Hyperoxia during CPR associated with improved survival

There has been some great discussion on the web recently about the potential danger of hyperoxia. Mike McEvoy was interviewed on EMS 12-lead about the emergency and ICU evidence, and a vigorous discussion took place on EMTLIFE about the same topic. Rogue Medic weighed in as well, asking the question "How many hundreds of thousands of patient have we killed with oxygen and our refusal to require evidence of improved outcomes?" 

In the midst of this heated dialogue about reactive oxygen species, a recent study was published that may be an important addition to this discussion. 

The new paper
The study, "Increasing arterial oxygen partial pressure during cardiopulmonary resuscitation is associated with improved rates of hospital admission," was conducted in an Austrian EMS system. The ambulances are staffed with physicians, and are equipped with portable ABG analyzers as well. They retrospectively analyzed all non-trauma cardiac arrest calls over a 7-year span, and found 145 patients that had received ABG analysis of the PaO2 during the code. The ABGs were obtained after compressions, intubation, and 100% oxygen had been started. 

After dividing the patients into low, intermediate, and high levels of PaO2, they examined which patients had survived to hospital admission (HA), as well as the "cerebral performance category"(CPC) of the longer-term survivors. About half of the 145 cardiac arrest patients with an ABG had ROSC.

It turned out that patients with intermediate (61-300 mm Hg) or high (> 300 mm Hg) levels of PaO2 were incrementally more likely to survive to hospital admission. This positive association did not extend to showing a significant improvement in cerebral performance in survivors, however, despite a suggestive trend.

How does this fit with prior studies?
One retrospective study conducted in the ICU demonstrated an association between post-ROSC hyperoxia and increased mortality, while a similar trial showed no consistent association. No other trials have looked at pre-ROSC PaO2, however.

The accompanying editorial attempts to explain this seeming paradox - that hyperoxia pre-ROSC increases survial, but worsens survival post-ROSC - but I wonder if it is premature to try and reconcile these studies. Given all the known limitations of retrospectively-obtained data, and of this trial in particular, perhaps we should await controlled trials that more clearly define the role of oxygen levels and survival. As the authors note, "Reasons for the benefit of higher oxygen tensions during CPR can more easily be hypothesized than explained." Given the conflicting data, it might behoove us to proceed cautiously in modifying the targets for oxygen delivery in cardiac arrest.

Can capnography help the intubated trauma patient?

I had a conversation with a paramedic student recently, regarding a dyspneic patient who had either CHF or COPD. While I was explaining the utility of checking for JVD and other old-fashioned tests, she replied with "How about we just check the end-tidal CO2?" The dialogue about capnography continued something like this:

Paramedic: "The waveform might show some shark-fining, which would point to COPD, but if the nebs weren't working, and the patient looked shocky, you might worry that poor perfusion from CHF is producing a falsely low PetCO2, and you could start nitro."

Me: "After we intubate them, yellow means yes!"

I realized I wasn't contributing much to that discussion. 

Around the same time, I found a new study that looked at the use of end-tidal capnography to adjust ventilations for trauma patients, and the results were intriguing. So, in the interest of sounding smarter to the paramedic students, I plunged into the world of EMS capnography.

(Very) Brief review of capnography for EMS
People breath in oxygen, and breath out carbon dioxide. The level of carbon dioxide in the arterial blood is a very important number, and it's written as PaCO2, or the Partial pressure, in the artery, of CO2. Typically it runs around 35 - 45 mm Hg.

Google images - who knew?

This number is the one that counts when we're adjusting the rate or volume on a ventilator. Only problem is that we need to use a needle to draw an arterial blood gas (ABG), and then use a sizable machine to analyze the blood. It hurts too!

Of course, we breath out carbon dioxide, so we can also check the Partial pressure of CO2 as we breath out, especially the very last bit, at the end of the tide of airflow; the PetCO2. The point labeled "D" in the figure below marks the point at which PetCO2 is measured.

source

Usually the PaCO2 and the PetCO2 are pretty close to one another, and the arterial level is typically only 3-5 points higher than the end-tidal level. Another way to put it:

                   PaCO2 - PetCO2 ≤ 5 mm Hg

Well, usually that is ...

EMS and end-tidal capnography
EMS has been able to use capnography to do some important things. Rather than copy an extensive list, I'll turn this over to Peter Canning, over at Street Watch. He has compiled a list of the 10 Things Every Paramedic Should Know About Capnography, and it's a great focused summary. (Meaning, 90% of what I know comes from this article!). 

Of course, capnography can used to confirm intubation. It can also be used in cardiac arrest to check for ROSC, or assist in deciding on termination of efforts. Some evidence suggests that it has a role in diagnosing obstructive lung disease (asthma, COPD), as well as various other problems. 

First on his list, however, is its utility in monitoring ventilations, avoiding hypo- and hyperventilation. In the hospital, this is simple: If the PaCO2 from the ABG  is less than 35, the patient is being hyperventilated, and either the rate or the tidal volume needs to be decreased. With capnography, the numbers and waveform would look like this:

source

If the PaCO2 over 45, one of those parameters needs to be increased, because the patient is being hypoventilated. 

source

Because capnography is so much simpler and faster than using ABGs, it has been hoped that EMS could be able to modify ventilations just as well as in-hospital people can. So what does this recent study tell us about that potential?

"Utility of Prehospital Quantitative End Tidal CO2?"

Missouri EMS researchers wanted to test how well patients could be ventilated by EMS after intubation. They choose to focus on patients who had suffered either severe trauma or burns, and ended up with 160 patients (87% trauma, 13% burn-relate) who were transported to a level 1 trauma center. Overall, these were serious trauma cases - 75% had a GCS < 8 prior to intubation, and 1 out of 5 died in the hospital.

Paramedics had been trained in the use of end-tidal capnography to avoid hyper- or hypo-ventilation. During transport EMS recorded the PetCO2 levels, and adjusted the ventilations accordingly. Upon arrival to the ED, ventilations were maintained at the same rate and volume that EMS had used, and an ABG was obtained. The PaCO2 from this ABG was then analyzed against the end-tidal reading obtained during transport.

On average, the prehospital PetCO2 (34 mm Hg) was significantly lower than the ED PaCO2 (44 mm Hg); i.e. PaCO2 - PetCO2 = 10 mm Hg.  

This overall  difference between the PaCO2 and the PetCO2 only got larger when the sicker subsets of patients were examined. 

• Patients who died during hospitalization: PaCO2 - PetCO2 = 17 mm Hg.
• Patients with a pH < 7.2: PaCO2 - PetCO2 = 20 mm Hg.

So, were these results expected? What has the rest of the prehospital capnography literature showed? And how should we use capnography in the future?

FIrst, I'll review the studies that suggested that end-tidal capnography was potentially very accurate, and then I'll go over the studies that highlighted problems in applying it to the EMS patient population.


Studies that showed benefit of capnography
A 2003 helicopter study had suggested that capnography could help prevent hypoventilation in severely injured patients. Randomly assigned helicopter medics were able to use PetCO2 monitoring to adjust ventilation of intubated trauma patients, and ABGs were checked upon arrival to the trauma center. The patients were a mix of general poly-trauma, with high injury severity scores. The study suggested a big benefit - medics who had access to the ETCO2 monitor were far more likely to avoid hypoventilation and achieve normoventilation (although there was no change in hyperventilation).

Next, end-tidal CO2 was used to manage ventilations in a 2004 ground EMS study conducted in San Diego. The researchers enrolled 291 patients with severe head injury who had been intubated. Most of the patients had ventilation managed through standardized setting, but for about 1/2 of the patients the paramedics had ventilator management protocols that targeted a PetCO2 of 30-35 mm Hg, and avoided a PetCO2 < 25 mm Hg. As in the present study, the PaCO2 was confirmed by ABG after arrival at the ED. The use of the end-tidal capnography resulted in about 8% less hyperventilation.
 
Studies that suggested problems with it.
A 2005 study was conducted by a French EMS agency that uses specialist physicians, and uses ambulances equipped with ventilators, end-tidal capnography, and portable ABG analyzers as well. They looked at 100 patients that had been intubated over the course of 16 months, and examined how the PetCO2 levels corresponded to the PaCO2. An important note: only the PaCO2 values were used to adjust the vent. The patients were a mix of medical and trauma. They found that, even though, on average, the PaCO2 was the same as the PetCO2, there was significant variability in individual patients.

•  For 27% of the patients: PaCO2 - PetCO2 > 10 mm Hg
•  For 2% of the patients: PaCO2 - PetCO2 < -10 mm Hg

In other words, over a third of the patients would have been ventilated using false setting, had PetCO2 been used. In graph form:

An ED-based study from 2009, by Korean researchers, looked at 66 patients with severe head injury who had been intubated in the ED, and were mechanically ventilated. ABGs were obtained simultaneously with PetCO2 readings, and the paired values were compared. In general there was a good correlation between the two methods, and the PaCO2 exceeded the PetCO2 by less than 4 mm Hg, on average. However, this relationship broke down in the sicker patients; e.g. those with acidosis, greater injury scores, hypotension, or chest trauma.

A second 2009 trial conducted on 180 trauma patients who were intubated in the ED showed an extremely poor relationship between PaCO2 and PetCO2 obtained simultaneously. In the subset of patients with an isolated mild head injury the correlation was somewhat better. Noentheless, the authors concluded that: 

If the recommendations for ventilation to an PetCO2 of 35 mm Hg to 40 mm Hg were implemented in this population, 80% of patients would have a PaCO2 > 40 mm Hg and 30% would have a PaCO2 > 50 mm Hg.

Not good!

How to use the results of this new study.
Although these studies employed a variety of protocols (for example, different definitions of hyperventilation), 2 common threads  emerge. 

The first is that if the trauma is either mild, or limited to the head, then PetCO2 is probably an accurate surrogate for PaCO2, and can be used to modify ventilations. On the other hand, if a person has sustained trauma to multiple organ systems, or is showing any signs of shock, then the PetCO2 may (or may not) significantly underestimate the PaCO2 - there's no way to know. You're flying blind, vent-wise.

It turns out that this is sort of a common theme in using end-tidal capnography - it works best in patients with a single problem, but loses utility when the patient gets complex. Specifically, PetCO2 is no longer accurate when the patient has problems both with ventilation and perfusion.

Take asthma and CHF as another example. Both can present with hypoxia, true, but asthma usually only involves the lung, a ventilation problem. A number of studies have shown that certain qualitative aspects of the waveform - the "shark's fin" - may serve as a way to demonstrate improvement or worsening. 

Source

On the other hand, CHF can involve both the pulmonary and the cardiac systems, at the least. There can be a complex relationship between things that drive the PetCO2 down (like poor perfusion from systolic failure) and those that drive it up (such as impaired ventilation from coexistent COPD). 

With these complexities, it isn't at all clear how to use end-tidal capnography in CHF, despite the advice offered in some EMS magazines. The best research that described using capnography to diagnose CHF versus COPD/asthma comes from a single article in the Croatian Medical Journal. Not something to hang your hat on.

By way of contrast, you can use PetCO2 to predict the degree of metabolic acidosis in pediatric DKA (studies here and here), or in pediatric gastroenteritis. These are strictly problems of metabolic acidosis (which would be reflected in the PetCO2), while only in the rare extreme cases would perfusion be affected.

The Bottom Line
The current study is consistent with prior studies, and it appears that end-tidal capnography is not yet reliable enough to use in severely traumatized or burned patients. Using capnography in this population runs the risk of underestimating the PaCO2, leading to hypoventilation.

COPD: Is EMS Killing Patients with Oxygen? (2)

The Tasmanian Study
Disclosure - the lead author used to be my partner. We were both medics at the same two agencies (volunteer and hospital) in New Hampshire. He never seemed to run out of energy, and it doesn't seem like that's changing!

And since we're talking about Tasmania, we might as well get this out of the way...

Dr. Austin at work.

Effect of high flow oxygen on mortality in chronic obstructive pulmonary disease patients in prehospital setting: randomised controlled trial was well done, especially for EMS research. It's unfortunate that the bar is so low in our corner of health care research, but Austin et al. are working to change that.

The study utilized an EMS service that covered both rural and urban settings, but transported to only one hospital. Paramedics were supposed to identify patients with "breathlessness and a history or risk of COPD," and enroll those patients that they suspected of having an acute exacerbation of COPD.

Most of the treatment was the usual stuff - salbutamol (aka albuterol), ipratropium, dexamethasone, etc. CPAP was not available. The only difference had to do with how oxygen got delivered.

Prior to treatment, patients were randomized to receive either titrated oxygen, delivered by nasal cannula to maintain a sat around 90%, or high-flow oxygen, administered by NRB mask.  Transports lasted an average of 45 minutes, and then the folks in the ED treated as they saw fit.

Before we look at the results, you have to understand two huge issues:

• Only about half of the patients had pre-existing COPD, as judged by a pulmonologist. There's no description of what the rest had. Perhaps many of them had a new diagnosis of COPD, while others where actually pneumonia or CHF with wheezing.
• There were a lot of protocol violations: 56% of the titrated-oxygen group got high-flow oxygen, while 21% of the high-flow group didn't get high-flow.

With those facts in mind, they found:

Note how the results are broken down by all patients versus only those patients with previously confirmed COPD. The most important result is that there's a 5% difference in mortality for all of the patients.

So, what's the problem? It seems to be pretty clear evidence that paramedics should be withholding oxygen in suspected COPD exacerbations, but there are a few reasons why we should be cautious in using these results on the streets tomorrow.

1. The results seem almost too remarkable to be true.
They show 5% increase in mortality between the high-flow and the titrated oxygen groups.  That's a "number needed to harm," or NNH, of 20. In other words, for every 20 patients you give a NRB to, 1 will die.

That's a really high number! Usually you have to do a lot of work in medicine to show that kind of effect. For example, you have to treat 20 STEMI patients with aspirin and streptokinase (versus placebos) to save one life.

And you know all that fuss about rushing the STEMI patient into the cath lab? We have to treat 50 STEMI patients with angioplasty (versus thrombolysis), to save 1 life.

Given this context, the result almost seems "too good to be true." Given that there is little randomized controlled data out there with similar results, and given the good-quality data that conflicts with it (as with the ICU studies described in the prior post), it is reasonable to be skeptical about the conclusions, and await validating data.

2. The oxygen therapy was brief.
The intervention only lasted during transport, on average about 45 minutes. Compare that with the Gomersall study in my last post, where patients in the ICU received higher levels of oxygen for 2 days. While certain brief EMS interventions can have important long-term consequences (defibrillation anyone?), the onus is on the authors to make the case that it was solely the EMS intervention that differed between the two treatment groups. They concede that it was difficult to isolate the effect of the EMS intervention, and state that:

"Unfortunately, collection of data on in-hospital management was beyond the scope of the study, so we cannot dissect the effect of prehospital and in-hospital oxygen administration."

3. Paramedics don't treat "confirmed" COPD
Although withholding oxygen in the cases of "confirmed" COPD looked very effective, the unfortunate truth is that paramedics apparently misidentified half of the patients as COPD.

Now, this may not be entirely true. Some of the "non-confirmed" cases may have been a first-time COPD attack, and the medics were spot on. Or, perhaps some of the cases were actually, say, CHF, or pulmonary emboli. We aren't given any data on what this sizeable subgroup was ultimately diagnosed with.

Remember, though, that most other causes of hypoxia are treated with oxygen. With that in mind, we are left to consider the effect on mortality for the patients who did not have COPD. It does not seem likely that their treatment was improved by leaving them hypoxic.

By the way, this is the same reason you don't treat "hyperventilation" with a paper bag. Seriously, you aren't doing that, are you?

"Med control, we have a problem."

4. There was no difference in rates of mechanical ventilation.
In general, it's hard to show a difference in mortality for any given therapy, because deaths are far less frequent than other bad events. For instance, you need to treat 13 patients in CHF with non-invasive positive pressure ventilation (NIPPV) to prevent 1 death, but you only have to treat 8 patients to prevent an intubation (figures from The NNT).

In fact, we also know that NIPPV cuts the rate of mortality of COPD exacerbations by about 42%, but drops the intubation rate more, by 52%.

So it's kind of odd to see this profound difference in mortality, but no difference in the rates of mechanical ventilation. Similarly, there's no difference in the average length of hospital stay here, another sensitive measure of the effectiveness of a therapy.

5. Oxygen is not the issue.
Ultimately, the study is not asking a relevant question. We know that COPD is primarily an issue of impaired ventilation, and that NIPPV is the treatment of choice for severe exacerbations, along with steroids, bronchodilators, and antibiotics.

In a way, designing a study that only looks at levels of oxygen delivery in COPD exacerbation is sort of like designing a CHF study that only focused on fluid restriction, but didn't use nitroglycerin or NIPPV.

Yes, it would probably show a benefit of fluid restriction, but that's not the problem!

The Bottom line

1. Suppressing the "hypoxic drive" is a rare entity, and concern for it should not drive management.
2. Withholding oxygen could be very dangerous if you are wrong about the diagnosis.
3. Treat bad COPD with bronchodilators, steroids, antibiotics, NIPPV, and, if hypoxic below their baseline, oxygen. Intubate PRN.
4. We're looking forward to more high-quality studies from Tasmania!

COPD: Is EMS Killing Patients with Oxygen? (1)

So this was one of the first bits of medical "lore" I learned as a spanking-new EMT-B. ("Lore" meaning things that are not found in textbooks, but that everyone knows are true.) You don't give oxygen by NRB to a COPD patient.

And you always leave a note. (Hurry up Netflix!)

Well, it turns out a lot of things I learned at that time have been discarded; MAST pants, the EOA, the long spine board (soon, soon...). How has high-flow oxygen fared?

Are we killing patients?

A recent study suggests that the old-time myth was true. In fact, if this study is corroborated, we may need to change our practice somewhat. The one-sentence summary of Effect of high flow oxygen on mortality in chronic obstructive pulmonary disease patients in prehospital setting: randomised controlled trial is:

For high flow oxygen treatment in patients with confirmed chronic obstructive pulmonary disease in the prehospital setting, the number needed to harm was 14; that is, for every 14 patients who are given high flow oxygen, one will die.

That's a really high number.

By comparison, when we rush a STEMI patient to the cath lab, or give then thrombolytics, we have to treat 50 patients to save one life. This study suggests that we can save far more lives by using a nasal cannula instead of a face mask. Can this be true?

Let's first take a look at some older evidence, giving us some context. In the following post we'll go over the new article that has generated so much interest, and see if some practical conclusions can be drawn.

Two ICU studies
The authors of Influence of inspired oxygen concentration on deadspace, respiratory drive, and PaCO2 in intubated patients with chronic obstructive pulmonary disease wanted to study this (possible) myth of the hypoxic drive. They studied the effect of increasing oxygen delivery to a very sick bunch of COPD-ers, 12 patients who had already been intubated after a bad COPD episode. These patients were just starting to recover, being weaned from the vent, able to breathe or their own, and were likely to be extubated in the next day or so. 

The researchers bumped up the oxygen level to 70% for 20 minutes, and checked what happened with the vent and the blood gas. Although 70% doesn't seem high, it is actually right about what we are delivering with a standard non-rebreather mask at 15 lpm (See Weingart's article for explanation; PDF if you prefer).

So what happened? Apnea? Bradypnea? Failure of the hypoxic drive????

They, uh, got more oxygen. That's it. In this population of sick-sick-sick patients, nothing happened.


Okay, you say, perhaps they were a delicate population, but they had been getting beta-agonists and steroids, and probably antibiotics for a few days. That's the whole reason, after all, that they were being weaned off the vent - they were now better. 

Maybe we should instead look at patients who are not yet intubated, but might be if just one more thing tips them over.

"Our study included one patient - this guy"

So, back to the ICU. The authors of Oxygen therapy for hypercapnic patients with chronic obstructive pulmonary disease and acute respiratory failure: a randomized, controlled pilot study. looked at patients right on the brink, with PaCO2 > 50 mmHg, and PaO2 < 50 mmHG - members of the 50/50 club! They were not yet intubated, but it could go either way.


The investigators titrated oxygen up with Venturi masks to two different levels, either a PaO2 of > 50 mm Hg in 17 patients, and > 70 mm Hg in another 17 patients. They got all all the usual meds, of course. As expected, some of these very sick patients had to be intubated, and some even died. The rates of intubation and death, however, were the same in each group. And no CO2 retention either!

So why all the fuss? As an editorial in Critical Care Medicine pointed out, the main evidence for the harm of oxygen in COPD exacerbations comes from the 70s, a time before noninvasive ventilation, routine steroid use for COPD, and even Atrovent was barely a year old! Another editorial from the same journal lays into the medical-education complex for perpetuating this lore:

"One sample of medical mythology is the commonly told story that the administration of oxygen to a patient with chronic obstructive lung disease will shut down the patient's hypoxic respiratory drive and lead to apnea, cardiorespiratory arrest, and the subsequent death of the patient. ... It is not clear where this fallacious information comes from, but it seems to enter the medical information database at an early age, at the medical student or resident level, almost like a computer virus corrupting the appropriate function of the equipment. In addition, this myth becomes very difficult to extinguish during the career of the physician, even with clear factual information of long standing. The danger here is that this medical mythology will inappropriately influence treatment decisions in patients."

The Bottom Line
This is hardly a comprehensive review of all the literature out there on the topic, but most of the other studies are observational, and it's really hard to draw firm conclusions from that sort of data. I haven't bothered to review them here, because why waste our tine if higher-quality studies have been done?

With this background, I'll discuss the important EMS study from 2010 in the next post.