Pulmonary hemorrhage in neonatal respiratory distress syndrome: Radiographic evolution,course,complications and long-term clinical outcomes

Abstract

BACKGROUND:

Pulmonary hemorrhage (PH) is occasionally seen in premature infants after surfactant treatment for respiratory distress syndrome (RDS). These infants receive frequent chest radiographs (CXR) during and after hospitalization enabling long-term radiographic-clinical correlation.

OBJECTIVE:

To chart the natural evolution of CXR findings of PH in RDS and correlate radiographic patterns to supplemental oxygen requirement.

MATERIALS AND METHODS:

Retrospective review of clinical notes for gestational age (GA), birth weight (BW), intraventricular hemorrhage (IVH) and oxygen requirement were performed. CXRs were reviewed at 4 time-points; during PH, 28 days postnatal age, 36 weeks and at farthest available clinical follow-up.

RESULTS:

18 infants born (2003–2016), GA (24–30 weeks); BW (482–1590 grams) were included. Mean onset of PH was 1.94 (0–5) days. 9/18 (50%) had IVH. 3 died during PH; all had IVH. During PH, CXR showed whiteout 9/18 (50%); patchy opacities 5/18 (27%); diffuse haziness 1/18 (6%) and no change 3/18 (17%). At 28 days postnatal age, CXR showed fine-interstitial (FI) markings 14/15 (93%) and whiteout 1/15 (7%). At 36 weeks,12/14 (85%) had FI and 2/14 (15%) developed cystic-interstitial changes. At farthest follow-up, FI 3/13 (23%); coarse-interstitial 4/13 (30%); peri-bronchial cuffing 5/13 (38%); normal 1/13 (9%) and the majority had hyperinflation 9/13 (69%). At discharge, 9/14 (64%) required home-oxygen and 5/14 (36%) were on room-air. At farthest follow-up, 6/14 (42%) required home-oxygen and 8/14 (58%) were on room-air.

CONCLUSION:

Premature infants that survive PH may later develop chronic lung disease of prematurity with an evolving interstitial pattern on CXR that clears overtime as they outgrow the need for supplemental oxygen.

Keywords

Pulmonary hemorrhage respiratory distress syndrome chronic lung disease of prematurity

1 Introduction

Neonatal respiratory distress syndrome (RDS) is a leading cause of morbidity and mortality in preterm infants. RDS is caused primarily by deficiency of pulmonary surfactant in the immature lung with a rising risk with increasing prematurity [1 –3]. Clinical trials in preterm infants with RDS have shown that endotracheal administration of exogenous surfactant is effective in reducing mortality by improving oxygenation and ventilation, and thus reducing rates of developing bronchopulmonary dysplasia (BPD), also known as chronic lung disease of prematurity (CLD) [4 –9]. Thus, the American Academy of Pediatrics (AAP) recommendation has historically been that all patients with RDS should receive surfactant at or soon after birth [10].

Studies have shown that pulmonary hemorrhage (PH), is a major complication in premature babies with RDS, which typically presents after surfactant administration [11]. Although the precise mechanism of PH remains unclear, it has been suggested that left-to-right shunting through the patent ductus arteriosus (PDA) may cause hemorrhagic pulmonary edema [12 –15]. Furthermore, exogenous surfactant administration requires endotracheal intubation risking barotrauma. For these reasons, and the 2014 AAP’s recommendation, current at the time of writing, that “continuous positive airway pressure therapy (CPAP) started at or soon after birth with subsequent selective surfactant administration may be considered as an alternative to routine intubation with prophylactic or early surfactant administration in preterm infants” has more recently led to the clinical hesitation to routinely administer surfactant [16].

On radiography, acute PH in RDS manifests as new alveolar opacities, however, to our knowledge, the evolution of radiographic patterns and long-term pulmonary sequelae have not been studied from an imaging standpoint [17 –19]. Additionally, patients with PH typically require prolonged and aggressive respiratory support via mechanical ventilation and is considered as a major insult to the immature lungs. This leaves a gap in our knowledge and in the imaging literature, about the long-term course of survivors of RDS complicated with PH, and how these patients present, both clinically and radiographically, weeks, months to years after the initial event. The purpose of our study was to characterize and chart the chest radiography (CXR) patterns in a cohort of premature infants with RDS who developed PH after receiving surfactant, and to longitudinally follow the evolution of radiographic findings and clinically correlating with need for supplemental oxygen requirement over months to years.

2 Materials and methods

This is a retrospective, institutional review board-approved and HIPAA (Health Information Portability and Accountability Act) compliant study. Premature infants with RDS who developed PH after surfactant administration, born at an urban tertiary care academic medical center, with advanced Level IV Neonatal intensive care unit (NICU) capability (highest level of neonatal care according to AAP) were included in the study [20]. Premature infants with RDS who suffered traumatic intubation or did not develop PH, infants with neonatal pneumonia, meconium aspiration syndrome, underlying congenital heart disease, or a known bleeding disorder were excluded. Electronic health records from 2003 to 2016 with keywords including RDS and PH was searched using free text radiology report and image search application known as RENDER, a web-based query interface with password-protected access available on the hospital intranet within the institution’s firewall.

Patient records were assessed for gestational age (GA) in weeks, birth weight (BW) in grams (g), and year of birth. Prematurity and BW based categorizes were created using the world health organization (WHO) classification into Extremely Preterm (EPT; < 28 weeks), Very Preterm (VPT; 28–32 weeks) or Preterm (PT 32–37 weeks), Extremely Low Birth Weight (ELBW; BW < 1000 g), Very Low Birth weight (VLBW; BW 1000–1500 g), or Low Birth Weight (LBW; BW 1500–2500 g). Clinical notes were reviewed for antenatal steroid therapy (AST), duration of PH, PDA, IVH and total days of intubation.

To clinically assess for CLD, the medical records were reviewed for need for supplemental oxygen at 28 days postnatal age, 36 weeks, at discharge, and at farthest available clinical follow-up. The time-point of 28 days postnatal age and 36 weeks were chosen in accordance with the most current definition of BPD/CLD, developed in 2001 at the NICHD/NHLBI/ORD BPD workshop [21]. This definition requires oxygen dependency for ≥28 days and at 36 weeks postnatal age or at discharge from the hospital, whichever comes first for infants born under 32 weeks gestation [21, 22]. This is also in accordance with the convention at the institution where the study was conducted. CXR for study participants were reviewed by three fellowship trained full-time pediatric radiologists with 12 years, 13 years and 25 years of clinical experience. Agreement was reached by consensus.

All CXRs were reviewed from birth till the last available clinical follow-up. CXR findings at 4 time-points; during the initial PH event; at 28 days postnatal age and 36 weeks when infants were clinically assessed for CLD; and at farthest available long-term clinical follow-up were recorded for imaging findings. The 1st time-point sought to categorize CXRs for interval change by examining existence and laterality of new opacities suggestive of acute PH and degree of parenchymal opacification (diffuse haziness, patchy alveolar to complete whiteout of both lung fields). Follow-up CXRs at 28 days postnatal age (2nd time-point) and 36 weeks (3rd time-point), were chosen to comply with the NIH consensus definition of BPD/CLD as above [21]. Finally, CXRs at farthest available clinical follow-up (4th time-point) were reviewed. Radiographic findings in the 2nd, 3rd and 4th time-points sought to categorize CXRs for imaging findings of CLD. Imaging descriptors included both of “new” BPD and “old” BPD imaging patterns of CLD described in literature [8 , 21–23]. These included interstitial markings (fine, coarse, peribronchial cuffing and cystic changes), parenchymal bands and architectural distortion. Additionally, note was made of any overlying opacities ranging from diffuse haziness, patchy alveolar to complete whiteout. The degree of lung inflation was categorized into normal, hyperinflation and hypoinflation. Signs of complications of barotrauma including pneumothorax and pulmonary interstitial emphysema (PIE) during their hospital course was recorded; and when available, findings on computed tomography (CT) of the chest was documented.

3 Results

Clinical characteristics of our patient population are presented in Table 1. 18 premature infants (10 male; 8 female) born from 2003–2016 met the inclusion criteria for the study. These infants were born at a mean GA of 26 2/7 weeks (24–30) and mean BW of 824 g (482–1590). Birth weight based categories of all 18 infants were the following: ELBW (15), VLBW (2) and LBW (1). Classification based on gestational age was: EPT (16) and VPT (2). 14/18 infants had received AST. The mean onset of PH was 1.94 days (0–5) with an average total duration of 2 days (1–10). 9/18 (50%) infants developed IVH in conjunction with PH. Echocardiography was performed to evaluate for PDA after the onset of PH, either on the same or the next day. PDA was discovered in 16 patients, 1 had no PDA, and 1 died without undergoing echocardiography. A total of 5/18 (28%) died in our cohort at farthest follow-up; of which 3/18 (16%) died soon after birth, during the acute PH event. These three cases also had IVH and were born on or before 2010 and were recorded as direct PH mortality. 14 of the survivors were discharged home, while 1 was transferred to another facility at 4 weeks of age. Post discharge and during clinical follow-up of the 14 survivors of PH, there were two fatalities at 5 months and 1 year of age respectively. The cause of these late deaths was unrelated to pulmonary complications with end stage respiratory failure and were recorded as indirect PH mortality.

Table 1
Clinical Data

n Yr born GA Wks Preterm Category Sex Wgt Gms BW Category AST IVH Grade* Barotrauma PDA PH Start Day PH End Day Intubation Days O₂ at 36 Wks GA O₂ at DC O₂ at Furthest FU Outcome

1 2003 25 4/7 EPT M 870 ELBW Y Grade 1 BL PIE Y 2 4 68 Y Y Y Alive at furthest Follow-up

2 2004 26 EPT M 970 ELBW N Grade 1 N N 2 4 5 NA NA NA Alive at furthest Follow-up

3 2005 27 3/7 EPT F 580 ELBW Y Grade 3 N Y 0 1 75 Y N N Indirect Mortality

4 2005 27 EPT M 1080 VLBW N Grade 1 N UN 2 2 2 NA NA NA Direct Mortality

5 2006 25 3/7 EPT F 560 ELBW Y N N Y 4 5 82 Y Y Y Alive at furthest Follow-up

6 2008 27 1/7 EPT F 980 ELBW Y Grade 3 BL PIE / L PTX Y 2 3 92 Y Y Y Alive at furthest Follow-up

7 2009 26 6/7 EPT F 800 ELBW Y Grade 2 N Y 2 2 2 NA NA NA Direct Mortality

8 2009 24 2/7 EPT F 580 ELBW N Grade 4 N Y 1 3 47 Y Y N Alive at furthest Follow-up

9 2010 26 EPT M 910 ELBW N N N Y 0 1 67 Y N N Alive at furthest Follow-up

10 2010 28 VPT M 1050 VLBW Y Grade 4 N Y 2 2 2 NA NA NA Direct Mortality

11 2010 25 5/7 EPT M 880 ELBW Y Grade 4 L PIE / L PTX Y 2 3 55 Y N N Alive at furthest Follow-up

12 2011 26 6/7 EPT F 770 ELBW Y N N Y 2 3 30 Y Y Y Alive at furthest Follow-up

13 2012 24 4/7 EPT M 700 ELBW Y N N Y 2 3 28 Y N N Alive at furthest Follow-up

14 2013 30 2/7 VPT M 1590 LBW Y N N Y 1 2 5 Y Y N Alive at furthest Follow-up

15 2013 24 4/7 EPT F 482 ELBW Y N N Y 2 3 30 Y Y Y Alive at furthest Follow-up

16 2014 25 3/7 EPT F 600 ELBW Y N L PIE Y 2 10 46 Y N N Alive at furthest Follow-up

17 2015 26 6/7 EPT M 730 ELBW Y N N Y 2 12 46 Y Y N Alive at furthest Follow-up

18 2016 24 3/7 EPT M 710 ELBW Y N N Y 5 8 66 Y Y Y Indirect Mortality

n	Yr born	GA Wks	Preterm Category	Sex	Wgt Gms	BW Category	AST	IVH Grade*	Barotrauma	PDA	PH Start Day	PH End Day	Intubation Days	O₂ at 36 Wks GA	O₂ at DC	O₂ at Furthest FU	Outcome
1	2003	25 4/7	EPT	M	870	ELBW	Y	Grade 1	BL PIE	Y	2	4	68	Y	Y	Y	Alive at furthest Follow-up
2	2004	26	EPT	M	970	ELBW	N	Grade 1	N	N	2	4	5	NA	NA	NA	Alive at furthest Follow-up
3	2005	27 3/7	EPT	F	580	ELBW	Y	Grade 3	N	Y	0	1	75	Y	N	N	Indirect Mortality
4	2005	27	EPT	M	1080	VLBW	N	Grade 1	N	UN	2	2	2	NA	NA	NA	Direct Mortality
5	2006	25 3/7	EPT	F	560	ELBW	Y	N	N	Y	4	5	82	Y	Y	Y	Alive at furthest Follow-up
6	2008	27 1/7	EPT	F	980	ELBW	Y	Grade 3	BL PIE / L PTX	Y	2	3	92	Y	Y	Y	Alive at furthest Follow-up
7	2009	26 6/7	EPT	F	800	ELBW	Y	Grade 2	N	Y	2	2	2	NA	NA	NA	Direct Mortality
8	2009	24 2/7	EPT	F	580	ELBW	N	Grade 4	N	Y	1	3	47	Y	Y	N	Alive at furthest Follow-up
9	2010	26	EPT	M	910	ELBW	N	N	N	Y	0	1	67	Y	N	N	Alive at furthest Follow-up
10	2010	28	VPT	M	1050	VLBW	Y	Grade 4	N	Y	2	2	2	NA	NA	NA	Direct Mortality
11	2010	25 5/7	EPT	M	880	ELBW	Y	Grade 4	L PIE / L PTX	Y	2	3	55	Y	N	N	Alive at furthest Follow-up
12	2011	26 6/7	EPT	F	770	ELBW	Y	N	N	Y	2	3	30	Y	Y	Y	Alive at furthest Follow-up
13	2012	24 4/7	EPT	M	700	ELBW	Y	N	N	Y	2	3	28	Y	N	N	Alive at furthest Follow-up
14	2013	30 2/7	VPT	M	1590	LBW	Y	N	N	Y	1	2	5	Y	Y	N	Alive at furthest Follow-up
15	2013	24 4/7	EPT	F	482	ELBW	Y	N	N	Y	2	3	30	Y	Y	Y	Alive at furthest Follow-up
16	2014	25 3/7	EPT	F	600	ELBW	Y	N	L PIE	Y	2	10	46	Y	N	N	Alive at furthest Follow-up
17	2015	26 6/7	EPT	M	730	ELBW	Y	N	N	Y	2	12	46	Y	Y	N	Alive at furthest Follow-up
18	2016	24 3/7	EPT	M	710	ELBW	Y	N	N	Y	5	8	66	Y	Y	Y	Indirect Mortality

Yr: year, GA: gestational age, Wks: weeks, Wgt: weight, GM: grams; M: male, F: female, Y: yes, N: no, UN: unknown, IVH: intraventricular hemorrhage, L: left, BL: bilateral, PIE: pulmonary interstitial emphysema, PTX: pneumothorax, DC: discharge, NA: not applicable, FU: follow-up, O₂: oxygen, PDA: patent ductus arteriosus, PH: pulmonary hemorrhage, EPT: extremely preterm, VPT: very preterm, ELBW: extremely low birth weight, VLBW: very low birth weight, LBW: low birth weight. *Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500gm. J Pediatr 1978; 92:529-34.

Radiographic findings for all 18 patients are presented in Table 2. All patients CXRs prior to development of PH demonstrated typical findings of RDS with diffuse granular opacities with improved aeration and clearing of both lung fields after surfactant administration. CXR at 1st time-point, during acute PH showed diffuse whiteout in 9/18 infants (50%), bilateral patchy alveolar opacities in 5/18 (27%), diffuse haziness in 1/18 (6%) and no interval change in 3/18 (17%). Interestingly, all 3 direct PH fatalities died on second day of life and showed complete whiteout of both lung fields consistent with overwhelming PH (Fig. 1).

Table 2

Chest radiograph findings

n	Year Born	At time of PH	28 Day PNA	GA	Follow-up CXR	Follow-up	Outcome
1	2003	Whiteout	FI, Hazy	FI, Hazy	FI, HI	8 Mo	Alive at furthest Follow-up
2	2004	Whiteout	FI, Hazy	Transferred	Loss to FU	1 Mo	Alive at furthest Follow-up
3	2005	No change	FI, Hazy	FI, Hazy	CI, HI	1Yr 2 Mo	Indirect Mortality
4	2005	Whiteout	NA	NA	NA	Day 2	Direct Mortality
5	2006	Whiteout	FI, Hazy	CC	PC, HI	6 Yr	Alive at furthest Follow-up
6	2008	Whiteout	WO	CC	PC, HI, BLO Perihilar	4 Yr	Alive at furthest Follow-up
7	2009	Whiteout	NA	NA	NA	Day 2	Direct Mortality
8	2009	Hazy	FI, Hazy	FI	PC, HI	5 Yr	Alive at furthest Follow-up
9	2010	Whiteout	FI, Hazy	FI, Hazy	FI, Hazy	3 Mo	Alive at furthest Follow-up
10	2010	Whiteout	NA	NA	NA	Day 2	Direct Mortality
11	2010	No change	FI	FI, Hazy	Clear	2 Yr	Alive at furthest Follow-up
12	2011	Patchy BL alveolar	FI	FI	PC, HI, BLO R Lung	4 Yr	Alive at furthest Follow-up
13	2012	Patchy BL alveolar	FI, Hazy	FI	FI, HI	7 Mo	Alive at furthest Follow-up
14	2013	No change	FI, Hazy	FI	CI, BLO L Lung	10 Mo	Alive at furthest Follow-up
15	2013	Patchy BL alveolar	FI, Hazy	FI, Hazy	CI, Hazy	3 Mo	Alive at furthest Follow-up
16	2014	Whiteout	FI	FI, Hazy	PC, HI, BLO R Lung	2Yr 4 Mo	Alive at furthest Follow-up
17	2015	Patchy BL alveolar	FI, Hazy	FI, Hazy	CXR not done at clinical follow-up	1Yr 6 Mo	Alive at furthest Follow-up
18	2016	Patchy BL alveolar	FI, Hazy	FI, Hazy	CI, HI	3 Mo	Indirect Mortality

PH pulmonary hemorrhage, PNA postnatal age, GA gestational age, CXR chest X-ray, NA not applicable, WO whiteout, NC no change, FI fine interstitial markings, CI coarse interstitial markings, PC peribronchial cuffing, HI hyperinflation, BLO band like opacity, BL bilateral, L Left, R Right, Yr year, Mo month, CC cystic interstitial changes, FU follow-up.

Fig.1

28 weeks old male who died on day 2 with severe pulmonary hemorrhage. Pre-surfactant CXR on 1d (A) shows intubated patient with diffuse granular opacities and air bronchograms consistent with RDS. Post-surfactant CXR on 1d after 4 hours (B) shows marked improved aeration of both lung fields. CXR on day 2 (C) shows diffuse whiteout consistent with overwhelming PH. Head ultrasound, coronal image on day 2 due to sudden drop of hematocrit (D) shows right sided germinal matrix hemorrhage with intraparenchymal extension (white arrow) consistent with Grade IV germinal matrix hemorrhage.

After the initial PH event, surviving infants 15/18 (83%) received frequent CXRs over the course of their hospitalization. The CXR findings of acute PH were noted to be transient; as survivors showed clearance of the opacities within the next few days correlating with clinical resolution of PH. 4/15 survivors (26%) developed signs of barotrauma, including pneumothorax and pulmonary interstitial emphysema.

At 28 days postnatal age, 14/15(93%) surviving infants had CXRs revealing a background of fine interstitial markings with overlying diffuse haziness in 11; and the remaining 1 infant showed diffuse whiteout of both lung fields. At 36 weeks, at time of institutional clinical assessment for CLD, 1 infant had been transferred to another hospital and was lost to follow-up. Of the 14 infants who continued to be hospitalized, 12/14 (85%) CXRs revealed a background of fine interstitial markings, with 8 also showing overlying diffuse haziness. This finding suggests that typically by 28 days, most patients develop a fine interstitial pattern throughout both lung fields, similar to early changes of “new” BPD/CLD and the diffuse haziness possibly represents excess lung fluid. The remaining 2/14 (15%) patients showed diffuse cystic interstitial changes at 36 weeks. One of these two infants had a clinical course complicated by barotrauma, with bilateral PIE and a large left sided tension pneumothorax requiring chest tube decompression. Furthermore, both these patients with cystic interstitial changes had later required tracheostomies for prolonged dependency on mechanical ventilation. In addition, these patients were the only two in our sample to have chest CT imaging available, performed at 6 and 7 months of age respectively. Chest CT in both cases revealed cystic interstitial changes with architectural distortion and parenchymal scarring, similar to the “old” BPD pattern (Fig. 2).

Fig.2

27 weeks old female with evolving patterns of CLD with up to 4-year clinical follow-up. Serial CXRs on day 1 shows typical granular pattern of RDS, pre-surfactant (A) and immediate post-surfactant (B) with improved aeration of both lung fields. On day 2, the patient developed PH with complete whiteout (C). On day 4 (D), changes of barotrauma including left sided pneumothorax (black arrow) and PIE (white arrow) was observed. On 28 days (E), diffuse whiteout over a background of coarse interstitial markings and post-surgical changes from PDA ligation was seen. At 36 weeks (F) bilateral cystic interstitial changes were present. Patient needed a tracheostomy for prolonged dependency on ventilator and CXR (G) and non-contrast enhanced chest CT with axial (H) and coronal (I) images done at 6 months show cystic changes (curved arrow) with scarring, architectural distortion and air-trapping throughout both lung fields. CXR at 4year follow-up with frontal (J) and lateral (K) projections, demonstrate hyperinflation with mild perihilar band like opacities, representing residual scarring. The cystic changes have resolved, and the patient has been decannulated as she no longer requires a tracheostomy.

Regarding the long-term clinical radiographic follow-up, 13 of 14 patients surviving past the birth admission had a post discharge CXR available. The one patient that did not have a corresponding CXR at farthest clinical follow-up of 18 months was on room air and doing well from a pulmonary standpoint. Farthest available follow-up imaging in these 13 patients was obtained at ages ranging from 3 months to 6 years. Hyperinflation was the most common imaging finding and was present in 9/13 (69%). This manifested as flattened domes of the diaphragm and prominent retrosternal airspace. Regarding interstitial changes, CXRs showed fine interstitial markings in 3/13 (23%); coarse interstitial markings in 4/13 (30%); peribronchial cuffing in 5/13 (38%) and clear lungs in 1/13 (9%). Parenchymal band-like opacity was seen in 4/13 (30%) and it is unclear whether these band-like opacities represented focal areas of parenchymal scarring or areas of subsegmental atelectasis, as patients with CLD are also known to have a higher risk of airway hypersensitivity and mucous plugging. The 2 patients who later died of indirect causes had CXRs obtained preceding their deaths; at 3 months and 14 months of age, both showing coarse interstitial markings and hyperinflation. Finally, 1 patient had clear lungs on CXR at 2 years of age (Fig. 3).

Fig.3

Serial CXR’s of a 25 weeks old female with no oxygen requirement at discharge and at 2-year clinical follow-up. CXRs on 1d (A) show a mild granular pattern suggestive of RDS. On day 2 (B) patient developed PH with complete whiteout. Day 3 (C) changes of bilateral pulmonary interstitial emphysema (white arrow) suggestive of barotrauma was observed. At 28 days (D) and 36 weeks (E), fine interstitial markings with diffuse haziness was seen respectively. CXR at 2-year follow-up frontal (F) and lateral (G), show near normal findings with mild hyperinflation and peribronchial cuffing.

At farthest follow-up (up to 6 years), no patient later developed cystic interstitial changes and architectural distortion as described with “old” BPD. In particular, the 2 patients who had developed cystic interstitial changes on CXR at 36 weeks, and also on CT at 6 & 7 months, had follow-up CXRs at 4 years and 6 years of age respectively, both showing resolution of cystic changes with residual hyperinflation and peribronchial cuffing.

All infants in the cohort were clinically assessed for CLD at 36 weeks per the formal BPD/CLD definition and per institutional protocol. All of the 14/14 (100%) infants surviving to 36 weeks, required supplemental oxygen. At discharge, 9/14 (64%) infants required supplemental home oxygen, while 5/14 (36%) were on room air. Regarding oxygen requirement at farthest available clinical follow-up (including the two patients that later died of indirect causes); 6/14 (42%) required home oxygen and 8/14 (58%) were on room air.

4 Discussion

PH is a rare but significant event encountered early during RDS management. It typically presents soon after birth following endotracheal administration of surfactant; and premature infants born at < 28 weeks gestational age and with extremely low birth weight < 1000 g are at highest risk. Similar profile was also seen in our study as we found only a total of 18 documented cases since 2003 in our database, and the majority were ELBW and EPT. This relatively low occurrence of PH at our institution may stem in part from change in management trend of premature infants with RDS, including the AAP’s updated recommendation for selective surfactant therapy rather than routine intubation for prophylactic surfactant administration. Nevertheless, it is a significant complication, given the heightened concern for ongoing respiratory failure requiring prolong mechanical ventilation support and thus risk of developing CLD in the survivors.

Secondly, all 3 direct PH fatalities in our study that died soon after birth, also had IVH. This observation was similar to as Li et al reported [25]. Furthermore, all three were born between 2003–2010, and the lack of direct PH related fatalities since 2011, may suggest improved management of several co-morbidities including IVH [4, 5]. Furthermore, we also observed a high association of PH with PDA (16 of 18 infants), and possibly early recognition of hemodynamically significant PDA may be an important step in managing PH [13 –15]. Worthy of mentioning is that all 3 direct PH mortalities could not undergo PDA ligation as they were hemodynamically unstable for surgical intervention. 4 of our cases had spontaneous closure of the PDA in the first week of life and the remainder underwent surgical ligation at 4 to 22 days of age.

Thirdly, we found an association between PH and the clinical development of CLD. All infants met the diagnostic criteria for CLD by oxygen requirement at 36 weeks and the majority showed characteristic radiographic findings of fine interstitial markings and diffuse haziness as seen in “new” BPD. Only 2 patients developed early cystic changes with architectural distortion similar to “old” BPD. Interestingly, we found that these radiographic abnormalities gradually cleared and resolved over time up to farthest available clinical follow-up of 6 years. These 2 patients also had the lengthiest total days of intubation and required tracheostomies for prolonged ventilatory support, and currently continue to require supplemental oxygen at farthest available clinical follow-up.

In terms of clinical course, it is known that most infants with CLD see a gradual improvement of their pulmonary function over the first two to four months of life, and are weaned from intubation to CPAP, to high flow nasal cannula, until they can finally maintain oxygenation on room air [22]. In our study, 5 patients were on room air by the time of discharge, and 8 at the farthest clinical follow-up. 5 patients are still requiring home oxygen, and amongst these are the two patients who developed cystic changes and needed tracheostomies to support prolonged dependency on mechanical ventilation. Hence, we found an association between cystic changes with prolonged clinical oxygen requirement. We believe that these cystic changes may reflect more severe lung damage from a combination of mechanical ventilation-mediated lung injury and oxygen toxicity due to high inspired oxygen concentrations as seen in old BPD [23, 24].

Meanwhile, only 1 patient had a clear CXR at 2 years follow-up. This patient was discharged on room air and continued on room air at farthest follow-up. 7 of our patients, i.e. about half of our cohort, had farthest clinical follow-up prior to two years of age. Thus, we speculate, that some of these infants who are doing better from a pulmonary standpoint and not requiring supplemental oxygen may eventually show clear lungs.

Chronologically, the evolution of radiographic findings observed in our cohort was the following: the majority of patients had fine interstitial markings suggestive of CLD at 28 days postnatal age and 36 weeks respectively, and by 2 years developed fine or coarse interstitial markings throughout both lungs fields. As the years advanced, we found gradual clearing of the peripheral lung fields with residual perihilar peribronchial cuffing. Isolated parenchymal band-like opacities were seen in a few cases, which may represent focal areas of parenchymal scarring or subsegmental atelectasis from mucous plugging, as radiographs were obtained during periods of acute exacerbation of respiratory symptoms, mostly due to viral bronchiolitis. Finally, radiographic signs of hyperinflation were by far the most common imaging finding seen in the majority of cases at farthest available clinical follow-up. The finding of air trapping, is in keeping with the known alterations of airway physiology described in CLD, which includes a combination of obstructive airway disease and airway hypersensitivity, both of which are major determinants of long-term pulmonary outcomes [21 –23].

We acknowledge that this study has limitations including small sample size and retrospective design. Like all retrospective analyses, it is limited by availability of follow-up data for patients; in particular, patients who appeared at their last clinical encounter to be improving but had no subsequent CXRs, were unable to be assessed by imaging. A reasonable inference is that patients who were clinically improved did not require follow-up imaging. Meanwhile, the follow-up CXRs available in the remainder of our patients were obtained during episodes of acute respiratory illness, complicating the imaging findings of CLD with overlying pneumonia, as only the sicker infants in the sample continued to undergo imaging over time. Furthermore, in the current era of advancements in sophisticated imaging technologies including quantitative, physiologic and functional imaging of the lungs, imaging analyses in our study only included plain radiography. Nonetheless, plain radiography is by far the most common imaging modality utilized in children with RDS and CLD for purposes of routine management and clinical follow-up, and is generally a reliable indicator of disease extent. Modalities like CT and magnetic resonance imaging require ionizing radiation and sedation respectively, particularly in younger children, and may not be appropriate for routine follow-up.

Despite these limitations, we believe that the study also has several strengths. To our knowledge, it is the first to explore the long-term radiographic evolution of RDS complicated by PH in recent times, and correlate imaging findings with clinical oxygen requirement. Despite the small sample size, we were able to obtain clinical data over a substantial period of time and observed a noteworthy difference in survivors of PH during that time frame. We hope that our experience and this data can provide information to clinicians as well as patients families regarding the range of outcomes possible in premature infants with RDS complicated by PH.

5 Conclusions

This single center small study found that amongst all different categories of prematurity, the extremely premature infant i.e. (GA < 28 weeks; BW < 1000 g) may be at increased risk of PH, which may occur soon after surfactant administration for treatment of RDS. We found more survivors of PH in the last decade, which may reflect changing management trends of RDS particularly in an advanced NICU setting.

Radiographically, RDS with PH presents as new alveolar opacities, associated with acute clinical deterioration of respiratory status, and the survivors may later show typical clinical and radiographic signs of CLD. On CXR, the most common findings of early CLD included mild interstitial involvement that gradually resolves overtime. Cystic interstitial changes with architectural distortion suggested more severe lung injury, correlating with prolonged dependency on mechanical ventilation and subsequent home oxygen requirement. Fortunately, we found that this imaging finding may be temporary in children, unlike adults with extensive architectural distortion and cystic honeycombing changes reflecting end stage lung disease. We believe that these evolving imaging findings highlight the unique ability of lungs in infants and young children to remodel and heal with time.

Although future outcomes cannot be derived or generalized from our small sized study, we hope that it does however bring to light our experience particularly from an imaging standpoint due to relative paucity of data on this topic in the imaging literature. While PH undeniably is a major complication in clinical management of RDS, our experience suggests that cautious optimism is reasonable in absence of other comorbidities, as some infants may survive and recover without developing lifelong CLD and supplemental oxygen requirement. Furthermore, with recent changes in clinical practice of RDS using CPAP and selective surfactant administration particularly in a level IV NICU setting, may reflect our decreasing incidence of PH and possibly improved long-term pulmonary outcomes for the survivors.

Disclosure statements

Financial disclosure statements

All the authors and the affiliated institution have no conflicts of interest.

Human research statement

This retrospective study was conducted with institutional review board approval in accordance with the ethical standards of all applicable national and institutional committees and the World Medical Association’s Helsinki Declaration.

Footnotes

Acknowledgments

We would like to thank Noemi Chavez and Eleni Balasalle for their expertise with the preparation of the manuscript.

References

Rodriguez

, Martin

, Fanaroff

. Respiratory distress syndrome and its management. In: Fanaroff

, Martin

, editors. Fanaroff and Martin’s Neonatal Perinatal Medicine: Diseases of the Fetus and Infant. 7th ed. St. Louis, MO: Mosby; 2002.

Goldenberg

, Culhane

, Iams

, Romero

. Epidemiology and causes of preterm birth, Lancet. 2008;371(9606):75–84.

Robertson

, Sniderman

, Laros

Jr , Cowan

, Heilbron

, Goldenberg

, et al. Neonatal morbidity according to gestational age and birth weight from five tertiary care centers in the United States, 1983 through 1986. Am J Obstet Gynecol. 1992;166(6 Pt 1):1629–41; discussion 41-5.

Martin

. Pathophysiology, clinical manifestations, and diagnosis of respiratory distress syndrome in the newborn January 31, 2018 [cited 2016 August 23, 2016]. Available from: https://www.uptodate.com/contents/pathophysiology-clinical-manifestations-and-diagnosis-of-respiratory-distress-syndrome-in-the-newborn.

Pramanik

. Respiratory Distress Syndrome January 31, 2018 [cited 2015 January 16, 2015]. Available from: https://emedicine.medscape.com/article/976034-overview#a4.

Lopez

, Gascoin

, Flamant

, Merhi

, Tourneux

, Baud

, et al. Exogenous surfactant therapy in 2013: What is next? Who, when and how should we treat newborn infants in the future? BMC Pediatr. 2013;13:165.

Northway

Jr , Rosan

, Porter

. Pulmonary disease following respirator therapy of hyaline-membrane disease. Bronchopulmonary dysplasia, N Engl J Med. 1967;276(7):357–68.

Bancalari

, Claure

. Definitions and diagnostic criteria for bronchopulmonary dysplasia, Semin Perinatol. 2006;30(4):164–70.

Edwards

, Hilton

, Merritt

, Hallman

, Mannino

, Boynton

. Respiratory distress syndrome treated with human surfactant: Radiographic findings, Radiology. 1985;157(2):329–34.

10.

Engle

. American Academy of Pediatrics Committee on Fetus and Newborn. Surfactant replacement therapy for respiratory distress in the preterm and term neonate, Pediatrics. 2008;121(2):419–32.

11.

Towers

, Wung

. Rationing ventilatory support for the newborn with respiratory distress, Pediatr Radiol. 1995;25(8):638–40.

12.

Pandit

, Dunn

, Colucci

. Surfactant therapy in neonates with respiratory deterioration due to pulmonary hemorrhage, Pediatrics. 1995;95(1):32–6.

13.

Alfaleh

, Smyth

, Roberts

, Solimano

, Asztalos

, Schmidt

, et al. Prevention and 18-month outcomes of serious pulmonary hemorrhage in extremely low birth weight infants: Results from the trial of indomethacin prophylaxis in preterms, Pediatrics. 2008;121(2):e233–8.

14.

Garland

, Buck

, Weinberg

. Pulmonary hemorrhage risk in infants with a clinically diagnosed patent ductus arteriosus: A retrospective cohort study, Pediatrics. 1994;94(5):719–23.

15.

Lewis

, McKeever

, Rutty

. Patent ductus arteriosus as a natural cause of pulmonary hemorrhage in infants: A medicolegal dilemma, Am J Forensic Med Pathol. 2004;25(3):200–4.

16.

Respiratory support in preterm infants at birth. Committee on Fetus and Newborn, American Academy of P. Respiratory support in preterm infants at birth, Pediatrics. 2014;133(1):171–4.

17.

Odita

. The significance of recurrent lung opacities in neonates on surfactant treatment for respiratory distress syndrome, Pediatr Radiol. 2001;31(2):87–91.

18.

Bick

, Muller-Leisse

, Troger

, Jorch

, Roos

, Meyer zu Wendischhoff

, et al. Therapeutic use of surfactant in neonatal respiratory distress syndrome. Correlation between pulmonary X-ray changes and clinical data, Pediatr Radiol. 1992;22(3):169–73.

19.

Dinger

, Schwarze

, Rupprecht

. Radiological changes after therapeutic use of surfactant in infants with respiratory distress syndrome, Pediatr Radiol. 1997;27(1):26–31.

20.

Levels of neonatal care. American Academy of Pediatrics Committee on Fetus And Newborn, Pediatrics. 2012;130(3):587–97.

21.

Jobe

, Bancalari

. Bronchopulmonary dysplasia, Am J Respir Crit Care Med. 2001;163(7):1723–9.

22.

Eichenwald

. Pathogenesis and clinical features of bronchopulmonary dysplasiaJanuary 31, 2018 [cited 2017 September 26, 2017]. Available from: https://www.uptodate.com/contents/pathogenesis-and-clinical-features-of-bronchopulmonary-dysplasia/print.

23.

Coalson

. Pathology of new bronchopulmonary dysplasia, Semin Neonatol. 2003;8(1):73–81.

24.

Naik

, Kallapur

, Bachurski

, Jobe

, Michna

, Kramer

, et al. Effects of ventilation with different positive end-expiratory pressures on cytokine expression in the preterm lamb lung, Am J Respir Crit Care Med. 2001;164(3):494–8.

25.

, Yu

, Wang

, Zhang

, Shen

, Yuan

, et al. A prediction score model for risk factors of mortality in neonate with pulmonary hemorrhage: The experience of single neonatal intensive care unit in Southwest China, Pediatr Pulmonol. 2008;43(10):997–1003.