Saturday, January 7, 2012

Specious Ideas: CO2 is bad for you and acidosis is a killer too

Here's some bad acid and CO2
Certain biases in the way we think in medicine are so pervasive and so misleading that I'm going to make a special section on this blog for them:  The Specious Ideas Series.  We'll begin with my favorite, the approach to hypercarbia (high CO2) and acidosis (low pH) both in and out of the ICU.

Patients with serious illness often have derangements in PaCO2 (arterial partial pressure of CO2; hereinafter abberviated CO2) and pH.  But it does not necessarily follow that patients are experiencing physiological stress because of CO2 and pH, and it likewise does not follow that actions to normalize these values will make patients better.  Indeed, such actions are often not only a distraction and a waste of time, but they can also be harmful.

How do I know that acidosis and hypercarbia (high CO2) are not harmful per se?  Because I have seen countless patients tolerate them with no sequelae (if you are a health care provider, you probably have too), and because the issue has received some study.  For example:
  • Patients with diabetic ketoacidosis (DKA) frequently have a pH < 7.0, and they almost always recover without sequelae if the DKA is treated and reversed.  While it is true that the acidosis contributes to their feeling ill, and that they have Kussmaul respiration as a result of it, it is also true that we do not treat the acidosis with bicarbonate; rather we work to treat the underlying disease.
  • In a 1959 study entitled Apneic Oxygenation in Man, the authors demonstrated the apparent safety of allowing pH to fall acutely to as low as 6.7, and CO2 to climb to as high as 250; again, there were no apparent sequelae.  (I recognize that "apparent" is the operative word here; there are no "apparent" sequelae to climbing everest without supplemental oxygen unless you do intensive neuropsychological testing - then subtle post-summit performance differences become apparent.)
  • Patients with COPD and asthma, and the obese with OHS (obesity hypoventilation syndrome - "Ondine's curse") can have chronically elevated CO2 in the range of 80-100, with acute increases on top of that and they too survive without apparent sequelae (for an extensive review, see:  Feihl and Perret, AJRCCM, 1994).
If pH and CO2 are not harmful, then why is there such a focus on them, and why do we get arterial blood gas samples (ABGs)?  Here we need to make a pivotal distinction between the ABG as a tool for evaluation versus a target for intervention.  An ABG in a patient with COPD who is struggling mightily to breathe allows us to determine if he is losing that struggle and is at risk of stopping breathing altogether.  In such a patient with, say, a respiratory rate of 30 and a pH of 7.19, the ABG may tip you off that something needs to be done to reverse the underlying disease process and prevent respiratory arrest.  (It is also true that such a patient can be managed without the ABG by simply assessing the work of breathing alone - "mightily" is a strong word.)  The ensuing action may be the use of nebulizers, BIPAP, endotracheal intubation, etc.  Follow up ABGs may be used to determine the effectiveness of nebulizers and BIPAP, but they are not necessary - reduction in respiratory rate, work of breathing, and heart rate are adequate markers of improvement, obviating ABGs in many cases.  (To my astonishment, respiratory rate, which is for me the most important vital sign in a patient with a pulmonary problem, is often not even reported among vital signs when I inquire about them.)  In this case, the ABG has been used as a tool for evaluation of the patient, one piece of data among many that together allow a holistic assessment of his respiratory status.

Now, suppose that the patient described above has been intubated and sedated, has a respiratory rate of 15 (ventilator set at 15) and a repeat ABG shows a pH of 7.21.  Is intervention indicated now?  Should the respiratory rate be increased?  For me, the answer is a qualified "no".  We have already recognized that there is a serious underlying disease, and we have already taken corrective action to prevent further deterioration and respiratory arrest - we have secured the airway.  He cannot now have a respiratory arrest, breathe a sigh of relief, my friends.  (But he can "Code" if he's overventilated in the course of efforts to normalize his ABGs.)  Now, if there are no other underlying conditions that limit tolerance of high CO2 and low pH (such as certain cardiac, pulmonary or intracerebral conditions the discussion of which is beyond the scope of this post in part because they are infrequently relevant) we can relax and await the abatement of the underlying disease.  Eventually, pH will improve in parallel with reduction in sedation, which is blunting respiratory drive, and relief of obstruction, which is causing both increased Vd/Vt (dead space) and CO2 production.

If this same patient were mechanically ventilated and continued to breathe (with apparent difficulty) at 30 breaths per minute with pH 7.21, I still would not intervene by changing ventilatory parameters to normalize the pH, but rather would try to determine why his minute ventilation and work of breathing are so high - is it because of dead space?  Agitation from fear and pain causing increased CO2 production?  Increased CO2 production from high work of breathing from ongoing obstruction?  It's easy to normalize blood gasses as a target for intervention; it is much harder to use them as a tool for evaluation to figure out the underlying problem and treat it specifically.  (Another good use as a tool for evaluation - the unresponsive patient with agonal respirations.  In this case, usually the ABG just confirms what you already know, and delineates the severity.)

The only time that I use AGBs as a target for intervention is when a neurosurgeon tells me to (that's another blog post altogether), or when we have a patient paralyzed and the intrinsic ventilatory control system has been rendered totally non-functional.  Yet most physicians, nurses, and respiratory therapists conceive of the ABG is seen as a target for intervention so they scurry about all night adjusting ventilatory parameters to create "normal" ABG results.  The benefit of this is uncertain, but the financial costs and drawbacks in terms of blood loss are clear.  The most troublesome cost may be an opportunity cost - the missed opportunity to identify the underlying problem.  For me, it is all just a waste of time because I am confident that as the underlying disease resolves (assuming I am specifically treating it), the ABG will revert towards normal.  There is no need to force it to be normal any more than there is a need to leukopherese a patient with pneumonia and leukocytosis.

A more pernicious result of the irresistable siren call to normalize CO2 is sometimes suffered by the obese.  The "can't breathe" versus "won't breathe"distinction is often neglected, in large part because of  ignorance of respiratory rate as a vital sign, and absense of the skill needed to perform a holistic assessment of the pattern of respiration and work of breathing.  The ABG is used as a woefully inadequate surrogate for this assessment.  Because of their abnormal ventilatory response to CO2 and pH, the obese (and the burgeoning population of patients addicted to narcotics, benzodiazepines, and every other manner of sedative) will often have an uncompensated respiratory acidosis when they present with a variety of acute medical problems. For various reasons, these patients can have low pH with absolutely no signs of resultant physiologic or metabolic stress - they simply need the acute problem that upset their fragile state of chronic compensation to be identified and treated.  (Patients with chronic renal failure who have lost the ability to compensate for chronic respiratory acidosis often are walking around with pH < 7.2 too.)  In any case, it makes no sense to react to a pH of 7.21 in a patient with a normal respiratory rate because such a patient can breathe but simply won't breathe - it's a lack of will rather than a lack of ability.  (In life in general, lack of will is more common than lack of ability.)  Intubate them and turn up the rate all you like, you're not going to teach them to breathe faster.  You can lead a horse to the water, but you can't make him drink.

All too often, an obese patient with pneumonia or cellulitis, respiratory rate of 18 and pH of 7.21 will suffer the indignity and danger of a profligate and superfluous intubation because of failure to make the can't breathe/won't breathe distinction.  Such a patient is prone to complications during difficult intubation; atelectasis during rapid sequence intubation; difficulty recruiting atelectatic lung after intubation because of low respiratory system compliance; requirement of high PEEP to maintain oxygen saturations; provider confusion between atelectasis, ARDS and pneumonia; prolonged ventilation and "failure to wean"; oversedation and accumulation of sedative drugs; risk of tracheostomy; and so on.  Their worsening status after the intubation is mistaken as evidence of evolution of the underlying disease rather than a result of the misguided interventions.  To date I have extubated about a dozen such patients directly from PEEP of 16-20 (to CPAP) so that I could stop sedation and allow them to recover from Status Iatrogenicus.  Not one of them has had to be reintubated.  But because of the uncertainties involved and the recalcitrant nature of atelectasis in the extremely obese, especially when sedated and intubated, it is a harrowing process to slow down this runaway train, which was set into motion by misunderstanding of basic respiratory physiology.

One phenomenon worthy of mention that we have observed occurs in obese patients with untreated obstructive sleep apnea (OSA) who present with some acute problem or decompensation and are ventilated with non-invasive positive pressure ventilation (NIPPV), because of hypoxemia or respiratory acidosis.  Occasionally these patients become totally unresponsive on NIPPV, initiating the sounding of alarms for impending respiratory arrest ("CO2 narcosis" - I love that one).  (They have often received sedatives acutely and/or chronically as well, which complicates the situation and its interpretation in several ways.)  An ABG is obtained, the pH is low and the patient is intubated.  The patient remains unresponsive and an extensive and expensive search for causes of coma is initiated, but is unrevealing.  In about a half a dozen such patients to date, we have ruled out everything except sleep - apparently the colloquialism "I was in a coma" is not far off mark.  (We have yet to prove this with an EEG, but that is in the works.  Other experimenters are already on the trail - see:  Aldrich; and Verma.)  My hypothesis regarding these patients is that they are wildly sleep deprived from untreated OSA; they come to the hospital for decompensation of some sort; they receive needed relief with oxygen, morphine, lasix, etc;  their untreated OSA is treated by the NIPPV; this results in a deep, much needed, restorative sleep, the only good sleep they have had in years; they are difficult to arouse from this deep sleep; they are intubated for CO2 narcosis; they continue to sleep and be virtually unarousable for days.  That's right, obese people are being intubated for sleeping too soundly - keep the perpetrators of this injustice away from my house on Saturday morning!

I do not mean to suggest here that acidosis is never harmful or should not induce concern; that CO2 narcosis does not occur; that patients with pH 7.21 do not require urgent evaluation and treatment; that most patients with pH < 7.3 simply have lack of respiratory drive; that "he's just sleeping" should be the default response to a call about an unresponsive obese patient with CO2 of 100.  I only wish to underscore the fact that abnormalities on ABGs must interpreted in the greater context of the patient's illness if subsequent interventions (or lack thereof) are going to benefit patients.

I'll conclude this diatribe with a little limerick:

Don't let the A-B-G lead you to seethe
Go see the patient and watch him breathe!
     If he's easy and slow
     Just let him go
Your laryngoscope, young man, you need not unsheathe!

But sometimes despite pH of seven-four-0
The patient's tripodding - you already know
     His rate is too fast
     For long he won't last
Very soon he'll be smoking an eight-point-0!

(And Yes, I'm going to recycle that battery.)


  1. I wish I had time to make a more profound comment, but it's late and I'm in the middle of a Peruvian desert packing to go home. Your other blog is good, but you are seriously on your way to a Pulitzer prize with this one....

  2. I got in a shouting match in the Bview CICU with an anesthesiologist over this exact issue.

    Good on you Dr. Aberegg...

  3. Was this written by the guy who uses "concentrated bicarb" in ARDS patients in an attempt to get them to tolerate lower tidal volumes and/or HFOV?

  4. Indeed it was. I'm very fond of concentrated bicarb. For very specific and sound reasons. We still should do that trial, I have the folder for the design on my desktop still

  5. We just had another patient with untreated OSA this week who was comatose for 36 hours on BIPAP. EEG confirms sleep. Time for a case report/series.

  6. i am always in a quandary about what to do in a patient with a bmi of 80, who is transferred to the icu for altered mental status, gets better on bipap, but intermittently goes into coma like you described, either on or off the bipap. I have sent atleast one of these patients for a trach, in the hope that that would atleast relieve their sleep apnea. But I don't know if that was a misguided decision. I don't know how to safely say that those who are comatose will not just become profoundly acidotic and bradycardic in their 'sleep-coma' and die, especially when they are not monitored at home.

    1. There is no satisfactory boilerplate approach to these patients. When ventilatory control is askew because of obesity, it is much harder to figure out what is going on with a given patient, and it requires thoughtful consideration of all the variables. If an obese patient is maintaining "adequate" minute ventilation (as evidenced by PaCo2 <100, respiratory rate greater than or equal to 12, let's say), careful observation and NIPPV may be sufficient. If they are distressed (as evidenced by tachypnea and high work of breathing) to sustain marginal ventilation, that is a different story altogether. If the respiratory rate slows to lower than 10-12 breaths, I would look for another cause of depression of ventilatory drive (medications, thyroid). Recall that if you weigh 400#, your CO2 production is through the roof - you can do the calculations or measure it next time you have a Swan-Ganz in a patient (if you're still committing that offense against humanity) - you will find that CO2 production and required alveolar ventilation are much higher than a person of normal weight. I point this out to reinforce the notion that it is a mismatch of CO2 production and ventilation in these patients, rather than an inability to maintain normal ventilation levels like advanced COPD, where high Vd/Vt (dead space fraction) makes normal MINUTE ventilation inadequate for NORMAL CO2 production - and these patients are distressed with high respiratory rate, and generally high respiratory drive. If THAT person becomes comatose with a CO2 over 80-100, especially if they have a NORMAL respiratory rate, they may be on the verge of arresting. Clearly, the situation is highly nuanced adn requires expert level skill to properly assess. I want to point out that the way many of us were taught to make the assessment (High CO2, unarousable patient) may not ALWAYS yield the correct assessment which is best for the patient. As I described in the post about "empiricism" last year, if you always intubate such patients, you will pat yourself on the back for "saving" them - even if they have a trach 2 weeks later because of "fatalectasis", high PEEP requirements and "failure to wean".

      Finally, the decision to trach a patient for refractory OSA is a difficult one, that I as a non-sleep specialist would not like to make. If that decision is being made solely on the basis of mental status and ventilatory status, however, the above mentioned cautions and caveats would still apply.