Extracorporeal Membrane Oxygenation for Diffuse Alveolar Hemorrhage Caused by Idiopathic Pulmonary Hemosiderosis: A Case Report and a Review of the Literature

Abstract

Diffuse alveolar hemorrhage (DAH) is a life-threatening condition presenting with hemoptysis, anemia, and diffuse radiographic pulmonary infiltrates; it causes acute respiratory failure. Idiopathic pulmonary hemosiderosis (IPH) is a rare cause of DAH occurring predominantly in children. Bleeding is often considered to be a contraindication for extracorporeal membrane oxygenation (ECMO) due to systemic anticoagulation. We present an 8-year-old girl with DAH caused by IPH. Unfractionated heparin was administered to maintain an activated clotting time of 150 to 180 seconds. The DAH resolved with immunosuppressive therapy, and the patient survived to decannulation. ECMO may be applied as a rescue therapy for DAH even with systemic anticoagulation.

Introduction

Diffuse alveolar hemorrhage (DAH) is a life-threatening condition presenting with a clinical constellation of hemoptysis, anemia, and diffuse radiographic pulmonary infiltrates; it causes acute hypoxemic respiratory failure. ¹ The underlying causes are classified by the presence or absence of pulmonary capillaritis. ² Idiopathic pulmonary hemosiderosis (IPH) is a rare cause of DAH occurring predominantly in children. It is characterized by diffuse infiltrates, iron deficiency anemia, and hemosiderin-laden macrophages in the sputum, bronchoalveolar lavage (BAL) fluid, or lung tissue; it is not associated with pulmonary capillaritis. ² Progressive acute respiratory failure from IPH can lead to mortality despite maximal medical therapy, including invasive mechanical ventilation, and immunosuppressive therapy. Prolonged use of medical devices and excessive immunosuppression together may cause a secondary infection, which can also lead to mortality.

Extracorporeal membrane oxygenation (ECMO) for bleeding is historically contraindicated due to the systemic anticoagulation required to avoid clot formation in the ECMO circuit and thromboembolism. ³ However, there is an increasing number of reports about the treatment for bleeding in a limited variety of conditions, such as DAH ⁴ and acute respiratory distress syndrome, due to multiple trauma. ⁵ However, only a few instances of DAH in children caused by IPH and treated with ECMO have been reported. ^{6 7} We report, herein, a case of an 8-year-old girl with severe respiratory failure caused by DAH who underwent ECMO.

Case Report

An 8-year-old girl with a history of bronchial asthma and iron deficiency anemia was seen at another hospital for cough, wheeze, and mild hemoptysis at 2-day duration. She was tachypneic and her peripheral oxygenation saturation (SpO ₂ ) was 68% without oxygen. Her chest X-ray revealed bilateral diffuse infiltration. She was transferred to a tertiary hospital near her residence, intubated, and mechanically ventilated. Antibiotics and high-dose methylprednisolone were administered, and plasma exchange was performed for suspected pulmonary capillaritis. Despite intensive therapy, her respiratory condition continued to deteriorate over the next several hours, and she was transferred to our hospital, the National Center for Child Health and Development (NCCHD), for further workup and possible treatment by ECMO.

Upon arrival to our hospital, she was placed on pressure-controlled conventional mechanical ventilation with a peak inspiratory pressure of 45 cm H ₂ O, positive end-expiratory pressure of 14 cm H ₂ O, and fraction of inspired oxygen (F _I O ₂ ) of up to 1.0 but showed severe respiratory effort and poor chest rise. The tracheal aspirate was bloody, and bilateral diffuse crackles were heard on auscultation. Her vital signs were the following: heart rate of 166 beats/min, blood pressure of 90/50 mm Hg, and SpO ₂ of approximately 90%. Blood gas analysis showed severe hypoxemia (pO ₂  = 75 mm Hg, PaO ₂ /FiO ₂ [P/F] ratio = 75, oxygenation index = 36), and severe respiratory acidosis (pH = 6.966, pCO ₂  = 313 mm Hg). Complement and anti-double-stranded DNA antibody levels were within the normal range. Further investigation showed negative autoimmune antibodies to anti-nuclear antibody (ANA), anti-neutrophil cytoplasmic antibody (ANCA), and anti-glomerular basement membrane (GBM). A polymerase chain reaction analysis of the tracheal aspirate for viruses and mycobacterium was also negative. The chest X-ray and computed tomography are shown in Fig. 1 .

Fig. 1.

Chest X-ray and CT scan. ( A ) Chest X-ray demonstrates bilateral pulmonary infiltrates. ( B , C ) CT scan shows lobar ground-glass opacities and predominant consolidation in both lungs. CT, computed tomography.

With the consent of her parents, the patient was placed on venovenous (VV) ECMO (Heart lung machine HAS, Senko Medical Instrument Mfg., Tokyo, Japan). A 14 Fr drainage cannula (FEM-II, Edwards Life Science, Irvine, CA, United States) was inserted from the internal jugular vein to the right atrium and a 12 Fr reinfusion cannula (FEM-II, Edwards Life Science, Irvine, CA, United States) was inserted from the left femoral vein percutaneously to achieve 1.2 L/min of blood flow with sweep gas at 6 L/min and F _I O ₂ of 1.0. Systemic heparin was used for anticoagulation to keep the activated clotting time (ACT) within 150 to 180 seconds, and platelets were transfused to maintain ≥ 100,000/mm ³ . Prothrombin time/international normalized ratio (PT/INR), fibrinogen, and antithrombin levels were also monitored to maintain values of < 1.5, ≥ 200 mg/dL, and ≥ 70%, respectively. Packed red blood cells (RBCs), platelets, and fresh frozen plasma were transfused as appropriate. Laboratory data regarding coagulation and doses of heparin before and during the first ECMO treatment are shown in Table 1 . The mechanical ventilation was switched to high frequency oscillation ventilation with 22 cm H ₂ O of mean airway pressure. High-dose methylprednisolone therapy for a total of 3 days and plasma exchange for a total of 4 days were continued to treat a suspected underlying autoimmune pathophysiology without a definitive diagnosis. Broad-spectrum antibiotics were also administered until the tracheal aspirate culture returned negative. The alveolar hemorrhage started to improve remarkably after 5 days, and the patient was successfully weaned from ECMO after 11 days.

Table 1. Laboratory tests regarding coagulation and dose of heparin infusion before and during the first extracorporeal membrane oxygenation.

Days on first ECMO	Unit	Before ECMO	1	2	3	4	5	6	7	8	9	10	11
ACT	(sec)	NA	163	158	155	158	164	164	158	161	159	164	166
Platelets	(×10 3 /µL)	299	148	112	141	96	154	139	142	162	123	135	130
PT/INR		1.37	1.39	1.53	1.44	1.43	1.4	1.33	1.31	1.21	1.27	1.3	1.25
aPTT	(sec)	30	81.4	111	107	96.6	120	86.2	88.8	71.3	80.2	70.7	81.6
Fibrinogen	(mg/dL)	415	244	253	267	236	248	296	296	296	309	273	312
D-dimer	(µg/mL)	28.9	8	15.1	8.9	6.3	12.7	11	15	7.2	5.9	4.6	4.1
Antithrombin	(%)	NA	72.1	74.2	NA	75.7	75.4	75.3	85.7	98.7	110.5	118.3	NA
Dose of heparin	(Units/kg/h)		15	20	40	35	31	26	30	28	24	27	30

Abbreviations: ACT, activated clotting time; aPTT, activated partial thromboplastin time; ECMO, extracorporeal membrane oxygenation; NA, not applicable; PT/INR, prothrombin time/international normalized ratio.

Six days after ECMO removal, the alveolar hemorrhage worsened with leukocytosis and elevated C reactive protein which led gradually to respiratory compromise. She was again placed on VV-ECMO the next day, followed by conversion to venoarterial ECMO (16 Fr right internal jugular vein [PCKC, Toyobo, Osaka, Japan], 14 Fr left femoral vein, 14 Fr common carotid artery; blood flow 1.4 L/min; F _I O ₂ 1.0, sweep gas 3 L/min) due to right cardiac failure from secondary pulmonary hypertension. Systemic anticoagulation and laboratory test targets were set at the same ranges as in the first ECMO treatment; the monitored values are shown in Table 2 . High-dose methylprednisolone and plasma exchange were restarted, and hydroxychloroquine was added. The alveolar hemorrhage and pulmonary hypertension improved significantly, and she was decannulated 12 days after the second ECMO run. Even though she encountered no complications during two courses of ECMO, she experienced sepsis due to gastrointestinal mucormycosis after discontinuation of ECMO and died on hospital day 39.

Table 2. Laboratory tests regarding coagulation and dose of heparin infusion before and during the second ECMO.

Days on second ECMO		Before ECMO	1	2	3	4	5	6	7	8	9	10	11	12	13
ACT	(sec)	NA	182	169	169	163	162	152	160	162	160	163	161	161	160
Platelets	(×10 3 /µL)	323	130	131	158	136	125	113	118	114	120	133	142	111	128
PT/INR		1.17	1.52	1.41	1.23	1.22	1.18	1.36	1.47	1.31	1.27	1.27	1.29	1.27	1.28
aPTT	(sec)	28.8	121	114	105	90.3	52.6	65.6	88.3	52.4	59.1	92.7	54.2	56.6	64.7
Fibrinogen	(mg/dL)	480	290	404	302	205	202	240	373	394	341	366	427	511	480
D-dimer	(µg/mL)	17.9	2.9	1.4	1.1	2.1	4	5.3	2.7	3.5	2.2	3.4	7.2	8.1	4.9
Antithrombin	(%)	NA	74.8	NA	NA	NA	NA	68	69.8	82.5	93.5	106	87.1	83.9	69.5
Dose of heparin	(Units/kg/h)		20	30	26	26	28	26	21	27	28	30	27	29	29

Abbreviations: ACT, activated clotting time; aPTT, activated partial thromboplastin time; ECMO, extracorporeal membrane oxygenation; NA, not applicable; PT/INR, prothrombin time / international normalized ratio.

Her BAL fluid during the second ECMO showed hemosiderin-laden macrophages, and a necropsy of the lung revealed no sign of capillaritis ( Fig. 2 ). After other etiologies, such as infection, known autoimmune diseases, and cardiovascular causes were excluded, the patient's DAH was seen to be consistent with IPH.

Fig. 2.

Histopathological findings of necropsy samples. ( A ) Hematoxylin eosin stain and ( B ) Berlin blue stain demonstrate hemosiderin-laden macrophages in the alveolar spaces and no sign of capillaritis.

Discussion

Diffuse alveolar hemorrhage is a clinicopathological syndrome characterized by usually diffuse but sometimes focal bleeding from the alveolar capillaries and is differentiated from bleeding from the bronchial circulation. ⁸ The condition presents with hemoptysis, anemia, and pulmonary infiltrates on chest radiographs and is sometimes life-threatening. ^{1 8} Because one-third of cases lack hemoptysis, BAL is generally required to confirm diagnosis based on the increasing RBC count in serial BAL samples. Once the diagnosis is established, the underlying etiology must be determined to start treatment. In our case, BAL was first performed while on the second ECMO. However, it might be better to perform it earlier because BAL is required both to confirm the diagnosis and to differentiate the cause of DAH, as it yields good negative predictive value for a pathogen, such as Pneumocystis jirovecii . ⁹ Exploration of extrapulmonary involvement in addition to serologic autoantibodies, infectious organisms, and diagnostic imaging are essential. DAH in children is classified by the presence or absence of pulmonary capillaritis. ² A lung biopsy is sometimes indicated in DAH of indeterminate cause.

Idiopathic pulmonary hemosiderosis is a rare cause of alveolar hemorrhage which occurs predominantly in children. Hemosiderin-laden macrophages in BAL samples or lung tissue are suggestive, but not specific to IPH, and other etiologies, such as vasculitis, must be excluded. IPH is not associated with pulmonary capillaritis and the involvement of organs other than the lungs is absent. Corticosteroids are mainly chosen as the first-line therapy and other immunosuppressants, such as hydroxychloroquine, mycophenolate mofetil, and azathioprine, are sometimes added, given the suspected association with autoimmune diseases. ¹⁰ To our knowledge, there are only two case reports of IPH successfully treated with ECMO. ^{6 7} While both cases were maintained by systemic heparin for anticoagulation, the DAH was successfully controlled with corticosteroid and the patients were weaned off ECMO without any bleeding complications.

Extracorporeal membrane oxygenation for DAH treatment was first reported by Siden et al ¹¹ in 1994 and an increasing number of cases successfully treated with ECMO are being reported both among adults and children. In a literature review of 21 adult cases, ⁴ ANCA-associated vasculitis was the most common etiology, followed by systemic lupus erythematosus. As reported, 95% of patients were successfully decannulated from ECMO and 90% survived to discharge. To our knowledge, there are 26 published cases ^{6 7 11 12 13 14 15 16 17 18 19 20 21 22 23} of DAH managed with ECMO in children, including our case ( Table 3 ). The most common etiology was IPH followed by a cardiac etiology. All the cases were treated with anticoagulation but were weaned off ECMO. Bleeding complications were reported in two cases, that is, a mediastinal hemorrhage after a median sternotomy was reported in one case, ¹⁹ and transient exacerbation of alveolar hemorrhage in IPH was reported in the other case. ⁷ Bleeding was eventually controlled in both cases and the patients survived to decannulation. These findings suggest that ECMO may be strongly indicated for refractory respiratory failure caused by DAH unresponsive to maximal therapy.

Table 3. Summary of pediatric cases of DAH treated with ECMO.

Age a	6.5	(0.1–13)
Male sex (%)	12	(46)
Etiology (%)
IPH	5	(19)
Cardiac/CPB use	4	(15)
WG	3	(12)
SLE	3	(12)
MPA	2	(8)
Sepsis	2	(8)
Others	7	(27)
P/F ratio before ECMO a ( n  = 14)	52	(44–61)
Mode of ECMO (%) [ n  = 24]
VV	13	(50)
VA/BiVAD + ECMO	13	(50)
Duration of ECMO a ( n  = 23)	5	(4–6.2)
Anticoagulation (%) [ n  = 22]	22	(100)
ACT target range a ( n  = 17)
Upper limit	180	(180–190)
Lower limit	160	(160–160)
Survival to decannulation (%) [ n  = 17]	26	(100)
Survival to discharge (%) [ n  = 25]	24	(96)

Abbreviations: ACT, activated clotting time; BiVAD, bi-ventricular assist device; CPB, cardiopulmonary bypass; DAH, diffuse alveolar hemorrhage; ECMO, extracorporeal membrane oxygenation; IPH, idiopathic pulmonary hemosiderosis; MPA, microscopic polyangiitis; P/F ratio, PaO ₂ /FiO ₂ ratio; SLE, systemic lupus erythematosus; VA, venoarterial; VV, venovenous; WG, Wegener's granulomatosis

^aData are shown as the median (interquartile range).

Anticoagulation management varies widely among ECMO centers ^{24 25} because no optimal regimen exists due to lack of controlled trials. In the reports of ECMO used for the treatment of DAH, the ACT target was set at various ranges from 130 to 150 seconds ¹⁷ to 200 to 220 seconds. ²¹ A strategy for avoiding systemic anticoagulation for a short period until hemostasis is achieved may be a choice. ²⁶ However, vigilance is required because blood flow during ECMO is slower in children than in adults and may predispose pediatric patients to developing thrombosis. In a case of DAH caused by a silicone embolism, the ECMO circuit without systemic anticoagulation had to be changed 36 hours after commencement due to high line pressure and visible clot formation. ²⁷ In our case, ACT, platelet, PT/INR, fibrinogen, and antithrombin targets were set at 150 to 180 seconds, ≥ 100,000/mm ³ , ≥ 1.5, ≥ 200 mg/dL, and ≥ 70%, respectively, without any complications. We have also monitored activated partial thromboplastin time (aPTT) to maintain values above normal range; however, we guided anticoagulation depending mainly on ACT because it requires less blood sampling and the baseline aPTT and the effects of heparin on aPTT vary significantly with age. ²⁸ Further study is needed to find the optimal method for monitoring and managing this condition.

We lost the patient to gastrointestinal mucormycosis which is one of the most severe forms of opportunistic infection. Only 25% of cases are diagnosed antemortem and the mortality rate reaches as high as 85%. ²⁹ Because Mucorales species cannot be detected in a blood culture, histopathological examination remains essential. ³⁰ Mucormycosis occurs most often in the patients with underlying diseases, such as malignancies, solid organ transplantations, and disorders or treatments, leading to immunosupression. ³⁰ Our patient was at high risk of fungal infection due to the extensive use of broad-spectrum antibiotics and aggressive immunosuppression. It should be noted that while immunosuppressive therapy may be essential for controlling the underlying disease, it also poses a substantial risk of opportunistic infections.

Conclusion

Although bleeding is often considered to be a contraindication for ECMO use, our case and previously published data suggest that ECMO may be applied as a rescue therapy for DAH even with systemic anticoagulation, allowing time for the diagnosis and treatment of the underlying disease. However, the optimal method of monitoring the condition and the target range for anticoagulation have yet to be determined.