Recently the concept of extracorporeal CO2 removal has become somewhat popular, with a goal of facilitating ultra low-tidal volume ventilation. Might there be other means to achieve the same goal? Let’s start with some basic concepts of mechanical ventilation…
Fundamental unanswered questions
Permissive hypercapnia refers to the concept of allowing the pCO2 level to increase to supernormal levels intentionally, in order to facilitate lung-protective ventilation. This is a fundamental, bedrock principle involved in safe ventilation of patients with ARDS or asthma. However, many fundamental questions regarding permissive hypercapnia remain unanswered. For example:
What is a safe pH range?
Respiratory acidosis is generally tolerated extremely well. Consequently, it’s unclear what constitutes an optimal pH. There is no well-defined pH cutoff below which patients obviously deteriorate. Case reports abound of patients with extreme respiratory acidosis who survived.
Certainly, the target pH should be adjusted on a patient-by-patient basis. This involves balancing the potential harm from more aggressive mechanical ventilation versus the potential harm from acidemia. For example, a young patient with profound asthma and stable hemodynamics might benefit from a lower pH target (prioritizing avoidance of barotrauma over hemodynamic optimization).
Nonetheless, some loose guardrails may be helpful. Specifically, what is a reasonable target pH for most patients?
permissive hypercapnia for the intubated patient with ARDS or asthma: what pH are you generally comfortable with?
— josh farkas 💊 (@PulmCrit) April 6, 2019
The twitter poll above illustrates uncertainty about this. The most common choice was a pH target of >7.20. This is consistent with recommendations in UpToDate and some modern study protocols.1 It also seems to be a largely arbitrary choice. To my knowledge there is no high-quality data regarding this.
What is a desirable pCO2 level?
This is also unknown. There are a variety of potential benefits from hypercapnia (e.g. anti-inflammatory effects, increased cardiac output). However, there are also potential harms from hypercapnia (e.g. right ventricular strain, renal vasoconstriction).
It’s tricky sorting out the clinical consequences of hypercapnia from those of respiratory acidemia (as the two conditions generally go hand in hand). In the absence of acidemia (e.g. if the acidemia is neutralized using alkali therapy), both the beneficial and the detrimental effects of hypercapnia may be diminished.
Elevated pCO2 levels are generally well tolerated. For example, it’s not uncommon to encounter a patient with severe COPD who happily lives at a pCO2 level of ~80 mm for years. Extremely high pCO2 levels may cause a sedative effect (CO2 narcosis), but this isn’t necessarily a bad thing for a patient who is intubated (indeed, it could help reduce the sedative requirement). Elevated pCO2 may increase intracranial pressure and pulmonary vascular resistance, so permissive hypercapnia is relatively contraindicated in patients with severe intracranial pathology or unstable pulmonary hypertension.
To my knowledge there is no reliable data regarding what the optimal pCO2 level is (or if one exists at all). To illustrate this equipoise, the TAME RCT is currently underway to evaluate whether mild hypercaponia (PaCO2 50-55mm) may be beneficial following cardiac arrest.
What is the role of IV bicarbonate to achieve pH targets?
Intravenous bicarbonate may be used as an adjunctive agent with permissive hypercapnia, in order to allow the pCO2 to rise without causing an excessive drop in pH. In short, when approaching a patient with refractory hypercapnia and respiratory acidosis, the easiest and safest way to improve the pH may be to slowly infuse 1-2 ampules (50-100 mEq) of sodium bicarbonate (as opposed to cranking up the tidal volume and respiratory rate, which would increase the risk of barotrauma).
The classic ARDSnet trial of low tidal-volume ventilation included the use of intravenous bicarbonate if needed to maintain target pH (see above protocol). The discussion section of this trial suggests that the use of IV bicarbonate was one potential reason that their trial was more successful than previous trials of lung-protective ventilation.2
Hypercapneic patients with ARDS commonly develop a metabolic alkalosis (e.g. bicarbonate level of 30-35 mM), even without any exogenous bicarbonate administration. This occurs for two reasons. First, these patients are often diuresed, leading to a contraction alkalosis. Second, when exposed to hypercapnia, the patient’s kidneys will naturally retain bicarbonate in efforts to increase the pH. This process of adaptation to hypercapnia usually occurs over several days.3 The administration of exogenous bicarbonate may merely serve to accelerate this physiologic adaptation.
One additional benefit of bicarbonate is that it will reduce the respiratory drive. This may make patients more comfortable on the ventilator, and subsequently reduce their work of breathing once extubated (thereby theoretically reducing the risk of re-intubation).
The graph above shows the minimum bicarbonate level which is needed for any specific pCO2 level, in order to maintain a pH above 7.20 (based on the Henderson-Hasselbach Equation). This illustrates that if we push the bicarbonate level to the 30-35 mM range, we can generally allow the pCO2 to rise quite high (80-90 mm). Thus, we can support patients with severe hypercapnia by using only a modest metabolic alkalosis (serum bicarbonate is normally 22-28 mEq/L, so 35 mEq/L is only mildly elevated in an absolute sense).
There is often concern that infusion of bicarbonate will increase the pCO2. Bicarbonate does indeed cause a transient elevation in pCO2, but this will be blown off over time (with a decrease back to the previous equilibrium pCO2 levels). Essentially, the additional pCO2 load increases the gradient driving pCO2 out of the body, thereby increasing pCO2 elimination until the pCO2 decreases back to baseline. Bicarbonate boluses shouldn’t be used, to avoid sharp increases in CO2 levels.
Hypoventilation managed without extracorporeal CO2 removal
Before the advent of extracorporeal CO2 removal, practitioners were occasionally forced to allow the CO2 level to climb extremely high (in patients who truly couldn’t be ventilated effectively). Although extreme elevation of pCO2 or markedly low pH might be scary, it doesn’t seem to kill patients. For example…
The patient with the pCO2 of 373 mm!
Garg et al. reported a case of a patient with hemoptysis who was nearly impossible to ventilate.4 His pH dropped to 6.88 with a pCO2 of 230mm. Extracorporeal CO2 wasn’t available, so he was given bicarbonate in order to maintain his pH. His ventilation continued to deteriorate even further, at one point with the following blood gas (pH 7.09, pCO2 373 mm, bicarbonate 110 mM):
The patient eventually recovered and did fine. This case illustrates that it is possible to survive despite profound hypercapnia (using bicarbonate to avoid extreme acidemia). Similar reports exist regarding patients with severe asthma who survived despite having pCO2 levels of ~200 mm.5 These examples challenge whether extracorporeal CO2 removal is truly necessary in patients with purely hypercapneic respiratory failure.
Ultra-protective mechanical ventilation without fancy technology
Regunath et al. reported a retrospective case series of 15 patients with ARDS who were managed using ultra-protective mechanical ventilation (tidal volumes of 2-5 cc/kg).6 73% of patients in this extremely sick cohort survived. Patients experienced substantial hypercapnia with reasonable pH values:
Interestingly, only a single patient in this cohort required exogenous bicarbonate to facilitate ultra-protective mechanical ventilation. Nonetheless, the Henderson-Hasselbach equation shows that on average these patients had a moderate metabolic alkalosis (average bicarbonate level 34 mM). This underscores that patients with severe ARDS will often develop an alkalosis without exogenous bicarbonate, as discussed above.
This study illustrates that it’s possible to perform ultra-protective mechanical ventilation without extra-corporeal CO2 removal. Indeed, this may not be difficult at all. It’s merely a matter of occasionally administering bicarbonate and accepting a lower pH range than we usually do (e.g. target pH >7.15 rather than >7.20).
Ultra-lung protective ventilation in ARDS facilitated by extracorporeal CO2 removal
Xtravent study (2013): Lower tidal volume strategy (~3 ml/kg) combined with extracorporeal CO2 removal versus ‘conventional’ protective ventilation (6 ml/kg) in severe ARDS
This was a multicenter RCT evaluating the use of ultra-lung-protective ventilation combined with extracorporeal CO2 removal in 79 patients with ARDS.1 The primary endpoint was ventilator-free days. No differences were seen in any of the endpoints:
Extracorporeal CO2 removal was generally well tolerated, although it did cause some bleeding (increasing the average erythrocyte transfusion requirement from 1.5 to 3.7 units). Extracorporeal CO2 removal reduced the dose of midazolam required, suggesting that unsatisfied respiratory drive (breathlessness) contributes to agitation.
SUPERNOVA trial (2019): Feasibility and safety of extracorporeal CO2 removal to enhance protective ventilation in acute respiratory distress syndrome
The SUPERNOVA study is the largest available study on extracorporeal CO2 removal, so it provides the greatest power to detect uncommon adverse events.7 It’s a single-arm feasibility study evaluating the safety of extracorporeal CO2 removal to facilitate ultra-protective ventilation in moderate ARDS. The primary endpoint was the ability to achieve ultra-protective ventilation (tidal volume 4 ml/kg and plateau pressure <25 cm water) without increasing PaCO2 by >20% from baseline or decreasing the pH below 7.30.
The study contained 95 patients, with ultra-protective ventilation achieved in 82% of them. The pH actually increased from baseline, indicating effective CO2 elimination. So, extracorporeal CO2 removal does indeed remove CO2.
Unfortunately, extracorporeal CO2 removal is expensive and potentially dangerous. It requires placing a large-bore catheter, anticoagulation, and exposure of blood to an artificial membrane (which may consume coagulation factors). Overall, 39% of patients had adverse events related to this procedure (including two major events: cerebral hemorrhage and pneumothorax; table above).
So we can remove CO2, but the question is: should we? ARDS patients don’t need to have a pH of 7.40. We won’t earn any gold stars for achieving euboxia. Ultra-protective ventilation could be achieved in a cheaper and easier fashion by accepting a lower pH target and slipping the patient a bit of bicarbonate. A less invasive strategy to ultra-protective ventilation would also be more accessible to thousands of hospitals which lack the technology to remove CO2 extracorporeally.
The studies that we need
Clinical trials are often inspired by new technologies and industry sponsorship, rather than the need to answer fundamental questions. We put the cart ahead of the horse. Unfortunately, without a firm understanding of fundamental issues, it’s impossible to understand how we should be using new technologies.
We desperately need a better understanding of pH and pCO2 targets in ARDS. For example, the above study design might help illuminate what our pH targets should be and whether exogenous bicarbonate is a beneficial way to reach such targets. If we could gain some understanding of our therapeutic targets, then we would be better positioned to use extracorporeal CO2 removal in a logical fashion. The horse needs to go in front of the cart.
- Permissive hypercapnia is a cornerstone of mechanical ventilation in difficult-to-ventilate patients (e.g. ARDS, asthma). However, the optimal degree of permissive hypercapnia is unknown (e.g. the safe pH range).
- Intravenous bicarbonate can be used to avoid severe acidemia in the context of permissive hypercapnia. This was utilized in the landmark ARDSnet trial and seems reasonable, but it too remains untested by rigorous trials.
- Extracorporeal CO2 removal is effective at decreasing CO2 levels, and thereby facilitating more gentle ventilation. However, in the absence of any clear therapeutic targets (regarding pH or pCO2), it’s unclear where this technology might fit in our therapeutic armamentarium.
- Available evidence regarding extracorporeal CO2 removal has failed to show benefit, but has demonstrated potential harms (e.g. increased risk of bleeding, including intracranial hemorrhage). Given that this technology is expensive and invasive, widespread incorporation should await evidence of benefit in RCTs.
- SUPERNOVA review by James Walter of TheBottomLine
- IBCC chapter – Disseminated Intravascular Coagulation (DIC) - January 18, 2021
- PulmCrit- RCTs don't justify using convalescent plasma or antibody cocktails - January 14, 2021
- PulmCrit – Six RCTs to answer one question: what is the role of tocilizumab in COVID-19? - January 12, 2021