We report the first known case of acute upper and lower airway obstruction due to retropharyngeal haemorrhage secondary to acquired haemophilia A. The patient required tracheal intubation and ventilation followed by surgical tracheostomy, prior to transfer to the intensive care unit for management of her coagulopathy with recombinant clotting factor products under the care of the haematologists.
Acute upper airway obstruction is a potentially life-threatening event requiring urgent intervention as a life-saving measure. Retropharyngeal haematoma is a rare cause of upper airway obstruction and is usually secondary to trauma, anticoagulation or more rarely aortic dissection . Acquired haemophilia A is a rare autoimmune disorder that has not previously been described in the aetiology of retropharyngeal haemorrhage .
We report the case of a 75-year-old female who presented with acute upper airway obstruction secondary to extensive mediastinal and retropharyngeal haematoma as a result of underlying acquired haemophilia A.
A 75-year-old female patient was referred to the on- call ENT services at our hospital with a 4-day history of worsening sore throat, dysphagia and intermittent chest pain. Her general practitioner had initially treated her with cefotaxime for suspected supraglottitis without success. In her past medical history she had undergone an aorto-bifemoral bypass and was hypertensive but was otherwise in good health. The patient was taking aspirin 75 mg once daily (OD) and dipyrimidol 100 mg three times daily (TDS) but was not formally anticoagulated. She was also taking simvastatin 20 mg daily and meloxicam 7.5 mg daily for osteoarthritis.
On presentation the patient had inspiratory stridor, was distressed, agitated, tachypnoeic and hypoxaemic breathing room air (peripheral oxygen saturations were 85%). The oxygen saturation improved to 93% and agitation was reduced by the use of high flow oxygen, nebulised adrenaline and salbutamol. Intravenous access was established and dexamethasone and cefotaxime were administered. Chest examination revealed decreased air entry at the left base with coarse crepitations but was otherwise unremarkable. Arterial blood gases taken at this time showed a Po2 of 8.9 kPa and a Pco2 of 6.5 kPa.
The patient tolerated nasendoscopy with difficulty, which revealed haemorrhagic, swollen pharyngeal tissues with some superficial sloughing. The vocal cords were not visualised, although the epiglottis and false cords were noted to be grossly distorted due to haematoma. Shortly after this the oxygen saturation fell to 85% and she again became more distressed, at which point anaesthetic help was requested and she was transferred to the operating theatre for further management.
On arrival in the operating theatre, the decision was made to intubate the trachea. With ENT surgeons present, a gas induction was attempted using 8% sevoflurane in 100% oxygen with the patient sitting in the most comfortable position for her. Direct laryngoscopy confirmed grossly abnormal pharyngeal anatomy and, although the epiglottis was visualised, only the arytenoids were recognisable below this. Suxamethonium 100 mg was given and, following the insertion of a bougie, the trachea was intubated using a size 6.5 tracheal tube (ETT); its position was confirmed using capnography. Neuromuscular blockade was achieved using 0.5 mg.kg−1 atracurium and anaesthesia was maintained using isoflurane in oxygen using volume-controlled intermittent positive pressure ventilation. Oxygen saturations improved to 95% and inflation pressures were between 20 and 25 cm H2O with a tidal volume of 500 ml. Haemodynamic parameters were stable throughout induction of anaesthesia and the patient was transferred to theatre without incident.
In the operating theatre, surgical exploration of the oropharynx and upper airway revealed extensive retropharyngeal haematoma throughout the nasopharynx, oropharynx and laryngopharynx including the vocal cords and the early portion of trachealis. The extent of upper airway obstruction necessitated the siting of a surgical tracheostomy, which was performed uneventfully. Flexible bronchoscopy was then performed to confirm position and for the purposes of tracheal toilet. Bronchoscopy proved to be technically difficult due to significant ingress of trachealis into the tracheal lumen at all levels from the tracheostomy down to the carina.
The tracheostomy was secured, an arterial line was sited and the patient was transferred, sedated and ventilated, to the ICU for further management. On the ICU a central line was sited and sedation was maintained using propofol and alfentanil infusions. Hypotension required treating with intravenous fluids and noradrenaline. Early investigations revealed a coagulopathy with an activated partial thromboplastin time ratio (APTTR) of 2.2 and an International Normalised Ratio (INR) of 1.2 which was treated initially using fresh frozen plasma (FFP) and platelets. With concerns over the extent of the haematoma in the upper airway, the patient was then transferred to the computed tomography (CT) scanner where a high resolution CT scan with contrast was performed on the head, neck and thorax (Figs 1 and 2). This demonstrated extensive mediastinal haematoma and oedema associated with compression of the trachea at all levels from the carina distally through to the nasopharynx. It was determined radiologically that this was venous bleeding and following a cardiothoracic opinion it was felt that associated aortic pathology could be excluded.
Bleeding persisted from around the tracheostomy site despite the clotting products, whereupon further haematological advice was sought. The post-FFP APTTR showed a deterioration in the coagulopathy at 2.98 with an INR of 1.6.
A 50/50 test with the patient's plasma and donor control plasma showed only a partial correction of the clotting defect. Factor VIII screens were undertaken on pre-FFP samples. These investigations revealed Factor VIII levels to be only 2% of the normal value. Other intrinsic factors were assayed and came back as normal (Factor IX 107%, Factor XI 71%, Factor XII 76%).
Factor VIII inhibitor assay was found to be strongly positive (Bethesda value of 67; normal is < 1), and the diagnosis of Acquired Haemophilia A was made.
The patient was treated initially using two doses of recombinant Factor VIIa 4 h apart. Tranexamic acid 1 mg was given to stabilise the clot along with methylprednisolone at 100 mg daily for immunosuppression. Factor VIIa therapy was switched to twice daily ‘factor eight inhibitor bypassing activity’ (FEIBA) after 24 h. This was reduced to a once daily regimen after eight days of treatment. The steroids were continued until the inhibitor assay returned to normal.
Sedation was stopped 24 h after admission to the ICU and was weaned from full ventilatory support to continuous positive airway pressure (CPAP) over the course of the next 48 h.
A repeat CT scan 1 week later (Fig. 3) showed a marked reduction in the amount of haematoma in the mediastinum, although the haematoma in the pharyngeal tissue remained to a greater extent. The coagulopathy was still present with an APTTR of 1.9, 10 days post presentation, although the Factor VIII assay had improved to 11%. The Factor VIII inhibitor assay was also tending towards normal with a Bethesda value of 22. There was still some oozy bleeding from the respiratory tract, which complicated full weaning, and consequently the patient continued to receive high dose steroids and FEIBA. The trachea was successfully decannulated on day 11 and she was discharged to the ward for further investigation and management of her coagulopathy.
In the first instance this patient presented a difficult challenge in terms of managing her acutely obstructing upper airway as not only was the diagnosis unclear at the time, but at presentation, she was highly agitated and unco-operative. It was this agitated state which severely limited the therapeutic options open to us. The findings at nasendoscopy suggested that even if she complied with an endoscopic procedure it was likely that the endoscope itself would obstruct her remaining airway. We also felt an awake tracheostomy under local anaesthesia would have been impossible as it requires a calm, compliant patient for a reasonable chance of success. The same is true of emergency percutaneous cricothyroidotomy devices. Furthermore, an emergency attempt at using a blind Quick-trak or other such rapid access device could have caused direct haemorrhage into the airway. We felt, in the circumstances, the only option left open to us was to attempt a gas induction with the ENT surgeons in close attendance, should an emergency tracheostomy have been required.
The unexplained extensive haematoma in the upper airway, coupled with the bizarre findings seen on bronchoscopy and her previous history of vascular surgery, led us to believe that there might be further bleeding in the mediastinum. When the CT scan confirmed these suspicions but ruled out an aortic source, it became obvious that there must be some other mechanism at work.
The incidence of acquired haemophilia in the UK has recently been put at around 1.5 cases per million per annum . The majority present with subcutaneous haematomas [3, 4] and fatal haemorrhage occurs in around 9% of cases. These deaths usually occur due to uncontrolled gastro-intestinal or pulmonary haemorrhage. We believe that upper and lower airway obstruction secondary to retropharyngeal bleeding associated with acquired haemophilia A has never previously been described.
Acquired haemophilia A appears to be idiosyncratic in around 63% of cases , although it has been shown to be associated commonly with malignancy, pregnancy  and certain autoimmune diseases (SLE/rheumatoid arthritis) and pemphigus. The mortality rate with first presentation of acquired haemophilia A is 8%. It is diagnosed on Factor VIII inhibitor assay, usually following deranged clotting factor screen results and always with the direct involvement of a haematologist in the UK. In this case it was the worsening coagulopathy following initial treatment with FFP that prompted further investigation.
A 50/50 test with the patient's plasma and donor control plasma showed only a partial correction of the clotting defect. If the problem was due to pure factor deficiency then this should have corrected completely. Partial correction points instead to the presence of a factor inhibitor. This is usually seen with lupus anticoagulant; however, in this case, as the clinical condition was that of bleeding and not of thrombosis, a lupus anticoagulant was thought unlikely. This was borne out by the further factor screens and a Factor VIII inhibitor assay, which ultimately delivered the diagnosis.
The treatment of initial bleeding in acquired haemophilia A includes three options:
• porcine Factor VIII with an 80% success rate ;
• FEIBA – Factor VIII inhibitor assay with a 76% success rate [7, 8];
• recombinant Factor VIIa with a 100% success rate .
Recombinant Factor VIIa was developed to stop bleeding in haemophilia A patients who developed antibodies to exogenous Factor VIII. The most important initiator of the clotting cascade is tissue factor binding with Factor VII. This activates Factor VIIa, which drives the extrinsic pathway to produce thrombin (IIa) and thereby conversion of fibrinogen to fibrin. Therefore by flooding the extrinsic pathway, as the intrinsic pathway is inactivated due to the inhibitor, clot formation can still be achieved.
Our patient was initially treated with two doses of recombinant Factor VIIa 4 h apart, the half life of which is 2½ h, before switching to FEIBA 24 h after the initial bleed; FEIBA has a longer half life and is less expensive than recombinant Factor VIIa. This was reduced to a once daily dose after eight days of treatment. Alongside the haematological management of clotting abnormalities it is also essential to modulate the immune process driving the pathology. High dose steroids were given at presentation and were continued until the inhibitor assay returned to normal. Other possible treatments include cyclophosphamide and intravenous immunoglobulin with steroid ; however, there is no firm evidence to demonstrate any improvement in efficacy with this approach [1, 10].
In conclusion, we have described the treatment of the first documented case of acquired haemophilia A presenting with acute upper airway obstruction.