poor sucking; grade II (moderate) includes lethargy, seizures, team arrived after the resuscitation was started, and marked abnormalities of tone, and requirement of tube feeding; in others, the obstetricians did not want to have the and grade III (severe) includes coma, prolonged seizures, severe infant enrolled. The outcome of these 107 infants was hypotonia, and failure to maintain spontaneous respiration. We not recorded. also registered survival after 28 days of life. Secondary outcome measures were Apgar scores at 5 minutes, Nineteen (6.6%) and 17 (5.3%) parents in the room heart rate at 90 seconds, time to first breath, time to first cry, air and oxygen groups, respectively, did not give duration of resuscitation, arterial blood gases and acid base status informed consent. These patients consequently were at 10 and 30 minutes of age, and abnormal neurologic examination not enrolled in the follow-up study, which is under- after 4 weeks, In addition, arterial oxygen saturation was recorded way. in some infants by pulse oximetry at 1, 3, 5, and 10 minutes of life. Time to reach arterial oxygen saturation of 75% also was recorded if possible. Umbilical cord blood was sampled in eight centers Baseline Characteristics (time 0), and four obtained arterial blood (three obtained blood by Of the 609 infants enrolled, 288 (47.3%) were re- needle stick and one by catheterization) and four obtained venous suscitated with room air and 321 (52.7%) with oxy- cord blood. gen. This should be related to the fact that there were Sample Size and Statistical Analysis 49.0% even dates and 51.0% odd dates in the enrollment period. Table 2 shows some important pre- and Based on the findings in the pilot study,7 to demonstrate a significant reduction in mortality and/or moderate to severe HIE perinatal factors. No significant differences between from 15% to 9%, we needed to enroll 920 infants. A reduction from the two groups were found. 15% to 8% would, however, require 648 enrolled infants (signifi- There were 133 females in both the room air group cance level 0.05; power 80%). (46.2%) and the oxygen group (41.4%) (NS). Gesta- All calculations were performed with SPSS for Windows 6.1.2 (SPSS Inc, Chicago, IL), and graphs produced by Microsoft Power tional ages ranged from 27 to 44 weeks in both Point Version 4.0 (Seattle, WA), GraphPad Prism (San Diego, CA), groups. Median gestational age (5 to 95 percentile) or SPSS. Mean and SD units are given, and parametric tests was 38.0 (32.0 to 42.0) weeks in the room air group (two-tailed t test) were used when the variables were approxi- and 38.0 (31.1 to 41.5) weeks in the oxygen group mately normally distributed. Median and 95% confidence inter- (NS). Birth weights ranged between 1000 and 5550 g vals (CIs) or 5 and 95 percentiles are given, and nonparametric tests (two-tailed Mann–Whitney U test) were used when the dis- in the room air group and between 1000 and 5270 g tribution of the variable was not normal. Repeated-measures anal- in the oxygen group. Median birth weights (5 to 95 ysis of variance (ANOVA) and ␹2 tests were used when appropri- ate. Logistic or Cox regression analysis was used when the comparison between the two groups was adjusted for gender, gestational age, and birth weight. Interim analyses were performed after 150 and 300 enrolled infants. Because there were no significant differences between the two groups with regard to primary outcome measure or mortality in the first week or within 28 days of life, the study continued.

RESULTS Patients Forms for 703 enrolled infants from 11 centers were received by the steering committee. Eighty-six of 94 infants from one center had been included without meeting the inclusion criteria for resuscitation, and all 94 infants from that center were excluded by the steering committee. Of the remaining infants, 16 had been allocated to the wrong group (10 born on even dates given oxygen and 6 born on odd Fig 1. Trial profile of the 609 infants meeting inclusion criteria.

Downloaded from www.aappublications.org/newshttp://www.pediatrics.org/cgi/content/full/102/1/ by guest on October 1, 2021 e1 3of7 TABLE 1. Eligible and Included Infants According to City lower in the room air group, with odds ratio (OR) ϭ 0.69. However, with a logistic regression multivari- Center Eligible Included RA/O2 ate analysis correcting for gender, birth weight, and Delhi 158 158 64/94 Bombay 132 123 52/71 gestational age, this difference was not significant. Madras 100 100 54/46 HIE grade II or III was not different between the two Cairo 87 87 47/40 groups. Chandigarh 47 47 19/28 Mansoura 34 34 21/13 Heart Rate and Apgar Scores Manila 110 26 16/10 Heart rates (Fig 2) in the room air and oxygen Tartu 30 26 10/16 Ϯ Ϯ Valencia 13 6 4/2 groups at 1 minute were (mean SD) 90 31 versus Tønsberg 6 2 1/1 93 Ϯ 33 and at 90 seconds were 110 Ϯ 27 versus Total 716 609 288/321 113 Ϯ 30. There were no significant differences between the two groups in heart rate over the first 30 Abbreviations: RA, room air; O2, oxygen. minutes of life (repeated-measures ANOVA). The number of infants with heart rates Ͻ60, 80, or 100 TABLE 2. Maternal and Fetal Variables beats per minute did not differ between the two Maternal or Fetal Variable Room Air Oxygen groups at any time. Previous pregnancies mean (SD) 2.2 (1.5) 2.3 (1.6) Median (5 to 95 percentile) Apgar scores in the Previous deliveries, mean (SD) 1.8 (1.3) 1.8 (1.2) room air and oxygen groups, respectively, were 5 (1 Maternal anemia, n (%) 58/265 (21.9) 66/300 (22.0) to 6.7) and 4 (1 to 7) (P ϭ .004) at 1 minute, 8 (4 to 9) Preeclampsia, n (%) 52/280 (18.6) 61/312 (19.6) and7(3to9)(P ϭ .12) at 5 minutes, and 8 (5 to 10) Vaginal delivery, n (%) 169/286 (59.1)* 202/320 (63.1)** ϭ Sedation, n (%) 41/272 (16.8) 43/312 (13.8) and 8 (3.5 to 9) (P .29) at 10 minutes (Fig 3). There Pain relief, n (%) 119/270 (44.1) 132/305 (43.3) were significantly more infants with very low Fetal bradycardia, n (%) 104/273 (38.1) 118/301 (39.2) 1-minute Apgar scores (Ͻ4) in the oxygen group Premature, n (%) 75/288 (26.0) 72/321 (22.4) (n ϭ 142/320; 44.4%) than in the room air group (n ϭ BW Ͻ2500 g, n (%) 115/288 (39.9) 137/321 (42.7) ϭ Meconium, n (%) 119/280 (42.5) 140/316 (44.3) 92/286; 32.3%) (P .001). However, at 5 and 10 Intubated, n (%) 73/288 (25.3) 82/321 (25.5) minutes of age, this difference between the groups did not reach significance. There were significantly No significant differences between the groups were found. Ͻ Maternal anemia: hemoglobin Ͻ10 g/L. more infants with 5-minute Apgar scores 7inthe Preeclampsia: systolic blood pressure Ͼ140–160 mm Hg and pro- oxygen group (n ϭ 102/321; 31.8%) than in the room teinuria (Ͼ0.3 g/24 h, or positive urinary findings for protein) air group (n ϭ 71/286; 24.8%) (P ϭ .03); however, at after 20 weeks of pregnancy. 10 minutes, no such difference between the groups Breech position: *5.6%, **8.1%. Prematurity: Ͻ37 weeks gestational age. was found. Meconium indicates meconium-stained amniotic fluid. Blood Gases, Oxygen Saturation, and Acid Base Status Blood gases and acid base status are given in Table percentile) were 2600 (1320 to 4078) g and 2560 (1303 4. In umbilical cord blood and at 10 minutes of life, to 3900) g (NS) in the room air and oxygen groups, no differences in Pao2 between the groups were respectively. found; however, at 30 minutes, Pao2 was significantly higher in the oxygen group than in the room Primary Outcome Measure, Mortality, and HIE group [89 (42) mm Hg; n ϭ 156 vs 74 (29) mm Hg; ϭ ϭ None of the 33 infants recruited from the three n 115; P .002]. Paco2 was not significantly dif- European centers died. Mortality among the other ferent between the groups in umbilical cord blood, at centers did not vary significantly. Death within 7 10 minutes, or at 30 minutes of life. Base deficit and days (or before discharge if discharged before 7 pH did not differ between the groups either in um- days), death within 28 days, and death within 7 days bilical venous or arterial cord blood or in arterial or and/or HIE, grade II or III (primary outcome mea- capillary blood at 10 or 30 minutes of life. There was sure) are shown for all centers combined in Table 3. no difference in base deficit at any time point be- Mortality before 28 days was borderline significantly tween survivors and nonsurvivors.

TABLE 3. Total Number of Registered Infants With Adverse Outcome

Event RA O2 Univariate Multivariate OR 95% CI OR 95% CI Death within 7 da 35/288 (12.2%) 48/321 (15.0%) 0.79 0.49–1.26 0.82 0.50–1.35 Death within 28 da 40/288 (13.9%) 61/321 (19.0%) 0.69 0.44–1.06 0.72 0.45–1.15 Death within 28 db 40/267 (15.0%) 61/294 (20.7%) 0.67 0.43–1.04 0.71 0.44–1.14 Death within7dorHIEgrade II or IIIa 61/288 (21.2%) 76/321 (23.7%) 0.87 0.59–1.27 0.94 0.63–1.40 HIE grade II or IIIa 47/288 (16.3%) 55/321 (17.1%) 0.94 0.62–1.45 1.04 0.67–1.63 a ϭ ϭ N 609 (univariate) and N 589 (multivariate). Death number (percent) in room air (RA) and oxygen (O2) groups and OR for various adverse events when comparing the room air group versus the oxygen group. The multivariate analysis is adjusted for gender, gestational age, and birth weight. The 73 infants in the room air group that had been switched to oxygen after 90 seconds because of so-called resuscitation failure (see text) were analyzed in the room air group. b Compared with children followed for 28 days. N ϭ 561 (univariate) and N ϭ 543 (multivariate). At 28 days, 21 infants were lost to follow-up in the room air group and 27 infants in the oxygen group.

4of7 RESUSCITATION OFDownloaded ASPHYXIATED from www.aappublications.org/news NEWBORNS WITH by ROOM guest on AIR October OR 1, OXYGEN 2021 air and oxygen groups, respectively (P ϭ .004), and time to first cry was 1.6 (1.5 to 1.7) minutes in the room air group and 2.0 (1.8 to 2.2) minutes in the oxygen group (P ϭ .006). Median (95% CI) duration of resuscitation was 2.0 minutes in both groups (P ϭ .09); at 30 minutes, 38 infants were on artificial ventilation. Median (95% CI) time to reach Sao2 of 75% was 1.5 (1.4 to 1.6) minutes in the room air group (n ϭ 103) versus 2.5 (1.9 to 3.1) minutes in the oxygen group (n ϭ 109) (P ϭ .27). At each time point, more room air infants had taken their first breath than O2 infants (Fig 4). All P values are adjusted for gender, gestational age, and birth weight by Cox regression Fig 2. Heart rates between 1 and 30 minutes after birth in infants analysis. resuscitated with room air or 100% oxygen. Mean and SD are In the room air group, there were 73/284 (25.7%) given. Open bars indicate room air group; hatched bars, oxygen so-called treatment failures who were switched to group. No significant differences between the two groups were oxygen 90 seconds after delivery. However, if the found. same criteria (bradycardia and/or central cyanosis after 90 seconds) were applied to the oxygen group, the number was 88/295 (29.8%), not significantly different from that for the room air group. These infants had a high mortality both at 7 days (33% vs 28%; NS) and at 28 days of life (41% vs 35%; NS) in the room air and oxygen groups, respectively.

DISCUSSION In this prospective, international multicenter study, we have not been able to verify the hypothesis that room air is superior to 100% oxygen for resuscitation of newborn infants. However, the study indicates that asphyxiated newborn infants can be resuscitated just as well with ambient air as with 100% oxygen. In fact, uncorrected neonatal mortality tended to be lower in the room air group compared with the oxygen group (OR ϭ 0.69; 95 CI ϭ 0.44– 1.06). However, when correcting for gender, gestational age, and birth weight, this reduction in mortality was not significant. Fig 3. Apgar scores at 1, 5, and 10 minutes after birth in infants The Apgar scores at 1 minute were significantly resuscitated with room air or 100% oxygen. Median values and 5 lower, and the time to first cry and first breath were to 95 percentiles are given. Open bars indicate room air group; hatched bars indicate oxygen group. *P ϭ .004. significantly longer in the oxygen group, compared with the room air-resuscitated group. There also were more infants with low Apgar scores at 1 and 5 Arterial oxygen saturation was measured with minutes of age in the oxygen group, indicating some pulseoximetry in all centers at 1, 3, 5, and 10 minutes possible unknown adverse effects by using 100% of age, however, in only a limited number of pa- oxygen for resuscitation. tients, and it did not differ significantly between the The theoretic background for this study was based two groups (repeated-measures ANOVA). Mean on our findings from the mid-1970s when we dem- (SD) Sao2 was in the room air and oxygen group, onstrated that the purine metabolite hypoxanthine respectively, at 1 minute [65 (11)% n ϭ 75 versus 61 accumulates in body fluids and tissues during hyp- (14)% n ϭ 67]; 3 minutes [82 (13)% n ϭ 87 versus 81 oxia.11,12 Because hypoxanthine is a potential oxygen (11)% n ϭ 91]; 5 minutes [86 (10)% n ϭ 102 versus 88 radical generator, we suggested that the injury often (10)% n ϭ 103]; and 10 minutes [90 (6)% n ϭ 112 noted in the posthypoxic reoxygenation period could versus 91 (7)% n ϭ 120]. be caused by an explosive generation of oxygen rad- 11 Median (5 to 95 percentile) Fio2 at 10 minutes was icals by the hypoxanthine–xanthine oxidase system. 0.21 (0.21 to 1.00) in the room air group and 0.23 (0.21 This hypothesis could explain the well known oxy- to 1.00) in the oxygen group (P Ͻ .0001). However, at gen paradox, a phenomenon describing that injury 30 minutes, the values were not significantly differ- related to hypoxia occurs, to a large extent, in the ent at 0.21 (0.21 to 1.00) in both groups. reoxygenation period.12–14 Because it had been shown in vitro that the amount of oxygen radicals produced Time to First Breath, First Cry, and SaO2 of 75%, by the hypoxanthine–xanthine oxidase system in- Duration of Resuscitation, Treatment Failure creases with the oxygen concentration as well as with Median (95% CI) time to first breath was 1.1 (1.0 to the hypoxanthine concentration,15 we, therefore, sug- 1.2) minutes and 1.5 (1.4 to 1.6) minutes in the room gested that resuscitation should be performed with

Downloaded from www.aappublications.org/newshttp://www.pediatrics.org/cgi/content/full/102/1/ by guest on October 1, 2021 e1 5of7 TABLE 4. Blood Gases and Acid Base Status 0 Min 10 Min 30 Min Room Air Oxygen Room Air Oxygen Room Air Oxygen

a b a d e f pO2 mm Hg 31 (17) 30 (22) 76 (32) 87 (49) 74 (29) 89 (42) * b c d g h pCO2 mm Hg 50 (15) 50 (19) 39 (14) 38 (13) 32 (13) 30 (11) pH 7.11 (0.14)i 7.12 (0.18)j 7.17 (0.11)k 7.15 (0.13)l 7.25 (0.10)m 7.26 (0.12)n BD mmol/L 14.7 (6.1)o 14.2 (6.7)p 14.0 (5.0)q 14.5 (5.4)r 11.2 (5.2)s 12.1 (5.8)t a n ϭ 107, b n ϭ 127, c n ϭ 69, d n ϭ 80, e n ϭ 115, f n ϭ 156, g n ϭ 116, h n ϭ 157, i n ϭ 166, j n ϭ 197, k n ϭ 134, l n ϭ 135, m n ϭ 192, n n ϭ 219, o n ϭ 154, p n ϭ 177, q n ϭ 132, r n ϭ 134, s n ϭ 188, t n ϭ 207. Mean (SD) values are given. Po2 and Pco2 are from arterial blood. pH and base deficit (BD) at 0 min are from either umbilical vein or artery and at 10 and 30 minutes after birth from either capillary or artery. *P ϭ .002 compared with room air value.

adults that 100% O2 in some circumstances reduces oxygen consumption,22,23 and there are indications that the energy metabolism of the brain might be affected.20,24,25 In premature animals and human infants, an opposite effect has been found because hyperoxia has been shown to increase oxygen consumption.26 In full-term newborn infants at 1 to 2 days of life, it was found that after a few minutes of hyperoxia (100% O2), the work of breathing and the metabolic rate were increased. Immediately after the onset of hyperoxic breathing, there was a decrease in minute ventilation. This response was, however, short-lasting and quickly reversed into an increase in ventilation.26 In one study on adult dogs, a decreased neurologic survival was found when 100% oxygen versus room air was given after cerebral ischemia.27 Furthermore, a reduced cerebral blood flow was detected at 2 Fig 4. Kaplan–Meier plot showing the proportion that had not hours of age in preterm infants supplemented with taken the first breath in room air- and oxygen-resuscitated newborn infants. Time to first breath was significantly longer in the 80% oxygen versus room air immediately after 28 oxygen-resuscitated group compared with the room air-resusci- birth. The brain of the newborn infant may respond tated group. In the oxygen group, 60/313 (19.2%) required Ͼ3 differently to 100% oxygen than the brain of adult minutes to take the first breath compared with 28/284 (9.9%) in ϭ ϭ animals, and whether resuscitation with 100% oxy- the room air group (OR 0.47; 95% CI 0.29–0.76). gen might be harmful obviously needs further investigation. as low a concentration of supplemental oxygen as In this study, we found it necessary to use a possible.12 quasirandomized method of allocating infants to the In a series of experiments in young and newborn two treatment groups, namely according to odd or pigs that we have performed in recent years to test even dates of birth. This was necessary to simplify whether resuscitation can be carried out with room enrollment in the different centers because delay in air, we were able to demonstrate that room air resus- resuscitation could be detrimental to the infants. citation normalizes blood pressure, acid-base vari- From a statistical point of view, this form of system- ables, hypoxanthine in plasma and brain microdia- atic assignment should, however, represent a satis- lysate, cerebral blood flow, and somatosensory factory procedure. The use of odd and even dates evoked potentials as efficiently as resuscitation with and related types of allocation has been used previ- 16–20 100% O2. In addition, brain morphologic changes ously in a number of other works, many of them 4 days after the hypoxic insult were similar in both studies of emergency procedures.29–33 groups of piglets.17 Nonetheless, because the study was not strictly The present study showed a delay in the onset of randomized, an important question when evaluating respiration in infants resuscitated with 100% oxygen the significance of the present results is whether the compared with room air. This could be one explana- two groups were similar, or if there was evidence of tion for the lower 1-minute Apgar scores found in the bias. The resuscitation teams at each participating oxygen group. Resuscitation with 100% oxygen per- center did its utmost to avoid any bias in the recruit- haps may induce lung atelectasis or depress ventila- ment. For instance, the number of infants included in tion or other vital functions of the infants. In new- each group corresponds closely with the number of born hypopneic lambs ventilated with 100% oxygen, even and odd dates in the enrollment period. A the minute ventilation already was reduced after 45 completely balanced randomization would have re- seconds compared with room air-ventilated ani- sulted in 298 room air- and 311 oxygen-resuscitated mals;21 thus, it is possible that resuscitation with infants, which is close to the 288 and 321 who were oxygen depresses ventilation quickly. actually enrolled into each group. Also, the similarity It is known from animal and clinical studies in in birth weight, gestational age, and severity of as-

6of7 RESUSCITATION OFDownloaded ASPHYXIATED from www.aappublications.org/news NEWBORNS WITH by ROOM guest on AIR October OR 1, OXYGEN 2021 phyxia evaluated by acid-base variables in umbilical 7. Ramji S, Ahuja S, Thirupuram S, Rootwelt T, Rooth G, Saugstad OD. cord blood of the two groups indicates that there has Resuscitation of asphyxic newborn infants with room air or 100% oxygen. Pediatr Res. 1993;34:809–812 been no or only minor selection bias. Furthermore, 8. Biros MH, Lewis RJ, Olson CM, Runge JW, Cummins RO, Fost N. because the study was not blinded, a bias in treat- Informed consent in emergency research. JAMA. 1995;273:1283–1287 ment is possible. For instance, we did not record the 9. US Food and Drug Administration. Dept of Health and Human Ser- ventilation pressure used or the postresuscitation vices. Protection of human subjects: informed consent. Federal Register. 1995;60:49086-103 care, thus, differences in handling of the infants in 10. Sarnat HB, Sarnat MS. Neonatal encephalopathy following fetal dis- the two groups cannot be ruled out. tress. A clinical and electroencephalographic study. Arch Neurol. 1976; The centers from developed countries participat- 33:696–705 ing in this study enrolled few patients because of a 11. Saugstad OD. Hypoxanthine as a measurement of hypoxia. Pediatr Res. low number of births and a low incidence of birth 1975;9:158–161 12. Saugstad OD, Aasen AO. Plasma hypoxanthine levels as a prognostic asphyxia. Because the mortality in the three Euro- aid of tissue hypoxia. Eur Surg Res. 1980;12:123–129 pean centers was zero in this study, the high mortal- 13. Saugstad OD. Hypoxanthine as an indicator of hypoxia: its role in ity in our study therefore reflects the present situa- health and disease through free radical production. Pediatr Res. 1988;23: tion in many developing countries. Therefore, the 143–150 14. Saugstad OD. Oxygen toxicity in the neonatal period. Acta Paediatr results of this investigation cannot be extrapolated Scand. 1990;79:881–892 directly to units with a low mortality. However, we 15. Fridovich I. Quantitative aspects of the production of superoxide anion still believe the data also may have important impli- radical by milk xanthine oxidase. J Biol Chem. 1968;243:5753–5760 cations for developed countries. 16. Poulsen JP, Øyasæter S, Saugstad OD. Hypoxanthine, xanthine and uric Whether there are differences in long-term neuro- acid in newborn pigs during hypoxemia followed by resuscitation with room air or 100% oxygen. Crit Care Med. 1993;21:1058–1065 logic outcome cannot be answered in the present 17. Rootwelt T, Løberg EM, Moen A, Øyasæter S, Saugstad OD. Hypox- study, but a follow-up investigation of the enrolled emia and reoxygenation with 21% or 100% oxygen in newborn pigs: infants between 18 and 24 months of age is under- changes in blood pressure, base deficit, and hypoxanthine and brain way. morphology. Pediatr Res. 1992;32:107–113 In conclusion, this study has not demonstrated 18. Rootwelt T, Odden J-P, Hall C, Ganes T, Saugstad OD. Cerebral blood flow and evoked potentials during reoxygenation with 21% or 100% O2 significantly improved survival by using room air in newborn pigs. J Appl Physiol. 1993;75:2054–2060 instead of 100% O2 for newborn resuscitation. How- 19. Rootwelt T, Odden J-P, Hall C, Saugstad OD. Regional blood flow ever, it indicates that resuscitation of asphyxiated during severe hypoxemia and resuscitation with 21% or 100% O2 in newborn infants can be performed with room air just newborn pigs. J Perinat Med. 1996;24:227–236 20. Feet BA, Xiang-Qing Yu, Rootwelt T, Øyasæter S, Saugstad OD. Effects as efficiently as with 100% O2. Room air-resuscitated of hypoxemia and reoxygenation with 21% and 100% oxygen in new- infants also seem to recover more quickly than do born piglets: extracellular hypoxanthine in cerebral cortex and femoral infants resuscitated with 100% oxygen, as assessed muscle. Crit Care Med. 1997;25:1384–1391 by Apgar scores, time to first breath, and time to first 21. Hutchison AA. Recovery from hypopnea in preterm lambs: effects of cry. Although the present study has some limita- breathing air or oxygen. Pediatr Pulmonol. 1987;3:317–323 22. Lodato RF. Decreased O2 consumption and cardiac output during nor- tions, it represents a first step in the evaluation of mobaric hyperoxia in conscious dogs. J Appl Physiol. 1989;67:1551–1559 newborn resuscitation routines. This study does not 23. Reinhart K, Bloos F, Konig F, Bredle D, Hannemann L. Reversible justify changes of the present routines in itself; these decrease of oxygen consumption by hyperoxia. Chest. 1991;99:690–694 should not be changed until studies in both devel- 24. Goplerud JM, Kim S, Delivoria-Papadopoulos M. The effect of postasphyxial reoxygenation with 21% vs. 100% oxygen on Naϩ,Kϩ: oped and developing countries confirm the present ATPase activity in striatum of newborn piglets. Brain Res. 1995;696: results. Nevertheless, this investigation shows that 161–164 optimal resuscitation routines should be assessed 25. Huang CC, Yonetani M, Lajevardi N, Delivoria-Papadopoulos M, more carefully, perhaps also in developed countries. Wilson DF, Pastuszko A. Comparison of postasphyxial resuscitation with 100% and 21% oxygen on cortical oxygen pressure and striatal dopamine metabolism in newborn piglets. J Neurochem. 1995;64:292–298 ACKNOWLEDGMENTS 26. Mortola JP, Frapell PB, Dotta A, et al. Ventilatory and metabolic re- This study was funded by the Laerdal Foundation for Acute sponses to acute hyperoxia in newborn. Am Rev Respir Dis. 1992;146: Medicine and the Norwegian Council for Research. 11–15 Professor Henry Halliday and Professor Gosta Rooth kindly 27. Zwemer CF, Whitesall SE, D’Alecy LG. Cardiopulmonary–cerebral re- revised this manuscript. suscitation with 100% oxygen exacerbates neurological dysfunction following nine minutes of normothermic cardiac arrest in dogs. Resuscita- tion. REFERENCES 1994;27:159–170 28. Lundstrøm K, Pryds O, Greisen G. Oxygen at birth and prolonged 1. World Health Organization. Child Health and Development: Health of the cerebral vasoconstriction in preterm infants. Arch Dis Child. 1995;73: Newborn. Geneva, Switzerland: World Health Organization; 1991 F81–F86 2. Levene MI, Kornberg J, Williams THC. The incidence of and severity of 29. Linder N, Aranda JV, Tsur M, et al. Need for endotracheal intubation postasphyxial encephalopathy in full-term infants. Early Hum Dev. 1985; and suction in meconium-stained neonates. J Pediatr. 1988;112:613–615 11:21–8 30. Appleton R, Sweeney A, Choonara I, Robson J, Molyneux E. Lorazapam 3. Levene MI. Management and outcome of birth asphyxia. In: Levene MI, versus diazepam in the acute treatment of epileptic seizures and status Bennett MJ, Punt J, eds. Fetal and Neonatal Neurology and Neurosurgery. epilepticus. Dev Med Child Neurol. 1995;37:682–688 Edinburgh, Scotland: Churchill Livingstone; 1995:427–442 31. Kulkarni M, Elsner C, Ouellet D, Zeldin R. Heparinized saline versus 4. American Heart Association, Emergency Cardiac Care Committee and normal saline in maintaining patency of the radial artery catheter. Can Subcommittees. Guidelines for cardiopulmonary resuscitation and J Surg. 1994;37:37–42 emergency cardiac care. VII. Neonatal resuscitation. JAMA. 1992;268: 32. Rizzolo SJ, Piazza MR, Cotler JM, Hume EL, Cautili G, O’Neill DK. The 2276–2281 effect of torque pressure on halo pin complication rates. A randomized 5. Saugstad OD. Resuscitation of newborn infants; do we need new guide- prospective study. Spine. 1993;18:2163–2165 lines? Prenatal Neonatal Med. 1996;1:26–28 33. Long MN, Wickstrom G, Grimes A, Benton CF, Belcher B, Stamm AM. 6. Campbell AG, Cross KW, Dawes GS, Hyman AI. A comparison of air Prospective, randomized study of ventilator-associated pneumonia in

and O2 in hyperbaric chamber by positive pressure ventilation in resus- patients with one versus three ventilator circuit changes per week. Infect citation of newborn rabbits. J Pediatr. 1966;68:153–163 Control Hosp Epidemiol. 1996;17:14–19

Downloaded from www.aappublications.org/newshttp://www.pediatrics.org/cgi/content/full/102/1/ by guest on October 1, 2021 e1 7of7 Resuscitation of Asphyxiated Newborn Infants With Room Air or Oxygen: An International Controlled Trial: The Resair 2 Study Ola Didrik Saugstad, Terje Rootwelt and Odd Aalen Pediatrics 1998;102;e1 DOI: 10.1542/peds.102.1.e1

Updated Information & including high resolution figures, can be found at: Services http://pediatrics.aappublications.org/content/102/1/e1 References This article cites 31 articles, 1 of which you can access for free at: http://pediatrics.aappublications.org/content/102/1/e1#BIBL Subspecialty Collections This article, along with others on similar topics, appears in the following collection(s): For Your Benefit http://www.aappublications.org/cgi/collection/for_your_benefit Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: http://www.aappublications.org/site/misc/Permissions.xhtml Reprints Information about ordering reprints can be found online: http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news by guest on October 1, 2021 Resuscitation of Asphyxiated Newborn Infants With Room Air or Oxygen: An International Controlled Trial: The Resair 2 Study Ola Didrik Saugstad, Terje Rootwelt and Odd Aalen Pediatrics 1998;102;e1 DOI: 10.1542/peds.102.1.e1

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://pediatrics.aappublications.org/content/102/1/e1

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has been published continuously since 1948. Pediatrics is owned, published, and trademarked by the American Academy of Pediatrics, 345 Park Avenue, Itasca, Illinois, 60143. Copyright © 1998 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Downloaded from www.aappublications.org/news by guest on October 1, 2021