Causes and management of massive haemoptysis

Haemoptysis and generic bleeding from the airway has come up a couple of times in this exam. 

• Question 2 from the first paper of 2012 asks specifically for possible causes and management in a guy who presents via the ED (i.e. it could be anything).
• Question 16 from the second paper of 2022 asked for investigations which could be used to localise the site of bleeding.
• Question 7 from the first paper of 2017 also asked for a list of possible causes, except this time the airway bleeding is from a recently fashioned tracheostomy, and you can't call it "haemoptysis" because strictly speaking the Greek word "ptusis" means "to spit", and the trache patient's bloody cough is bypassing the mouth and lips.

A huge list of differentials for the causes of haemoptysis is presented by the college in their model answer to Question 2 from the first paper of 2012. An even larger list is presented in the article on the role of bronchoscopy in the management of massive haemoptysis. I reproduce this table here.

Causes of Haemoptysis

Infectious Mycobacteria (particularly tuberculosis) Fungal infections (including mycetoma) Necrotizing pneumonia and lung abscess (Klebsiella pneumoniae, Pseudomonas aeruginosa,Staphylococcus aureus, Streptococcus pneumoniae, other Streptococcus spp. and Actinomyces spp.) Bacterial endocarditis with septic emboli Parasitic (paragonimiasis, hydatid cyst) Neoplastic Bronchogenic carcinoma Endobronchial tumors (carcinoid, adenoid cystic carcinoma) Pulmonary metastases Sarcoma Pulmonary Bronchiectasis (including cystic fibrosis) Chronic bronchitis Alveolar hemorrhage and underlying causes Vascular Pulmonary artery aneurysm (Rasmussen aneurysm, mycotic, arteritis) Bronchial artery aneurysm Pulmonary infarct (embolism) Pulmonary hypertension Congenital cardiac or pulmonary vascular malformations Airway-vascular fistula Arteriovenous malformations Mitral stenosis Left-ventricular failure

Vasculitis Wegener's granulomatosis Goodpasture's syndrome Behçet's disease Systemic lupus erythematosus Trauma Induced by diagnostic bronchoscopy (brushing/biopsy) Related to interventional pulmonology procedures (dilation, metallic stent placement, high-dose brachytherapy) Catheter-induced PA rupture Blunt or penetrating chest injury Transtracheal procedures Hematological Coagulopathy (congenital, acquired or iatrogenic) Platelet disorders Drugs and toxins Penicillamine Solvents Crack cocaine Trimellitic anhydride Bevacizumab Miscellaneous Endometriosis Lymphangioleiomatosis Broncholithiasis Cryptogenic Foreign body aspiration Lung transplantation

You call that "massive"?

Critical care literature is rarely given to pointless hyperbolae, and the use of words like "massive" or "catastrophic" is usually reserved for colloquial descriptions rather than precise clinical definitions. This also seems to be the case here. How much is "massive", and can you really ever quantify the volume of expectorated blood? It tends to go everywhere, it's not as if the patient is carefully depositing it into a measuring jug.

Historically, valiant efforts to hold that jug have been made. A great many article incorrectly reference Amirana et al (1968) as the origins of a low-volume definition for massive haemoptysis. These investigators reported on a series of TB patients from the 1960s, but did not get over-excited with his terminology - they described it as "significant" haemoptysis when their patient coughed up a total of 100ml per day. This volume was correctly described, as it hardly seems worthy of the term "massive".  At the other end of the spectrum,  a highly cited article by Corey et al (1987) instead suggests a volume of 1000ml over 24 hours, which seems better suited to descriptors like  "colossal" or "spectacular".

Of course, all these definitions are totally pointless. The LITFL article on haemoptysis wisely points out that definitions based on volume are meaningless in terms of defining "massiveness", as the more important issue is how life-threatening it is (by virtue of airway obstruction). In this case haemoptysis coming from the tracheostomy patient in Question 7 from the first paper of 2017 could be life-threatening with only 4ml of blood (that is the internal volume of a 8.0mm non-adjustable percutaneous tracheostomy tube). Rate of blood loss must also be viewed as important, as the loss of 1000ml over 24 hours may be better tolerated than the the loss of 1000ml over fifteen seconds. These considerations had prompted a certain W.H. Ibrahim to write a letter to the editor of European respiratory Journal, calling for an introduction of the term "life-threatening haemoptysis", as this might  "provide a fascinating and rich understanding of the condition".

Localising the site of bleeding

Question 16 from the second paper of 2022 asked the candidates tto "list the investigations that will assist with localizing the site of bleeding". 

• A chest Xray would usually be the first investigation performed for any such patient, not necessarily because it is valuable for the task of locating the bleeding site but because it tends to happen reflexively for intubated patients. That may be its only advantage - ubiquitous availability and the real likelihood that it will be performed anyway, and probably sooner than any of the other investigations. At worst it may reveal non-diagnostic  lung horror, and confirm the ETT position. In the best case scenario, it may reveal one lung to be the source (the other being perfectly clean), which may guide the choice of the next management step (eg. the insertion of a double-lumen tube). The main disadvantage is the lack of accuracy- Revel et al (2002) reported that Xrays identified the bleeding site in only 46% of their cases.
• Non-contrast CT could still be helpful because lung pathology leading to haemoptysis is often obvious and easily visible, for example a large malignant mass, cavitating abscess or a fungus ball. Moreover the lobe with the greatest amount of unexpected ground glass change or consolidation can be implicated as the bleeding site (apparently this strategy achieves "almost perfect" concordance with bronchoscopy findings, according to Seon et al, 2016)   If for whatever reason one is not inclined to give IV contrast,  a non-contrast CT is still possibly helpful. The disadvantages of going for this non-ideal investigation is the actual transport, which carries quite a significant risk for this patient. You are taking an acutely ill patient with a real potential for impending total airway obstruction for an investigation the findings of which might not be useful. This may represent a waste of precious time, and produce a deterioration. 
• CT angiography with contrast is the modern standard, now that the resolution of multidetector CT  scanners is well-matched with digital subtraction angiography. Unlike bronchoscopy, CT can also reveal extrapulmonary causes of bleeding, such as aneurysms of various sorts, and it can reliably detect a pulmonary embolism which bronchoscopy cannot. Accuracy is excellent, and is getting better with every technological iteration - Revel et al (2002) reported that CT angiography identified the bleeding site in 77% of cases, and this was using ancient bronze-age CT scanners from the late 1990s.  Disadvantages include contrast exposure, the perils of transport, and the nagging knowledge that the CT, while being able to point out the site of the bleeding, does not offer a means to fix it. Additionally, the volume of bleeding would have to be reasonably large in order for this to be detected by CT. Also, endobronchial masses (i.e. tumours growing off the inside of the bronchus) may not be detectable if the whole bronchus is filled with blood, and will therefore go unnoticed.
• Digital subtraction angiography was at one stage the gold standard, but has receded from this position, mainly because its main diagnostic advantage over other modalities was accuracy, and this has been lost with improvements in CT technology. This change in position occurred as early as 2004 (Remy-Jardin et al, 2004, pointed out that CT was better, albeit in a small group of patients). DSA can pinpoint small arteries as the source of bleeding, but because bronchial arteries are numerous, DSA may miss the culprit lesion because only some of the arteries are imaged during each screening. The harder the anatomy and the more elusive the lesion, the higher the dose of radiation and contrast (as opposed to the CT, where all the anatomy is imaged precisely and simultaneously). In short, the role of DSA has transitioned gradually towards being a therapeutic modality that compliments CTA. One advantage that DSA still has is the ability to define the feeding vessels of a tumour or AVM, thus helping target embolisation.
• Fiberoptic bronchoscopy now occupies a similar position to DSA, in the sense that its diagnostic accuracy is below that of CT, but it at least offers some therapeutic benefits. One may be able to delineate the source of the bleeding by visualising it directly (if it is in the proximal bronchial tree) or at least localise it down to the lung segment (if it is distal). It may even be possible to see the lesion (eg. an endobcronchial mass). The disadvantage is, it may not be possible to see anything, as blood obscures the field of view. Its possible to suction blood, clearing the airway, and some of the clots could be grabbed with the tip or using an instrument port, which means you can pull them out and unobstruct the bronchus. It is also possible to damage the mucosa and worsen the bleeding, particularly where coagulopathy is playing a role in the haemoptysis. Bronchoscopy definitely has a role to play (even if only in helping place the DLT) but its specific role in the diagnosis and localisation of the bleeding site is diminishing.

Management of massive haemoptysis

Unlike the dislodged tracheostomy or the patient suddenly impossible to ventilate, this thing does not have an agreed-upon algorithm of management. The brief summary of sensible-sounding management steps seen in the discussion section for  Question 2 from the first paper of 2012 is offered here in lieu of anything more official:

1) Control the airway.

• Intubate the patient with a large-bore tube to permit bronchoscopy
• If you are skilled and the pathology is unilateral, a dual-lumen tube could be considered
• Position the patient in a Trendelenberg position, or with the bleeding lung dependent.

2) Control the breathing.

• Ventilate the patient with the bad lung dependent, to prevent contralateral lung soiling
• Increase the PEEP, to get the benefit of whatever tamponade effect it might provide.

3) Control the circulation.

• Replace the lost blood and stabilise the hemodynamic variables

4) Control the bleeding

• Reverse any coagulopathy
• Perform bronchoscopy
  • Suck out any obvious clots
  • Place a balloon-tipped catheter to put pressure on the bleeder
  • Burn the bleeder with argon plasma (if you have the tools)
• Perform angio-embolisation if bleeding is not controlled. Angio-embolisation is a pretty cool modality, with a low complication rate.
• Send the patient to thoracotomy if angio-embolisation is impossible

5) Control the cause

• Antibiotics for tuberculosis and fungal abscesses
• Surgery or radiotherapy for cancers
• Immunosuppression for vasculitis
• Surgery for AVMs

Now, let's say, as in Question 16 from the second paper of 2022, that a single source of bleeding has been identified. What are your management options? A whole series of possibilities presents itself. Davidson & Shojaee (2020) have an excellent paper that details some of these options. This also permits an exploration of an alternative approach to the answer, worded and structured slightly differently. Below, the options are presented in order of both priority and invasiveness

1) Achieve lung isolation

• Intubate the patient with a large-bore endotracheal tube to permit bronchoscopy
  • And position the patient in a lateral position with the bleeding lung dependent
• Or: introduce a bronchial blocker and inflate the balloon in the main bronchus of the affected lung
• Or: intubate the patient with a double-lumen endotracheal tube
  • For a bleed originating in the left main bronchus, a right-sided DLT is called for, as a left-sided DLF would not permit surgical or bronchoscopic access
  • For any other site of bleeding a left-sided DLT would be appropriate and much easier to place
  • Some authors argue that these devices are not useful because the small aperture of each lumen does not allow the passage of a "proper" bronchoscope, of the sort that has ample instrument or suction ports.

2) Trial conservative management: some patients may improve (apparently, up to 17%, according to Valipour et al, 2005)

• Nebulised tranexamic acid - a trial by Wand et al (2018) found that 500mg three times a day was enough to significantly reduce the expectorated blood volume and produce a much better rate of resolution by day 5 (96% vs 50%). These were mainly stable awake patients with a history of malignancy or bronchiectasis; it is unknown whether this strategy would consistently work for proper haemoptysis, or whether it would be equally effective to be giving tranexamic acid via simpler and more convenient IV administration.
• Nebulised adrenaline has been mentioned in a number of authoritative sources, but has not been the subject of any big prospective studies. Recommended doses seem to range from 8mg to 0.5mg, which would be a surprisingly large range for a well-accepted therapy, suggesting that this one is not. Fortunately it is reasonably safe, in the sense that the drug has a very short half-life, and the haemodynamically unstable patient may even benefit from absorbing it systemically.
• Ice saline lavage, delivered from the end of a flexible bronchoscope, is listed among the "conservative management" techniques because it is "non-interventional", which is another way of saying that it's ineffective. It is occasionally presented as a measure of last resort in resource-poor settings. The literature ranges in its recommendations, and most people seem to use fridge-cold saline (4ºC) in 50-60ml flushes, each sucked out after 30-60 seconds. Apparently on average about 500ml is required, though some champions have used up to 2.5L. The reader is left to think on what effect this might have on dislodging delicate haemostatic clots and worsening the bleeding (or soiling the contralateral lung, or washing out surfactant, etc).

3) Interventional bronchoscopy techniques involve usually the obliteration of the bleeding site or the occlusion of the mucosa. Options include:

• Adrenaline injections
• Various gels, foams, sealants (including thrombin slurry)
• Temporary silicone plugs  
• Temporary bronchial stents
• Nd:YAG laser
• N-butyl cyanoacrylate glue

4) Rigid bronchoscopy should probably be mentioned, mainly because the examiners had complained about it in their comments to Question 16 from the second paper of 2022. In case the reader is wondering what the difference is, a rigid bronchoscope is a hollow metal tube which is introduced into the trachea instead of an endotracheal tube, rather than through it, and which does not have any of the advantages of the latter (i.e. it is not cuffed, does not ventilate beyond simple gas insufflation, and cannot offer controlled positive pressure). Of course the counterargument here is that a normal bronchoscopic swivel adaptor leaks horribly anyway, so what pressure and ventilation are you really getting? Anyway: access for instruments is the main advantage, and apart from ventilation the main disadvantage is the inability to reach upper lobe bronchi and more distal structures. Technques made available mainly via the rigid bronchoscope include:

• Argon plasma coagulation
• Electrocautery
• The placement of larger stents
• The retrieval of larger clots

Of these, size is the most important thing, as electrocautery and plasma are only limited by rapidly evolving flexible bronchoscope technology

5) Interventional radiology is mainly involved for bronchial artery embolisation. Coils, foam, glue, and tiny PVC particles have all been used. There is probably little else to discuss here other than to say that this technique is powerless to limit the damage from pulmonary arterial or venous bleeding.

6) Surgery is the last option, as it would usually be performed in an emergency and would usually result in severe morbidity. Lobectomy and pneumonectomy are often required, and the mortality rate for emergency procedures is quoted as 34%. This is the only option for pulmonary arterial or venous bleeding, and probably the preferred option for malignancies and large abscesses.