Ventilation strategies for COPD

This chapter deals with a common cause of ICU admission, the wheezy COPD patient with some mixture of respiratory acidosis, hypoxic respiratory failure and biventricular dysfunction. After a rather long break from asking specifically about COPD, CICM examiners came back with a rather intelligent question in 2018, featuring one of the best-constructed model answers.

In total, the SAQs asking about COPD have been: 

• Question 11 from the second paper of 2020
• Question 8 from the second paper in 2018
• Question 29 from the second paper of 2005
• Question 30 from the second paper of 2005
• Questions 1a, 1b, 1c  and 1d from the first paper of 2001 

In summary:

Investigations

• ABG
• EUC / CMP / FBC / CRP /BNP
• Chest Xray - looking for pneumonia
• Transthoracic echo - to assess the contribution from heart failure
• Blood/sputum cultures
• Urinary pneumococcal and legionella antigens
• Atypical pneumonia serology

Non-invasive management

• Oxygen therapy, aiming at a SpO₂ around 90%
• Anticholinergic bronchodilators in combination with beta-agonists, eg. ipratropium bromide plus salbutamol - at first 2-hourly, and gradually de-escalating
• Think about methylxanthines
• Think about antibiotics (3rd generation cephalosporin and a macrolide)
• Steroids IV (100mg of hydrocortisone q6h)
• Attack them with a physiotherapist and an incentive spirometer.
• Manage their heart failure. Consider digoxin in cor pulmonale, and be careful with diuretics (as right heart failure may worsen in the absence of adequate preload).
• Attention to electrolytes, particularly phosphate
• Nutrition, preferably low-carb and high-fat
• NIV to reduce respiratory muscle workload, correct respiratory acidosis, improve oxygenation.

Invasive management

• Consider intubation if their functional capacity suggests the possibility of a return to independence
• Intubate them if they are comatose, not benefitting from NIV, or unable to clear secretions
• Use a minimum level of support, and aim for longer expiratory time
• Match intrinsic PEEP with extrinsic PEEP
• Extubate them onto NIV as soon as is practical.

Assessment of severity of COPD

Question 29 from the second paper of 2005 asks about the assessment of the severity of COPD.

Historical features:

• exercise tolerance
• breathlessness with everyday activities
• presence of chronic cough
• high volume of sputum, suggestive of bronchiectasis
• haemoptysis, suggestive of malignancy
• home O2 requirement
• home CPAP requirement
• pattern of bronchodilator use
• pattern of steroid use
• frequency of hospitalisations
• previous mechanical ventilation
• anorexia and weight loss

Examination:

• features of malnutrition
• features of obesity (sedentary lifestyle)
• features of chronic steroid use
• central cyanosis
• breathlessness at rest
• hyperexpanded chest
• degree of air entry
• signs of right heart failure

Investigations:

• Bicarbonate levels
• Hb (polycythaemia)
• Spirometry, pre and post bronchodilator
• Formal lung function tests
• ABGs to determine degree of hypoxia and hypercapnea
• TTE (pulmonary pressures)
• High-resolution CT to assess the severity of emphysematous changes

The BODE index

• Predicts mortality among COPD outpatients
• According to the original article, "is better than the FEV1 at predicting the risk of death from any cause and from respiratory causes among patients with COPD"
• Consists of 4 categories, each worth a certain number of points, up to a maximum score of 10 moints (being the worst chances).

• The MMRC breathlessness scale is basically a subjective report of how breathless the patient feels; 0 is "doin fine" and 4 is "can't leave the house"

Mortality accordidng to the BODE index

This graph is the slightly modified data from the original article;  it demonstrates that among the most severe group, there is a 50% mortality at 36 months.

Acute respiratory failure in COPD

This comes in two flavours, distinguished by the extent of respiratory acidosis. You may either have hypercapneic respiratory failure, or your PaCO² may be normal with hypoxia.

Hypercapnic respiratory failure in COPD

These are patients with chronic bronchitis. They are typically obese, and they may be obtunded not only by the hypercapnoea but also by the presence of CNS depressant medications. They will almost inevitably have some sort of obesity-related sleep hypoventilation. The right heart failure in these people may not play such a major role, but there will likely be increased pulmonary artery pressure.

Normocapnic respiratory failure in COPD

These are the thin emphysema patients, with hyperxpanded chests and vigorously active accessory muscles. They are typically chronically hypoxic, and may be on home oxygen. Their right heart is failing, and pulmonary hypertension is almost inevitably present.

Infective exacerbation of COPD

About a half of these patients have an "infective exacerbation" of COPD. Streptococcus pneumoniae and Haemophilus influenzae account for 80% of the pathogens. The rest are atypicals and viruses, such as Moraxella, Mycoplasma pneumoniae, Pseudomonas, RSV, adenovirus, influenza and parainfluenza. There may or may not be an actual pneumonia.

In either case, one can see from the above that most of the pathogens are either Gram-positive (covered by beta-lactams and cephalosporins) or atypicals (covered by macrolides). This promotes the use of the traditional cocktail of ceftriaxone and azithromycin/roxithromycin in these patients.

Does it matter if there is also left heart failure?

It is a well-known and lamented feature of emergency resident life that COPD patients get passed back and forth between the respiratory and cardiology teams. It is frequently difficult to distinguish where the CCF ends and the COPD begins. The lifestyle risk factors which predispose them to COPD also create a fertile soil for coronary artery disease, and some degree of ischaemic cardiomyopathy is almost mandatory.

In addition, BNP levels can be useful to differentiate between the cardiac and pulmonary causes of heart failure, assisting the beleaguered ED resident. BNP issues forth in volumes from the insulted over-stretched atria of the heart failure patient, whereas in the exacerbation of COPD these levels may be quite low. (Apparently this distinction is useful and holds true provided your patient is under 70 yrs of age and has normal kidneys).

Thankfully, in the ICU the question of "which team to admit them under" is pleasantly remote. One accepts that the heart failure is part of the spectrum, and both the pulmonary and cardiovascular components of the respiratory failure will respond to the same treatments. The left ventricle may be diseased, and there may be features of congestive heart failure which will respond to positive pressure. Relief of bronchospasm and the resulting decrease in intrinsic PEEP will improve the diastolic failure by improving left ventricular preload. The increased cardiac output demands to power respiratory pump muscles will also be remedied by interventions which decrease the work of breathing.

Do I really need to restrict oxygen for these people? What is the evidence?

They do say that the chronic exposure to hypercapnia causes the hypercapneic respiratory drive to be suppressed, and that hypoxia is the primary driver of ventilation in severe COPD.

There are several other reasons as to why one may run into trouble with over-oxygenating their COPD patients. For instance, normoxia reduces their anxiety, and so their respiratory rate will decrease - which may be counterproductive. Similarly counterproductive is the effect of normoxia on the CO₂-carrying capacity of hemoglobin (i.e. by the Haldane effect, the deoxygenated hemoglobin molecules have higher affinity for CO₂).

In one Thoracic Society position statement, 100% O₂ was administered for 15 minutes; an average 23mmHg rise in CO₂ was recorded. Of this rise, only a 5mmHg rise was attributed to the Haldane effect. A major proportion was thought to be due to increased dead-space. How does this work?

Consider the emphysematous lung. Some emphysematous lung units a represented by eroded alveoli with little gas exchange surface, supplied by chronically obstructed bronchi. The pulmonary arteries running through these lung units are vasoconstricted because typically, the oxygen tension in these units is quite low. Consider what would happen if you wafted some oxygen though the obstructed bronchi and into these eroded alveoli. The ventilation has not become more vigorous- if anything, the reduced work of breathing associated with oxygen therapy has decreased the ventilation of these lung units. But the increased oxygenation has increased blood flow by decreasing the pulmonary vasoconstriction. The effect is that of increasing blood flow into lung units which have poor ventilation and limited gas exchange surface. That blood was being shunted way from effective lung units, and this decreased the overall rate of CO₂ removal.

In fact, this is exactly what happens in asthma, and for asthmatics the recommendation is to maintain a modest level of oxygenation, which permits good V/Q matching.

None of this means that oxygen should be withdrawn altogether if severe hypoxia persists. If the oxygen saturation remains below 88% one ought to up-titrate their oxygen delivery. Hypoxia is more rapidly fatal than hypercapnea.

The relevant features of their investigations

Yes, the hypercapnea on the blood gas will be the dominant feature. Each 10mmHg chronic rise of CO₂ from 40mmHg increases the HCO₃^- by 4mmol/L. In a patient with a random CO₂ of 70, one ought to expect a chronic HCO₃^- level of around 36.

Apart from the obvious blood gas features, one who is asked by an examiner to comment on a COPD X-ray would be wise to mention something about the features of hyperinflation as well as the features of increased pulmonary arterial pressure.

One may also talk about the paucity of lung markings, about atrial enlargement, and about the presence of bullae (if there are any bullae to be seen).

The relevant features of their formal lung function tests

The formal pulmonary function tests also lend us some information about the severity of COPD.

The TLCO (total lung carbon monoxide) uptake is a reflection of how much diffusing surface you have left; the fewer alveoli remain the lower the TLCO. In emphysema, this surface is destroyed, and the TLCO is usually ~ 80% below predicted.

Additionally, the plethysmographic lung volume measurements will demonstrate that the total lung capacity, FRC and residual volume are all increased ( the residual volume would be over 40% of total lung capacity!). This is a reflection of gas trapping ( the residual volume is all trapped gas).

This is somewhat helpful in differentiating the disorders. In asthma, the CO diffusing capacity will remain normal - alveolar tissue is not destroyed.

Everybody seems to get bronchodilators, but is the obstruction not chronic and irreversible?

That's right, every COPD patient seems to get massive asthma-like doses of salbutamol in hospital. There is indeed some small element of reversible airflow limitation here. However, a meta-analysis has demonstrated that over the long term, the greatest mortality reduction is due to anticholinergic drugs like tiotropium, and that beta-agonists are essentially no better than placebo.

The role of methylxanthines

It is generally thought (and sometimes even demonstrated in studies) that drugs like aminophylline and theophylline improve outcomes in severe stable COPD. The therapeutic window is narrow; blood levels of the low effective range are usually targeted (55-85 mmol/L). However, the higher effective range (85-110mmol/L) is thought to be required for the diaphragm stimulation and respiratory drive increase which you want from these drugs. At this range, the balance of useful effects to side-effects becomes dangerously even.

The importance of sputum clearance

If you cannot clear sputum, you cannot survive the exacerbation. Physiotherapy and those bubble-themed incentive spirometers tend to improve sputum clearance and encourage coughing. If all else fails, one may wish to suction these people's tracheas though a nasopharyngeal airway, or attempt actual bronchoscopy to remove large sputum plugs.

The COPD diet

Theoretically, a low carb high fat diet should decrease the whole-body CO2 production (Ohs Manual reports that the decrease is by 15%). There is some evidence to support this, but strong recommendations are nowhere to be found. Overall, one can make a point for ensuring adequate nutrition in general, given how malnourished these people can get.

The goal of non-invasive ventilation

Lets face it, your first impulse is to blow off some CO2 with NIV, or to give some positive pressure and improve oxygenation. Additionally, the PEEP should decrease the work of breathing due to dynamic hyperinflation and bronchospasm.

There are several studies of NIV in COPD, which have demonstrated a survival benefit. Their entry criteria for NIV are as follows:

• Resp rate > 28
• Respiratory acidosis: PaCO₂ > 45mmHg and pH < 7.35

In short, NIV is your most useful tool in normalising respiratory function in this group.

When NOT to intubate the COPD patient

This might seem like a completely valid practical question, which makes it all the more surprising to find it among the repertoire of CICM past paper SAQs. Question 8 from the second paper in 2018 presented the candidates with a severe COPD scenario, and asked: "Discuss in detail how you would make a decision about whether to offer invasive mechanical ventilation to this patient, should he fail the trial of NIV."  The majority of these people don't have end-stage disease, and the mere fact that you have some COPD should not be a barrier to having a trial of mechanical ventilation to treat some sort of reversible pulmonary pathology. The patient in the CICM scenario has disease qualitatively described as "severe", but is not on home oxygen (which is generally viewed as a marker of the "end stage" of COPD), and is therefore in that horrible limbo where it is not completely clear whether to offer or withhold intubation. 

In short, when deciding whether or not to offer intubation to a severe COPD patient, there are a few main questions to consider:

• Is there sufficient information to make a decision regarding whether or not to offer mechanical ventilation?
• Does invasive ventilation meet any of the definitions of futility?
• Will invasive ventilation improve medium-term mortality?
• Will invasive ventilation improve medium-term quality of life?
• Will invasive ventilation be a short-term process for this patient?
• Is invasive mechanical ventilation something the patient will agree with?

These will now be discussed in some detail.

Is there sufficient information to make a decision regarding whether or not to offer mechanical ventilation? If the answer is no, then it is generally better to err on the side of intubation. One buys time in this fashion. It is a broadly acknowledged fact that any plastic placed by an intensivist can be equally easily un-placed, i.e. if it emerges over the subsequent hours and days that the patient's background functional status and life expectancy is much worse than previously thought, a palliative one-way-extubation can be performed after some discussions with the family. Those discussions will also be much more civilised, as they will be performed in a calmer atmosphere - the decisions will not be urgent.  In both  Question 1b from the first paper of 2001 and Question 8 from the second paper in 2018 the college examiners are in favour of initial aggressive therapy (in the latter case, the actual phrase "buy time" is used in the model answer).

There is some evidence that jumping to conclusions does harm. Wilman et al (2007) found there was a significant amount of "prognostic pessimism" among ICU physicians who were asked to predict patient outcomes in this setting. "For the fifth of patients with the poorest prognosis according to the clinician, the predicted survival rate was 10% and the actual rate was 40%". The investigators concluded that some patients who might otherwise survive are probably being denied intubation. This is not helped by the fact that there are no strict criteria for this decisionmaking, and ICU specialists differ significantly in their opinion regarding the same patients (Wildman et al, again, this time 2003).

Does invasive ventilation meet any of the definitions of futility?  Under certain circumstances, intubation may meet the broadly accepted definition of "operational" futility, being a therapy which is “so unlikely to succeed that many people... would consider it not worth the cost”. It remains to be debated what precisely "to succeed"  means when it comes to intubating these people, but a sufficiently large number of people (both lay and professional) would probably agree that death is not "success". As such, patients should not be offered mechanical ventilation in any scenario where it will inevitably lead to death. That brings us to the next question.

Will invasive ventilation improve medium-term mortality? In other words, how likely is this patient to die if mechanical ventilation is offered, and what will that look like? This seems to depend on the pathology which drives the decision to intubate. Some scenarios and pathologies are associated with better outcomes. The following things can be confidently said:

• If you need intubation because you have "failed" NIV, your outcome is worse. Chandra et al (2012) found that the odds of death for such patients are almost seven times higher than for patients who "succeed" NIV, and about 60% higher than for patients who end up getting intubated without an NIV trial. The in-hospital mortality for these people was 33%; non-survivors were largely elderly (over 55% of them were aged 75 or older). Something similar was found by Stefan et al (2015). 
• If you need intubation for a "pure" COPD exacerbation, your outcome is better. Nevins et al (2001) found that the in-hospital mortality was only 12% for these people, as compared to 28% for the rest of the cohort (admittedly, a small cohort of 166 patients, but then only about 5% of acute COPD exacerbations lead to intubation). These were relatively robust patients (average FEV₁ was 1.24 for the entire group).
• If after 72 hours you still require mechanical ventilation, your outcome is worse. Again quoting Nevins et al (2001), the patients who were still invasively ventilated after 72 hours had an in-hospital mortality of 37%, whereas the rest (presumably, extubated by then) had a mortality of 16%. If you develop ventilator-associated pneumonia, the mortality increases to 57% (Rinaudo et al, 2018)
• The presence of comorbidities makes survival less likely, and among the comorbidities the most important one is malignancy, which basically doubles the mortality risk (Stefan et al, 2015). Other comorbidities associated with non-survival were cor pulmonale, chronic hypercapnia, and left ventricular failure (Menzies et al, 1989).
• Poor FEV1 and FEV1/FVC predicts mortality. This is the GOLD score (Mannino et al, 2006) and even though it uses data from twenty years ago, it is still mentioned by CICM examiners as a mortality prediction model. GOLD stage 3 or 4 (FEV₁/FVC<0.70 and FEV₁<50% predicted) is associated with a mortality of around 35% at ten years follow-up, but this is not specific to ICU patients (it was a community follow-up study). 
• Poor function in a broader sense predicts mortality. The BODE index (Celli et al, 2004) takes BMI,  MRC dyspnea score, six-minute walk distance and FEV₁ % predicted. It appears to be better at predicting mortality than FEV₁ alone.  If the patient is emaciated, breathless at rest and unable to walk more than 150m over six minutes, with an FEV₁ below 35% of the predicted value, BODE gives them a 20% chance of surviving the next two years. The 2018 NICE guidelines also recommend this index. In terms of ICU performance, Menzies et al (1989) found that the premorbid level of activity was the most important parameter. If the patient is unable to leave their house because of their symptoms, their mortality was 71%, going up to 75% if they were chronically bedbound or chairbound.
• Dependence on home oxygen seems to be a strong predictor of poor outcome. A recent study from Thorax (Hajizadeh et al, 2015) retrospectively observed a cohort of 4791 end-stage oxygen-dependent COPD patients who were intubated, and found that 23% died in the hospital, and 45% died in the subsequent 12 months, with 26.8% discharged to a nursing home within 30 days.

Will invasive ventilation improve medium-term quality of life? This is the question most effectively answered by the severity rating scales. 

Will invasive ventilation be a short-term process for this patient? And if not, what will be the likely outcome? This is a question often raised in discussion with respiratory physicians, who might recommend a "short-term trial" of invasive ventilation. Gadre et al (2018) found that the median duration of invasive ventilation in their cohort of 670 intubated COPD patients was 3 days, which certainly sounds like a short trial. The patients who were extubated early were mainly the "pure" COPD exacerbations, whereas the pneumonia were ventilated for longer. In general these authors found that the ICU mortality for their "pure" COPD cohort was actually lower than for other acute respiratory pathologies (9% ICU mortality, 17% hospital mortality). The authors concluded that the anxiety about intubating these people appears to be mainly driven by the results of older studies, from a bygone era of barbaric ventilator management. 

But let's say they end up stuck on the ventilator, as you feared. What's the outcome? Quinnell et al (2006) reported on a cohort of 67 such patients, with median FEV₁ of 0.6. Of these,  95% were weaned off the ventilator with a tracheostomy, and almost 90% survived to discharge. Their median post-discharge survival was 2.5 years, with one-year, 2-year, and 5-year survival rates were 68%, 54%, and 25%, respectively. ICU length of stay was relatively prolonged (around 30 days on average), and it was followed by another roughly 30-day stay in the "Respiratory Support and Sleep Centre", what sounds like some sort of ventilated nursing home designed for the purpose of weaning these people. In short, you're in for a long stay. 

Is invasive ventilation something the patient would agree with? Often this becomes obscured by the fog of hypercapnic narcosis, but the patient's decisionmaking surrogates may be able to enlighten you on what the level of insight has been. It will have been influenced by what their respiratory physician has been telling them. Sullivan et al (1996) found there was some considerable amount of "spin" put on this, positive or negative (depending on what the "respirologist" expected for the patient).

When asked, months after their ICU experience, COPD patients who survived intubation seem relatively comfortable. In spite of the burden of their symptoms, Wildman et al (2009) found the majority rated their health as same or better, and 96% responded that they would be willing to undergo similar treatment again. Less is known about the slow-to-wean, two-month-in-hospital cohort. We can only assume that they are like all the other long-term ventilation survivors, who (according to Huttmann et al, 2018) who were all functioning poorly and feeling horribly depressed. Of these, 32% reported that they would have elected to die in hindsight rather than receive invasive ventilation for a prolonged period.

Which COPD patients should be intubated

There are several features which promote intubation as a sensible step.

• Failure of NIV: The patient is already on NIV, but it has failed to reduce the work of breathing, and the patient is beginning to become fatigued which makes the ABG results look terrible.
• Coma: Level of consciousness does not permit NIV, or the airway reflexes are not preserved
• Pneumonia: There are copious secretions, and an ineffective cough (which is made worse by NIV- its very hard to clear secretions by coughing against a 10cm positive pressure)
• Their hypoxia is worsening, and NIV is not contributing usefully

How to ventilate the COPD patient

The objectives are to improve secretion clearance and to rest the respiratory muscles while removing as much CO₂ as practical. Therefore, the less support you use the better the outcome.

There are a couple of tricks which may be useful in this scenario:

Expiratory pause to measure dynamic hyperinflation:

• The expiratory pause manoeuvre measures intrinsic PEEP. As this rises to 8-10 cm, one might consider changing the I:E ratio to increase the duration of expiration

Match intrinsic PEEP with extrinsic PEEP

• Having established which the intrinsic PEEP is, one should set the machine to deliver a similar level of pressure, thereby neutralising the effect of intrinsic PEEP on the work of breathing.

Steep inspiratory rise time

• If you have little control over the I:E ratio (eg. in a patient-triggered mode) you may consider using a higher inspiratory flow rate (which on some ventilators is represented by a "rise time" inspiratory flow ramp). The greater the inspiratory flow, the shorter the inspiratory time; thus more time is allowed for expiration for any given respiratory rate. A good quick flow rate is about 100L/min.

Extubate early, and on to NIV

• As soon as practical, these people should be extubated and supported with NIV. This Cochrane review suggests this strategy can halve mortality and reduce the rate of VAP to one third. .

But, even with the best intentions, some of these people go on to have a prolonged period of ventilation, which culminates in a tracheostomy.

When and why to decannulate the tracheostomy

The tracheostomy experience is a terrible thing for secretion clearance; without an epiglottis, one cannot cough properly, and the whole process becomes counterproductive.

• As soon as you confirm the upper airway reflexes are intact
• When the secretions are under control (i.e. you suction the tracheostomy only every 2-4 hours)
• When the respiratory muscle strength has recovered sufficiently to make good cough efforts
• Obviously, when the patient is free from their dependence on the ventilator for PEEP and oxygen(i.e. they are on humidified blow-over room air)