Secondary Logo

Journal Logo

Case Reports

Severe Pulmonary Hemorrhage in a Neonate With Hypoxic Ischemic Encephalopathy and Sepsis Managed on Extracorporeal Membrane Oxygenation

Cardona, Vilmaris Quinones*,†; Joshi, Swosti*,†; Conway, Daniel*,†; Menkiti, Ogechukwu R.*,†

Author Information
doi: 10.1097/MAT.0000000000001520
  • Free


Hypoxic ischemic encephalopathy (HIE) results from perinatal asphyxia leading to primary and secondary brain injury. Controlled hypothermia (CH) is the standard of care for neonatal HIE, as it improves neurodevelopmental outcomes; however, it can lead to impaired hemostasis, bradycardia, and increased systemic vascular resistance.1

Bleeding concerns are increased in patients with HIE on CH with respiratory failure refractory to medical management who require extracorporeal membrane oxygenation (ECMO) because of necessary anticoagulation. Pulmonary hemorrhage (PH) occurs in 1–12 per 1,000 live births with reported mortality as high as 50%.2 It is postulated to be the result of LV failure from asphyxia, stress failure of the pulmonary capillaries with breakage in the endothelial barrier and from significant ductal shunting with high pulmonary blood flow.2 PH on ECMO, though rare, may cause irreversible lung damage and management should be deliberate to prevent fatality.3 We present the therapeutic strategies in a term neonate with HIE undergoing CH with Enterobacter cloacae pneumonia and sepsis complicated by severe PH on ECMO.

Case Description

A term 3, 134 g female was born vaginally at 41 weeks of gestation to a 24 years old primigravid female. Delivery was complicated by meconium stained, foul-smelling amniotic fluid, and prolonged rupture of membranes. At birth, the infant was limp and apneic, requiring intubation and FiO2 1.0. Apgar scores were 2, 4, and 5, at 1, 5, and 10 minutes, respectively. Initial laboratory tests revealed profound metabolic acidosis with pH 6.9, white blood cell count of 3,900 cells/mcL and coagulopathy with PTT 58.6 seconds and INR 2.36. She was started on ampicillin and gentamicin for clinical sepsis and transferred to our institution with diagnosis of HIE for CH therapy.

Over the following 24 hours, she required escalation of cardiorespiratory support for worsening hypoxemia and persistent pulmonary hypertension of the newborn (PPHN), systemic hypotension, and lactic acidosis. She required inhaled nitric oxide, intravenous infusions of dopamine (20 µg/kg/min), dobutamine (20 µg/kg/min), vasopressin (80 milliunits/kg/hour), and hydrocortisone. Severe coagulopathy was treated with 30 ml/kg of fresh frozen plasma (FFP). Worsening respiratory failure led to broadening of antibiotics to include cefotaxime, escalation of ventilator support, and intratracheal administration of surfactant for meconium aspiration syndrome. Initial echocardiography confirmed severe PPHN with normal function. Screening head ultrasound (HUS) was normal.

Progressive respiratory failure (oxygenation index: 46), severe hypoxic spells, and pressor resistant hypotension refractory to maximal medical therapy resulted in the use of ECMO as rescue therapy. During cannulation, the infant had further desaturation events despite hand ventilation. Given normal coagulation profile before cannulation, a standard loading dose of 100 units/kg of heparin was given per ECMO guidelines. Desaturations progressed and frank blood was initially noted in the endotracheal tube (ETT) which subsequently extended rapidly into the ventilator tubing. Platelets, FFP, and red blood cell (RBC) transfusions were given emergently while the positive end expiratory pressure (PEEP) was increased to tamponade the pulmonary capillaries. The infant received one dose of intravenous epinephrine for ensuing bradycardia as the right internal jugular vein was cannulated with a 10Fr venous cannula and the right carotid artery with 8Fr arterial cannula. She was successfully placed on veno-arterial (VA) ECMO at 36 hours of life using a S5 roller pump (Germany, Munich, GmbH) with a venous compliance reservoir, Quadrox-i pediatrics, and neonatal oxygenator (Maquet Getinge Group, Rastatt, Germany), and a Tygon S-95-E 1⁄4-inch tubing (LivaNova, Colorado, Unites States). The circuit was primed with Plasma-Lyte A, pH 7.4, albumin, sodium bicarbonate, calcium gluconate, and RBCs.

Massive PH led to elective clamping of the ETT for 5 hours and transfusion of blood products including FFP 10 ml/kg every 8 hours for the first 24 hours. Activated clotting time goals were set at 180–200 seconds for 48 hours (Table 1). Inotropes were successfully weaned in the first 24 hours while milrinone was introduced for severely decreased biventricular function with ejection fraction of 19% and worsening PPHN on postcannulation echocardiography (Table 2).

Table 1. - Coagulation Parameters and Heparin Requirement
D1* D2 D3 D4+ D5 D6 D7 D8 D9
PT 18.3 15.3 14.3 14.8 13.8 12.9 13 13.1 12.6
INR 1.47 1.17 1.08 1.12 1.02 0.94 0.95 0.96 0.92
ACT (seconds) 93–269 185–262 190–206 181–244 220–240 230–265 234–265 222–256 240–256
PTT 114.8 50.7 49.8 75.7 82.6 61.1 69.4 81.6 83.3
Platelet 128 115 94 113 73 98 79 85 103
Anti-Xa (IU/mL) <0.1 <0.1 <0.1 0.11 0.24 0.12 <0.10 0.1 0.33
AT3 activity (%) 46 47 <15 63 79 77 84 79 90
Heparin (units/kg/min) 17–28 15–28 15–18 15–27 22–28 25–31 22–28 22–28 22–28
*Fresh frozen plasma given every 8 hours.
†Circuit change and furosemide started.
ACT, activated clotting time; Anti-Xa, Anti-Xa factor assay; AT3, antithrombin III activity; D, day; ECMO, extracorporeal membrane oxygenation; INR, international normalized ratio; PT, prothrombin time; PTT, partial thromboplastin time.

Table 2. - Serial Echocardiogram Results
PFO RVSP RV Function PDA LV Function
Pre-ECMO L→R shunt Unable to estimate Normal Large, bidirectional shunt Normal
D1* R→L shunt Unable to estimate Severely decreased Large, R→L shunt Severely decreased (EF: 18.9%)
D2 bidirectional Unable to estimate Moderately decreased Small, L→R shunt Normal
D3 bidirectional 83mmHg+RAP Mild to moderately decreased Moderate, bidirectional shunt Normal
D4 R→L shunt 90mmHg+RAP Mildly decreased Not demonstrated Normal
D5 R→L shunt 96mmHg+RAP Mildly decreased Small, R→L shunt Normal
D6 L→R shunt 48mmHg+RAP Mildly decreased Small to moderate, L→R shunt Normal
D7 L→R shunt 52mmHg+RAP Mildly decreased Small to moderate, L→R shunt Normal
D9 L→R shunt Unable to estimate Mildly decreased Closed Normal
Off ECMO L→R shunt Unable to estimate Mildly decreased Closed Normal
*Postendotracheal tube clamping for 5 hours and milrinone started.
Dornase alfa twice daily.
D, day; ECMO, extracorporeal membrane oxygenation; EF, ejection fraction; LV, left ventricle; L, left; PDA, patent foramen ovale; PFO, patent foramen ovale; R, right; RAP, right atrial pressure; RV, right ventricle; RVSP, right ventricular systolic pressure.

On ECMO day 4, increased metabolic demands and oxygen consumption were evident by a decline in SVO2 to 58%, increase in sweep gas flow to 0.7 L/min despite sedation and ECMO pump flow of 145 ml/kg/min (Table 3). This prompted repeat cultures and broadened antibiotic coverage to cefepime. Lung conditioning and diuresis with furosemide were introduced for severe atelectasis and fluid overload, respectively. Later that day, a large clot was noted in the oxygenator warranting a circuit change. The patient did not tolerate this brief separation from ECMO and required bolus epinephrine for severe bradycardia. The ECMO circuit culture drawn returned positive for Enterobacter cloacae and cefepime treatment was continued.

Table 3 - ECMO Parameters and Oxygen Consumption
D1* D2 D3 D4 D5 D6 D7 D8 D9
Max pump flow (ml/kg/min) 125 134 145 120 120 153 132 129 83
FiO2 100 100 100 100 100 100 85-95 35-85 30
SVO2 (%) 65–86 80–87 71–86 58–72 50–76 50–76 70–81 77–84 67–83
Max sweep gas flow (L/min) 0.25 0.3 0.6 0.7 0.45 0.7 0.7 0.5 0.1
Lactate (mmol/L) 8.7 5.6 2.4 2.8 1.4 1.2 1.2 1 1.1
O2 consumption (ml/min) 10 13 11 18 24 24 19 10 12
O2 consumption (ml/kg/min) 3.3 4.2 3.6 5.8 7.8 7.9 6.2 3.2 4.2
*Paralyzed postcannulation
Antibiotics expanded and blood culture positive for Enterobacter cloacae. Circuit changed.
O2 consumption on VA ECMO estimated by cardiac output (pump flow) x 1.34 x hemoglobin x (outlet saturation-inlet saturation).
D, day; ECMO, extracorporeal membrane oxygenation; FiO2, fraction inspired oxygen; O2, oxygen; SVO2, mixed venous O2 saturation.

Despite robust diuresis, lung recruitment maneuvers and modified pulmonary toileting to avoid recurrence of PH, serial chest radiographs revealed persistent consolidation and air bronchograms likely due to Enterobacter cloacae pneumonia (Figure 1). Off-label use of dornase alfa 1.25 mg and albuterol twice daily for 48 hours with pulmonary toilet was done before contemplating bronchoscopy. This resulted in marked improvement in lung expansion and PPHN avoiding the need for bronchoscopy (Figure 1, Table 2). ECMO support was weaned and after a successful trial off, she was decannulated on ECMO day 9.

Figure 1.:
Serial chest and abdominal radiographs. Serial chest and abdominal radiographs. (A) Precannulation, (B) postcannulation, (C) ECMO day 6 (predornase alfa), (d) ECMO day 7 (24 hours after first dose of dornase alfa). ECMO, extracorporeal membrane oxygenation.

Brain magnetic resonance imaging performed at 14 days old demonstrated a grade two intraventricular hemorrhage, an interval increase from HUS grade 1 immediately after cannulation. She was discharged home in room air taking fortified oral feeds at 1 month of life.


Pulmonary hemorrhage in premature infants has been associated with respiratory distress syndrome and surfactant use, whereas in term neonates it has been related to low Apgar scores.4 Low birth weight, intraventricular hemorrhage, heart failure, and sepsis have been noted to increase PH-related mortality.5 Although a term infant, our patient had multiple risk factors for PH and PH-related mortality including perinatal asphyxia, coagulopathy, surfactant therapy, sepsis, heart failure, and IVH. The decompensation before cannulation suggests an acute change in lung mechanics likely related to evolving PH that became grossly evident during cannulation. It is important to note that the coagulation profile 1 hour before cannulation was normal; therefore, heparinization alone cannot explain the massive PH in this case. Postcannulation echocardiography revealed severely diminished left ventricular (LV) systolic function, which may have contributed to an increase pulmonary pressure and PH in the face of the discussed risk factors. VA ECMO allowed us to provide full support without dependence on the right ventricle and pulmonary blood flow while the heart and lungs recovered.

Although PH in our patient happened before ECMO initiation, bleeding on ECMO is multifactorial. In addition to bleeding risk from anticoagulation, dilutional coagulopathy and thrombocytopenia can occur if the ECMO circuit is primed with crystalloid solutions or RBCs.2 Our institutional guidelines call for priming the circuit with RBCs and transfusing patient plasma, platelets and RBCs as needed during cannulation. Due to massive PH, our patient received all products in addition to scheduled FFP transfusions for 24 hours. Additionally, the activated clotting time target was adjusted from our usual 200–220s to 180–200s, allowing for lower heparin use. Aminocaproic acid and recombinant factor VII were deferred for circuit thrombosis concerns and inability to engage the hemorrhaging lungs enough to survive a close interval circuit change. Inhalational therapies, like tranexaminic acid, were not considered given massive PH obstructed the endotracheal tube making it nearly impossible to deliver effectively.

Increasing PEEP and clamping the ETT were strategic mechanical interventions used to temporize bleeding while aggressively correcting recurrent coagulopathy. Clamping the ETT is described in adults with PH as supportive strategy during ECMO.3 However, this strategy has risks. In the setting of PPHN, complete lung collapse without delivering PEEP or delta pressure, traditionally provided in rest settings, results in significant intrapulmonary shunting which can lead to right ventricle failure requiring higher pump flow to meet metabolic demands. Furthermore, on VA ECMO high pump flow results in increased afterload which would be problematic on the LV as it leads to LV distension and increased wall stress.6 Increasing pump flow in this patient would have further compromised the struggling LV predisposing to LV failure, LV thrombosis or PH. Striking this balance was vital in our patient, reason why we planned a limited 5 hour ETT clamping and avoided this complication.

To our knowledge, only one other neonatal case has reported using ETT clamping for PH on ECMO. Vobruba et al., 2017, described a septic neonate with PH who required ETT clamping for 63 hours followed by multiple bronchoscopies to clear the lungs.7 In our case we were able to institute a noninvasive approach for mucus and clot clearance with dornase alfa, in light of pneumonia and PH, which allowed us to avoid bronchoscopy. Rescue treatment with dornase alfa has been shown to help with atelectasis and tracheal plugging in neonates.8,9 Benefits from this approach outweighed the concern for dislodging a clot as coagulopathy had been corrected and more invasive procedures such as bronchoscopy were avoided.

Although PH-related mortality is high, strategic interventions including ETT clamping, aggressive correction, and monitoring of coagulation parameters and pulmonary toileting with dornase alfa were crucial in the successful management of this septic neonate with PH on VA ECMO.


1. Jacobs SE, Berg M, Hunt R, Tarnow-Mordi WO, Inder TE, Davis PG: Cooling for newborns with hypoxic ischaemic encephalopathy. Cochrane Database Syst Rev. 31: CD003311, 2013.
2. Zahr RA, Ashfaq A, Marron-Corwin M. Neonatal pulmonary hemorrhage. Neoreviews. 13: e302–e306, 2012.
3. Willers A, Swol J, Kowalewski M, et al. Extracorporeal life support in hemorrhagic conditions: a systematic review. ASAIO J. 7: 476–484, 2020.
4. Bhandari V, Gagnon C, Rosenkrantz T, Hussain N: Pulmonary hemorrhage in neonates of early and late gestation. J Perinat Med. 27: 369–375, 1999.
5. Li L, Yu J, Wang J, 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. 43: 997–1003, 2008.
6. Cevasco M, Takayama H, Ando M, Garan AR, Naka Y, Takeda K: Left ventricular distension and venting strategies for patients on venoarterial extracorporeal membrane oxygenation. J Thorac Dis. 11: 1676–1683, 2019.
7. Vobruba V, Grus T, Mlejnsky F, Belohlavek J, Hridel J, Lambert L. Management of severe pulmonary hemorrhage in a neonate on veno-arterial ECMO by the temporary clamping of the endotracheal tube—a case report. Perfusion. 33: 77–80, 2018.
8. El Hassan NO, Chess PR, Huysman MW, Merkus PJ, de Jongste JC: Rescue use of DNase in critical lung atelectasis and mucus retention in premature neonates. Pediatrics. 108: 468–470, 2001.
9. Erdeve O, Uras N, Atasay B, Arsan S: Efficacy and safety of nebulized recombinant human DNase as rescue treatment for persistent atelectasis in newborns: case-series. Croat Med J. 48: 234–239, 2007.

pulmonary hemorrhage; extracorporeal membrane oxygenation; neonate

Copyright © ASAIO 2021