Pneumothorax in preterm babies has been associated with an increased risk of intraventricular hemorrhage (IVH)

Transcription

1 Identification of Pneumothorax in Very Preterm Infants Risha Bhatia, MB, MA, MRCPCH, Peter G. Davis, MD, FRACP, Lex W. Doyle, MD, FRACP, Connie Wong, RN, and Colin J. Morley, MD, FRACP, FRCPCH Objective To compare respiratory and other morbidities between very preterm infants with and without a pneumothorax and to determine whether infants at higher risk of pneumothorax can be identified early in their course. Study design Preterm infants at 23 to 28 weeks gestation with pneumothorax were compared with matched control subjects. Demographic and clinical data from birth through the first 72 hours were compared. Results Sixty-two (9.2%) of 675 infants had pneumothorax. There were no significant differences in the baseline maternal and infant characteristics. Mortality was significantly higher in the pneumothorax group (43%) versus control subjects (13%). There was no significant difference in continuous positive airway pressure or surfactant treatment or rates of intraventricular hemorrhage or bronchopulmonary dysplasia. Infants treated with early continuous positive airway pressure in the delivery room typically had pneumothorax on day 2 of life. Those who had pneumothorax had higher inspired fraction of oxygen before its diagnosis and over the first 12 hours of life than did control subjects. Conclusions Pneumothorax is associated with increased mortality and with severity of lung disease in the first day of life. It may be possible to identify babies at highest risk of pneumothorax on the basis of inspired fraction of oxygen in the first 12 hours of life. (J Pediatr 2011;159:115-20). Pneumothorax in preterm babies has been associated with an increased risk of intraventricular hemorrhage (IVH) 1 and death. 2 Although antenatal corticosteroids 3 and postnatal surfactant 4 have reduced the rate of pneumothorax, it remains an important complication, with reported rates of between 4% and 14% in the modern era. 5 The continuous positive airway pressure (CPAP) or Intubation (COIN) trial, 6 which randomly assigned preterm infants of between 25 and 28 weeks gestational age to CPAP or intubation in the delivery room (DR) reported an increased rate of pneumothorax in those babies who received CPAP as their initial mode of ventilation. However, an increase in pneumothorax was not observed in the CPAP arm of the recently reported SUPPORT trial. 7 Little is understood about neonatal pneumothoraces, when they occur, which babies are most susceptible to them, and their effect on the medium- and long-term outcomes of these babies. If the baseline risk variables were better understood, prevention may be possible. Therefore, the aims of this study were (1) to determine whether very preterm infants who are going to have a pneumothorax can be identified early; (2) to compare the timing of pneumothoraces in infants on CPAP compared with intubated infants; and (3) to compare respiratory and other morbidities between very preterm infants who do and do not have a pneumothorax. Methods This was a case-control study of preterm infants between 23 and 28 completed weeks gestation. Cases had development of a pneumothorax at The Royal Women s Hospital, Melbourne, Australia, between January 1, 2002, and June 30, The Royal Women s Hospital is a level III perinatal center with more than 5000 deliveries and 1400 admissions to the intensive care and special care nurseries per year. Infants less than 28 weeks completed gestational age who had a pneumothorax in the time period were identified and their case notes and chest radiographs (CXR) reviewed. A control subject was identified for each baby with a pneumothorax by taking the next baby born with a gestational age of 3 days and with a birth weight within 15% of the index case who did not have a pneumothorax. Infants were excluded if they had known congenital pulmonary malformations or if pneumothoraces occurred after surgery such as patent ductus arteriosus ligation. BPD CPAP CXR DR FiO 2 IVH ROC Bronchopulmonary dysplasia Continuous positive airway pressure Chest radiography Delivery room Inspired fraction of oxygen Intraventricular hemorrhage Receiver operating characteristic From the Neonatal Services (R.B., P.D., L.D., C.W., C.M.), The Royal Women s Hospital, Melbourne, Victoria, Australia; and the Department of Obstetrics and Gynecology (P.D., L.D., C.M.), University of Melbourne, Melbourne, Victoria, Australia Supported by the Australian National Health and Medical Research Council (Program Grant ). P.D. is supported in part by an Australian National Health and Medical Research Council Practitioner Fellowship. The authors declare no conflicts of interest /$ - see front matter. Copyright ª 2011 Mosby Inc. All rights reserved /j.jpeds

2 THE JOURNAL OF PEDIATRICS Vol. 159, No. 1 Demographic data included gestational age, sex, birth weight, mode of delivery, Apgar scores, multiple or singleton birth, type of anesthesia during delivery, duration of premature rupture of membranes, and details of antenatal corticosteroids or antibiotics. Details of the maximum respiratory support in the DR and inspired oxygen concentration (FiO 2 ) administered were recorded. The primary mode of respiratory support in the DR, the ventilatory variables, and the age of the baby at the time of an initial dose of surfactant were noted from nursing and medical notes and pharmacy records. Oxygen requirements and ventilatory modes were recorded at 4-hour intervals for the first 72 hours after birth. During the study epoch, all infants were resuscitated with 100% oxygen. FiO 2 was weaned during transport and arrival to the neonatal intensive care unit (NICU), where blended O 2 was available to target saturations of 88% to 94%. The need for intubation or CPAP was determined by the resuscitation team. The policy for very preterm infants receiving CPAP was to intubate and ventilate if there was an FiO 2 >0.6, apnea unresponsive to stimulation and methylxanthine treatment, an arterial ph of <7.25 with a PaCO 2 of >60 mm Hg, or metabolic acidosis unresponsive to treatment. Nasal CPAP was initially provided using the Dr ager Babylog 8000 plus ventilator (Dr agerwerk, Lubeck, Germany) in CPAP mode, at 5 to 8 cm H 2 O and a flow of 8 L/min. The ventilation mode used was assist control volume guarantee, using the same ventilator, at target tidal volumes of 3.5 to 6 ml/kg. The SensorMedics 3100A High Frequency Oscillatory Ventilator (SensorMedics, Yorba Linda, California) was occasionally used. Pneumothoraces were diagnosed either clinically (by transillumination) or on CXR. Details collected included the age at diagnosis, the maximal respiratory support at the time of diagnosis, and management of the pneumothorax. Follow-up data included survival, the duration of respiratory support and oxygen treatment, and rates of bronchopulmonary dysplasia (BPD) (defined as oxygen treatment at 36 weeks postmenstrual age), retinopathy of prematurity, necrotizing enterocolitis, and grade of IVH. The study was approved by the Royal Women s Hospital Research Committee. Statistical Analyses Data were compared using either the Pearson c 2 test or the Fisher exact test for categorical data as appropriate, t test for continuous variables, and the Mann-Whitney U test for nonparametric data or discrete variables. Receiver operating characteristic (ROC) curves were used to illustrate the relationship between FiO 2 requirement by age of the infant as a predictor of pneumothorax development. Infants who had pneumothoraces were compared with control subjects, and infants initially treated with CPAP were compared with those intubated in the DR. Stata v.10 for Windows (Stata Statistical Software, StataCorp LP, College Station, Texas) was used to analyze the data. A probability value of <.05 was considered significant. Results Over the 5½-year period, 675 babies were born between 23 and 28 completed weeks gestation; 62 (9.2%) infants had a pneumothorax. One (2%) infant was excluded because a matched control subject could not be identified. None had congenital pulmonary malformations. Data from 61 babies with a pneumothorax and 61 control subjects were analyzed. There were no substantial differences in the baseline maternal and infant characteristics of the infants in the pneumothorax and control groups as shown in Table I. There were no significant differences in the DR treatment between the groups, with 22 (36%) infants treated with CPAP and 38 (62%) with intubation and positive-pressure ventilation in both groups. One (2%) infant in each group had no respiratory support in the DR (Table II). All of the pneumothorax group were ultimately intubated; 52 (85%) before the diagnosis of pneumothorax, including 38 (62%) who were intubated in the DR, and 9 (15%) after the diagnosis of pneumothorax. Fifty-one (84%) of the control infants were intubated (P =.34). Pneumothoraces were diagnosed either by transillumination of the chest or by CXR at a median of 31 hours, with diagnostic CXR performed at a median of 32 hours (Table II). Thirty-six of 61 (59%) of the pneumothorax group and 34 of 61 (56%) of the control subjects received surfactant (Curosurf, Chiesi Pharmaceuticals Ltd, Parma, Italy) within 2 hours of birth (P =.71). Infants who had a pneumothorax received surfactant at a median of 0.92 hoursafterbirthcomparedwithcontrolsubjects,at0.62 hours (P =.05). Nineteen of 61 (31%) of the pneumothorax group had pulmonary interstitial emphysema on CXR compared with 5 of 60 (8%) of control subjects (P =.002). Table I. Baseline demographic characteristics Characteristic Pneumothorax (n = 61) Control subjects (n = 61) P value Maternal age* Gestational age* (wk) Birth weight* (g) Antenatal corticosteroid 54/60 (90) 56/60 (93).51 treatment, n (%) Antibiotics before 28/49 (57) 30/58 (52).58 delivery, n (%) Normal vaginal 25 (41) 29 (48).47 delivery, n (%) Prolonged ruptured 27 (45) 23 (38).42 membranes, n (%) Duration of ruptured 0 (0-205) 0 (0-10).15 membranes (h) Males, n (%) 34 (56) 28 (46).28 Singletons, n (%) 39 (64) 45 (74).24 Apgar score 1 min 5 (3-7) 6 (4-7).12 Apgar score 5 min 8 (6-9) 8 (8-9).36 Data were not available for all babies for antenatal steroid and antibiotic treatment. *Mean SD. Median (interquartile range). 116 Bhatia et al

3 July 2011 ORIGINAL ARTICLES Table II. Delivery room, subsequent treatment, and outcomes Characteristic Pneumothorax (n = 61) Control subjects (n = 61) P value Delivery room treatment CPAP, n (%) 22 (36.1) 22 (36.1) 1.0 Endotracheal tube, 38 (62.3) 38 (62.3) 1.0 n (%) No respiratory 1 (1.6) 1 (1.6) 1.0 support, n (%) Age at diagnosis of 31 ( ) N/A pneumothorax* (h) Age at intubation* (h) 0.17 ( ) 0.09 ( ) z.34 Intubated, n (%) 61 (100) 51 (84).34 Age at surfactant 0.92 ( ) x 0.62 ( ) k.05 treatment* (h) Surfactant <2 h of age, 36 (59) 34 (56).71 n (%) Age at first CXR* (h) 2.1 ( ) { 2.2 ( ) #.90 Age at diagnostic 32.2 ( )** N/A CXR* (h) PIE on CXR, n (%) 19 (31.2) 5 (8.3).002 Outcomes Mortality, n (%) 26 (43) 8 (13) <.001 BPD, n (%) 13/35 (37) 14/54 (26).26 Death or BPD, n (%) 39 (64) 22 (36).004 ROP grade 3 or 2/35 (6) 6/55 (11).55 worse, n (%) Any IVH, n (%) 23/56 (41) 19/60 (32).29 IVH grade 3 or 4, n (%) 12/56 (21) 5/60 (10).09 PIE, pulmonary interstitial emphysema; ROP, retinopathy of prematurity. *Median (interquartile range). n = 61; data were available for n patients. zn = 51; data were available for n patients. xn = 59; data were available for n patients. kn = 43; data were available for n patients. {n = 61; data were available for n patients. #n = 58; data were available for n patients. **n = 58; data were available for n patients. n = 60; data were available for n patients. The median (IQR) FiO 2 at 4-hour intervals, during the first 72 hours, is shown in Figure 1 for both groups. The median (IQR) maximum FiO 2 in the DR was 0.6 (0.3 to 1.0) in the pneumothorax group and 0.5 (0.3 to 0.6) in the control group (P =.23). The FiO 2 in the pneumothorax group was statistically significantly higher (P <.001) at each 4-hour interval, for the next 72 hours. Because we do not know when each pneumothorax actually occurred, we compared the FiO 2 at a mean (SD) 2.1 (1.1) hours before diagnosis with the FiO 2 of the control case at a similar time point using a paired t test in 54 of 61 pairs. Five pairs were not analyzed because these pneumothoraces were diagnosed after 72 hours of age and FiO 2 data were not collected. Paired data were missing for two cases. The mean FiO 2 for the pneumothorax group was significantly higher (mean, 0.59; SD, 0.3) than the control subjects (mean, 0.33; SD, 0.21) (mean difference, 0.26; 95% confidence interval, 0.19 to 0.32; P <.001). We also compared the FiO 2 at a mean (SD) of 6.2 (1.1) hours before diagnosis in 41 of 61 pairs. Again, there was a significantly higher FiO 2 between the pneumothorax group (mean, 0.39; SD, 0.18) and control subjects (mean, 0.28; SD, 0.14) (mean difference, 0.11; 95% CI, 0.04 to 0.19; P <.003). Twenty pairs could not be analyzed because in 14, FiO 2 2 to 6 hours before diagnosis was before birth, five were diagnosed after 72 hours when FiO 2 data were not collected, and, in one case, paired data were unavailable. ROC curves were constructed for all 4-hour intervals after birth. An ROC curve at 12 hours as a risk factor for pneumothoraces is shown in Figure 2 (available at because an elevated FiO 2 at 12 hours after birth was more strongly associated with a pneumothorax than the FiO 2 at any other time (Table III). The median (IQR) positive end-expiratory pressure or CPAP pressure at the time of diagnosis of a pneumothorax was6(5to7)cmh 2 O. The median (IQR) peak inspiratory pressure was 23 (19 to 27) cm H 2 O, and, for those infants treated with high-frequency oscillatory ventilation, the median (IQR) mean airway pressure was 15 (12.5 to 17.7) cm H 2 O. Tidal volume and blood gas data were not collected for this study. Mortality was significantly lower in the control group (13%) than the pneumothorax group (43%) (P <.001) (Table II). There were no significant differences in the incidence of IVH or BPD between the groups. There were no substantial differences between groups for causes of death (Table IV; available at Infants with Pneumothorax Who Were Started on CPAP in the DR Of the pneumothorax group, 22 of 61 (36%) received CPAP in the DR. These infants had a mean (SD) gestational age of 27.5 (1.5) weeks and a mean (SD) birth weight of 1053 (243) g. The median age when a pneumothorax was diagnosed was 31 hours. The median age at endotracheal intubation was 22 hours. Only 8 of 61 (13%) infants were receiving CPAP at the time of diagnosis of a pneumothorax. Details of the mode of ventilation at the time of diagnosis are shown in Table V (available at The first CXR were performed in all infants within 2 to 3 hours of birth and showed no evidence of pneumothoraces. Only 4 of 22 (18%) were intubated before the first CXR. Eighteen of 22 (82%) had radiological evidence of respiratory distress syndrome, and 3 of 22 (14%) CXR were normal. At the time the diagnostic CXR was performed, 20 of 22 (91%) infants were intubated. Seventeen of 22 (77%) survived. Three of the survivors (18%) had BPD. Three of 22 (14%) infants had grade 3 or 4 IVH, and all died. The median (IQR) FiO 2 for the infants initially treated with CPAP, over the first 72 hours, in the pneumothorax and control groups is shown in Figure 1. The median (IQR) FiO 2 in the DR was 0.3 (0.25 to 0.5) in the pneumothorax group and 0.4 (0.3 to 0.5) in the control group (P =.34). However, the differences in FiO 2 were significantly higher (P <.05) in the pneumothorax group at each 4-hour interval for the rest of the 72 hours. Identification of Pneumothorax in Very Preterm Infants 117

4 THE JOURNAL OF PEDIATRICS Vol. 159, No. 1 Infants with Pneumothorax Who Were Intubated in the DR Of the pneumothorax group, 38 of 61 (62%) were intubated and ventilated in the DR. These infants had a mean (SD) gestational age of 26.1 (1.6) weeks and a mean (SD) birth weight of 851 (215) g. The median time of diagnosis of pneumothorax was 22 hours. The first CXR was performed in all infants within 2 to 3 hours of birth. Eleven (29%) had pneumothoraces on the initial CXR. Thirty-six of 38 (95%) had radiological evidence of respiratory distress syndrome, and in 2 of 38 (5%), radiographs were normal. Seventeen of 38 (45%) had a chest drain in situ, and 21 of 38 (55%) did not. Only 1 of 38 (3%) was extubated and treated with CPAP at the time of diagnostic CXR. Seventeen of 38 (45%) survived. Ten of the survivors (59%) had development of BPD. Nine of 33 (27%) had grade 3and4IVH(Table V), and only 1 survived. No cranial ultrasound results were available for 5 of 38 infants. Discussion During the study period, 62 of 675 (9.2%) infants born at less than 29 weeks gestation had pneumothoraces, which is in keeping with rates of 7% to 12% reported in the United States 8,9 and the United Kingdom. 10 The 43% mortality rate in the infants with pneumothoraces compared with 13% in the control subjects highlights the fact that pneumothorax is a serious disease worthy of further investigation and prevention. The FiO 2 data suggest that infants who have a pneumothorax have worse lung disease than those who do not. There was no difference in the demographic characteristics of infants between the groups. Therefore, the only variable we identified to help determine which babies are most at risk of developing pneumothoraces appeared to be the FiO 2, especially FiO 2 at 12 hours of age. This may indicate more serious lung disease or may be due to an underlying pneumothorax. Distinguishing cause and effect is not possible from these data. Only infants with FiO 2 requirements >0.3 at 12 hours had a pneumothorax. These infants should be closely monitored for evidence of pneumothorax before acute deterioration occurs. A weakness of our study design is that we are unable to assess the effect of other important prognostic variables, such as gestational age, which have been used to match cases with control subjects. An alternative study design would have been to collect data on all subjects and then adjust for confounders with multivariable statistical techniques. We considered including more control subjects for each case but decided against this because the conclusions would have been unlikely to change and the gain in power would have diminished as each subsequent group of 61 control subjects was added. In addition, we would have had progressively more difficulty obtaining suitable control subjects. Several studies have suggested that infants treated with CPAP may have more pneumothoraces than those treated with head box oxygen or with positive-pressure Figure 1. A, FiO 2 over the first 72 hours in all infants in the pneumothorax group compared with all controls. B, FiO 2 over the first 72 hours in infants treated with CPAP only in the DR who had a pneumothorax compared with control subjects. Infants who had pneumothoraces appeared to have more significant lung disease as reflected by the increased oxygen requirements before diagnosis of the pneumothorax and for the first 72 hours from birth. 118 Bhatia et al

5 July 2011 ORIGINAL ARTICLES Table III. Receiver operating characteristic area under the curve for FiO 2 at 4-hour intervals during the first 24 hours in all infants and in those infants who commenced CPAP in the delivery room Time after birth (h) All infants Area under curve (95% CI)* CPAP in DR group Area under curve (95% CI) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) *CI not encompassing 0.5 (the null value) is statistically significant. ventilation. 14 The use of CPAP from birth in very low weight birth infants reduced the rate of intubation, as shown in the COIN trial 6 and more recently in the SUPPORT trial. 7 The COIN trial had a 9% pneumothorax rate in infants treated with CPAP in the DR and 3.1% in the intubated group. 6 However, the SUPPORT trial, which randomly assigned 1316 preterm infants between 24 and 28 weeks gestation to CPAP or intubation and surfactant treatment in the DR had an almost identical incidence of pneumothoraces in both groups; 6.8% in the CPAP group and 7.4% in the intubated and surfactant group. 7 Surfactant has been associated with a reduction in pneumothorax rates from about 20% to 10% in trials that are now about 20 years old. Our data, on contemporary very preterm infants, show that similar proportions of infants in the pneumothorax and control groups received surfactant within 2 hours of age, although it was administered at a median 20 minutes later in those infants who had pneumothoraces. We did not design this study to examine the effect of timing of surfactant and our finding that the pneumothorax group received surfactant 20 minutes later than control subjects is interesting but inconclusive. The rate of pneumothoraces in the recent CURPAP trial 19 was 6.7% in the group given surfactant within 30 minutes of birth versus 1.0% in the group that was treated with surfactant at a median of 4 hours. Conventional thinking is that pneumothoraces develop rapidly, and the diagnosis is made when the infant has decompensated enough such that there is a new respiratory problem that warrants investigation. The diagnosis of a pneumothorax depends on clinical suspicion, transillumination of the chest, and a CXR. It has been suggested that the time between onset of pneumothorax and clinical diagnosis is 1 to 10 hours. 20 This is not possible to corroborate from our data. All we can say is that of infants who had pneumothoraces, only 18% had one on their first CXR at about 2 hours. All these infants were ventilated in the DR, suggesting they had worse lung disease, although the resuscitation may have caused the pneumothorax. Forty-eight percent of infants with a pneumothorax already had a chest drain in situ when the diagnostic CXR was taken. It is unclear how many of these pneumothoraces were correctly diagnosed by clinical examination and transillumination. Not surprisingly, babies who had pneumothoraces were of significantly lower gestational age and birth weight if they were intubated and ventilated in the DR as compared with those who received CPAP in the DR. In babies started on CPAP, our data suggest that pneumothoraces develop in the second day, later than in those that are initially ventilated and receiving early surfactant. McIntosh et al 20 have described the use of transcutaneous carbon dioxide trend monitoring to facilitate early diagnosis. This may be impractical in very immature infants because the heated electrodes increase the risk of burns in these infants. It is essential to find new methods of assessing the risk of a pneumothorax or enabling early detection. Pneumothorax detection using pulmonary acoustic transmission 21 or computerized analysis of breath sounds 22 has been investigated in animal models and found to be promising, but neither have been studied in infants. Another method that may be useful is electrical impedance tomography, which continuously monitors regional lung impedance changes that correlate with intrathoracic changes in air content and ventilation. 23 Animal studies suggest that electrical impedance tomography may be useful for the detection and monitoring of pneumothoraces Previous studies have shown that pneumothoraces are associated with increased morbidity. 27 Hemodynamic changes at the time of pneumothorax are thought to lead to an abrupt increase in cerebral blood flow at the time of pneumothorax with the development of IVH soon after. 1,28 It is reassuring to note that our pneumothorax group did not have significantly higher rates of IVH than the control infants. IVH rates between infants who were intubated and ventilated or who received CPAP in the DR and had pneumothoraces were not significantly different. However, this study was not powered to look for significant differences in the rates of IVH, and therefore it is not possible to be confident about the effects of pneumothoraces on IVH. In conclusion, pneumothoraces significantly increase mortality in very preterm babies. Infants who have an increased FiO 2 in the first few hours are at increased risk of a pneumothorax and should be monitored closely. It was not possible for us to determine exactly when a pneumothorax occurred. Further research is needed to determine the etiology and timing of pneumothoraces before we can improve prevention. n We thank Drs Jennifer Dawson and Marta Thio and Ms Brenda Argus for statistical advice and editing the manuscript. Submitted for publication May 25, 2010; last revision received Nov 11, 2010; accepted Dec 13, Reprint requests: Dr Risha Bhatia, MB, MA, MRCPCH, Newborn Research, The Royal Women s Hospital, Locked Bag 300, Cnr Grattan Street and Flemington Road, Parkville VIC 3052 Australia. risha.bhatia@ thewomens.org.au Identification of Pneumothorax in Very Preterm Infants 119

6 THE JOURNAL OF PEDIATRICS Vol. 159, No. 1 References 1. Hill A, Perlman JM, Volpe JJ. Relationship of pneumothorax to occurrence of intraventricular hemorrhage in the premature newborn. Pediatrics 1982;69: Powers WF, Clemens JD. Prognostic implications of age at detection of air leak in very low birth weight infants requiring ventilatory support. J Pediatr 1993;123: Crowley P. Prophylactic corticosteroids for preterm birth. Cochrane Database Syst Rev 2000: CD Morley CJ. Systematic review of prophylactic vs rescue surfactant. Arch Dis Child Fetal Neonatal Ed 1997;77:F Horbar JD, Badger GJ, Carpenter JH, Fanaroff AA, Kilpatrick S, LaCorte M, et al. Trends in mortality and morbidity for very low birth weight infants, Pediatrics 2002;110: Morley CJ, Davis PG, Doyle LW, Brion LP, Hascoet JM, Carlin JB. Nasal CPAP or intubation at birth for very preterm infants. N Engl J Med 2008; 358: Finer NN, Carlo WA, Walsh MC, Rich W, Gantz MG, Laptook AR, et al. Early CPAP versus surfactant in extremely preterm infants. N Engl J Med 2010;362: Fanaroff AA, Stoll BJ, Wright LL, Carlo WA, Ehrenkranz RA, Stark AR, et al. Trends in neonatal morbidity and mortality for very low birthweight infants. Am J Obstet Gynecol 2007;196: Horbar JD, Badger GJ, Carpenter JH, Fanaroff AA, Kilpatrick S, LaCorte M, et al. Trends in mortality and morbidity for very low birth weight infants, Pediatrics 2002;110: Baumer JH. International randomised controlled trial of patient triggered ventilation in neonatal respiratory distress syndrome. Arch Dis Child Fetal Neonatal Ed 2000;82:F Ho JJ, Subramaniam P, Henderson-Smart DJ, Davis PG. Continuous distending pressure for respiratory distress syndrome in preterm infants. Cochrane Database Syst Rev 2002: CD Hall RT, Rhodes PG. Pneumothorax and pneumomediastinum in infants with idiopathic respiratory distress syndrome receiving continuous positive airway pressure. Pediatrics 1975;55: Berg TJ, Pagtakhan RD, Reed MH, Langston C, Chernick V. Bronchopulmonary dysplasia and lung rupture in hyaline membrane disease: influence of continuous distending pressure. Pediatrics 1975;55: Gittermann MK, Fusch C, Gittermann AR, Regazzoni BM, Moessinger AC. Early nasal continuous positive airway pressure treatment reduces the need for intubation in very low birth weight infants. Eur J Pediatr 1997;156: Milner AD. Surfactant and respiratory distress syndrome. Turk J Pediatr 1996;38: Yost CC, Soll RF. Early versus delayed selective surfactant treatment for neonatal respiratory distress syndrome. Cochrane Database Syst Rev 2000: CD Soll RF, Blanco F. Natural surfactant extract versus synthetic surfactant for neonatal respiratory distress syndrome. Cochrane Database Syst Rev 2001: CD Soll RF, Morley CJ. Prophylactic versus selective use of surfactant in preventing morbidity and mortality in preterm infants. Cochrane Database Syst Rev 2001: CD Sandri F, Plavka R, Ancora G, Simeoni U, Stranak Z, Martinelli S, et al. Prophylactic or early selective surfactant combined with ncpap in very preterm infants. Pediatrics 2010;125:e McIntosh N, Becher JC, Cunningham S, Stenson B, Laing IA, Lyon AJ, et al. Clinical diagnosis of pneumothorax is late: use of trend data and decision support might allow preclinical detection. Pediatr Res 2000; 48: Mansy HA, Royston TJ, Balk RA, Sandler RH. Pneumothorax detection using pulmonary acoustic transmission measurements. Med Biol Eng Comput 2002;40: Mansy HA, Royston TJ, Balk RA, Sandler RH. Pneumothorax detection using computerised analysis of breath sounds. Med Biol Eng Comput 2002;40: Victorino JA, Borges JB, Okamoto VN, Matos GF, Tucci MR, Caramez MP, et al. Imbalances in regional lung ventilation: a validation study on electrical impedance tomography. Am J Respir Crit Care Med 2004;169: Costa EL, Chaves CN, Gomes S, Beraldo MA, Volpe MS, Tucci MR, et al. Real-time detection of pneumothorax using electrical impedance tomography. Crit Care Med 2008;36: Hahn G, Just A, Dudykevych T, Frerichs I, Hinz J, Quintel M, et al. Imaging pathologic pulmonary air and fluid accumulation by functional and absolute EIT. Physiol Meas 2006;27:S Preis C, Luepschen H, Leonhardt S, Gommers D. Experimental case report: development of a pneumothorax monitored by electrical impedance tomography. Clin Physiol Funct Imaging 2009;29: Laptook AR, O Shea TM, Shankaran S, Bhaskar B. Adverse neurodevelopmental outcomes among extremely low birth weight infants with a normal head ultrasound: prevalence and antecedents. Pediatrics 2005;115: Goldberg RN. Sustained arterial blood pressure elevation associated with pneumothoraces: early detection via continuous monitoring. Pediatrics 1981;68: Bhatia et al

7 July 2011 ORIGINAL ARTICLES Table IV. Causes of death Cause of death Pneumothorax (n = 26) Control subjects (n = 8) Collapse, n (%) 2 (8%) 0 (0%) Intraventricular hemorrhage, n (%) 6 (23%) 1 (12.5%) Necrotizing enterocolitis, n (%) 5 (19%) 1 (12.5%) Respiratory distress syndrome, n (%) 9 (35%) 1 (12.5%) Sepsis, n (%) 3 (12%) 2 (25%) Withdrawal of treatment, n (%) 1 (4%) 2 (25%) Bronchopulmonary dysplasia, n (%) 0 (0%) 1 (12.5%) Table V. Pneumothorax details and morbidity by delivery room treatment for the pneumothorax group Pneumothorax group started on CPAP (n = 22) Pneumothorax group intubated and ventilated (n = 38) P value Gestational age* (wk) Birth weight* (g) Age at diagnosis (h) 31.3 ( ) 22 ( ).36 Age at intubation (h) 22 ( ) 0.05 ( ) <.001 Age at surfactant (h) 18.3 ( ) 0.6 ( ) <.001 Age at first CXR (h) 2.8 ( ) 1.9 ( ).03 Intubated on 1 st CXR 4 (18%) 38 (100%) <.001 Pneumothorax on 1 st CXR 0 (0%) 11 (29%).007 Intubated on diagnostic CXR 20 (91%) 37 (97%).56 Chest drain on diagnostic CXR 11 (50%) 17 (44.7%).79 Age at diagnostic CXR (h) 20/22 37/ ( ) 22.8 (2.8-45) Mode of ventilation at time pneumothorax diagnosed CPAP 8 (36.4%) 1 (2.6%) <.001 IPPV 12 (54.6%) 15 (39.5%) HFOV 2 (9.1%) 22 (57.9%) Survival, n (%) 17 (77%) 17 (45%).03 BPD, n (%) 3/17 (18%) 10/17 (59%).02 Any IVH, n (%) 7/22 (32%) 16/33 (48%).33 IVH grade 3 or 4, n (%) 3/22 (14%) 9/33 (27%).47 IPPV, intermittent positive pressure ventilation; HFOV, high-frequency oscillatory ventilation. *Mean (SD). Median (interquartile range). Figure 2. ROC curve of FiO 2 requirements at 12 hours in all infants in the pneumothorax group. The FiO 2 at points labeled a, b, c, and d are shown. Identification of Pneumothorax in Very Preterm Infants 120.e1