Is CPR effective? Can ice save a life? Rethinking Resuscitation

Given the importance of running effective resuscitations on critically ill patients one would think our rationale would be firmly grounded in the literature. Unfortunately, the reality is that much of what we do to resuscitate these patients has as much science behind it as burning witches or sacrificing virgins. So, the next time you’re ready to do a rain dance in the resuscitation room, focus on the evidence, make good choices and rely on tradition only where no better answers exist.

Myth #1: CPR is effective

Will we ever see a day without CPR? Perhaps not. More importantly, will anyone ever be able to prove that it makes a difference in outcomes? I seriously doubt it. One thing seems to make a difference in non-traumatic cardiac arrest: electricity! Generally speaking, those with the best chance of survival are those with a shockable rhythm who had access to prompt defibrillation. CPR sounds like the right thing to do while your waiting for the defibrillator to arrive. After all, you can’t just stand there! In an article by Sayre it was reported that CPR without ventilation, often a deterrent to bystander CPR, was just as effective as CPR with ventilations (Circulation 2008 Apr; 117:2167). Are they saying that circulating deoxygenated blood is an effective way to perfuse a pulseless patient? Nope. What they are saying, whether they realized it or not, is that ventilations didn’t matter because CPR just doesn’t work.

Myth #2: Bicarb improves outcomes in metabolic acidosis

I recently read a chart on a 32-year-old female with type I diabetes with pyelonephritis and sepsis. Her glucose was elevated at 668 and her serum CO2 (bicarbonate) was 15. Arterial pH was 7.19. Besides her IV fluids and insulin drip, “1 amp of NaHCO3–” was ordered as well. Intuitively, this makes sense. She has DKA and is in metabolic acidosis. So, if your “bicarb” is low, give some back, right? While “bicarb” has its place, there isn’t a shred of evidence suggesting that it improves outcomes in metabolic acidosis from any source (Chest. 2000; 117:260–267). If you’re treating metabolic acidosis, treat the underlying cause and hold the “bicarb.”


Myth #3: ABGs are necessary to measure acidosis

The patient above had an ABG to measure her actual pH. Although she is clearly acidotic, based on her serum CO2 level, there are some additional advantages to obtaining a blood gas. Knowing the actual pH is a more direct measure of acidosis than the serum CO2, and it gives us a good sense of how effective the patient is compensating with their respiratory system. It also gives us clues regarding the work of breathing and who is likely to experience respiratory failure from just getting way too tired from trying to compensate for the acidosis. A blood gas does all of these things. However, who ever said we have to check an “Arterial” blood gas? Plenty of physicians, for many generations have and still do. ABGs are not benign, require special skills to obtain and often special personnel to obtain them. In recent years, many have questioned this practice in lieu of a venous blood gas or VBG. It’s less expensive, safer, easier to obtain and takes less time. What do we give up with a venous sample? Not a thing! It has been conclusively determined that the differences in pH and pCO2 comparatively between an ABG and a VBG are negligible (Emerg Med J. 2006 Aug;23(8):622-4, Emerg Med J. 2007 Aug;24(8):569-71). The only expected difference is in the pO2. However, we can easily overcome that issue with pulse oximetry.

Myth #4: Early Goal-Directed Therapy is the wave of the future

The EGDT ship sailed from port with 90% of the specialty buying its preliminary findings published by Manny Rivers in 2001 (N Engl J Med. 2001 Nov 8;345(19):1368-77). It takes no stroke of genius to throw a potpourri of interventions at a single patient and find some benefit (Chest. 2007 Aug; 132(2):425). The real question is, of the myriad of treatments employed, which of them really have beneficial effects? Bill Cosby could have added a Jell-O pudding pop to the EGDT sepsis approach and claimed its beneficial effects too. For me, if its vanilla, I’ll take the pudding pop over centrally-measured ScvO2s any day. In a nutshell, although some say EGDT may have a benefit for the sickest sepsis patients who can be identified using the MEDS score, I say the shine is fading. We know that early antibiotics and fluid resuscitation are very important. Nothing new there. We also now know that lactate may be an early marker for poor cellular perfusion associated with sepsis. This is a piece of information that has been fruitful from the sepsis literature. Interestingly, many physicians don’t know that the serum lactate will rise much more quickly than the serum CO2 will decline. Thus, if you are assessing for sepsis with a CO2 and not a lactate, you’re likely to miss some early cases. Perhaps the most important reason why sepsis patients showed improved outcomes with EGDT sepsis is because we were paying closer attention. We are detecting them earlier and managing them more aggressively (Chest 2006 Feb;129 (2): 225-232 ).

Myth #5: Use Steroids for sepsis

What about steroids in sepsis? “Stress dose? No benefit. Then, low dose was believed to be the way to go (Ann Intern Med. 2004 July;141(1):47). Now, that appears to have been refuted (N Engl J Med. 2008 Jan; 358(2):111). We are bound to repeat history if we are ignorant of our past. We have been studying anti-inflammatory agents for sepsis for decades. Perhaps, they just don’t work.


Steroids may benefit critically ill patients with adrenal suppression. However, identifying those patients is challenging. Some data suggests that cortisol levels have no predictive value for those who will benefit from corticosteroid administration compared to those who won’t (Arch Surg. 2005 July; 140:661). The key is, if you suspect adrenal insufficiency (i.e. hypotension, refractory hypoglycemia, hyponatremia, vomiting and hyperkalemia, etc.) just treat it. Steroids are cheap! 

Where adrenal insufficiency has emerged as a critical consideration for EPs is in the case of etomidate for rapid sequence intubation (RSI). In continuous infusion, three days and beyond, adrenal suppression and increased mortality has been shown to be higher in trauma patients. Other patient subsets have experienced this phenomena as well (i.e. sepsis patients). What really isn’t clear is whether or not one dose for RSI will have that same result. My money is on the logical explanation. Although some patients do worse with etomidate, they tend to be the sicker patients who might already have developed adrenal insufficiency from sepsis, etc…, regardless of exposure to etomidate. Honestly, we don’t have the definitive answer yet. However, we should all be mindful of this potential phenomena and recognize that this response to etomidate is both cumulative and dose-dependant. Thus, repeat doses of etomidate in critically ill or injured patients is ill advised and consider the use of alternatives for patients with known septic shock (Chest. 2005 Mar;127(3):1031-8, Acad Emerg Med. 2009 Jan; 16(1):11).

Myth #6: Control that glucose!

For years, and most recently with the “Surviving Sepsis” campaign, the evidence suggested that hyperglycemia was a bad prognostic indicator for those with significant disease (i.e. sepsis, pneumonia, CAD, trauma, etc.). Well it might be, since sicker patients tend to have higher serum glucose levels, but hyperglycemia may not be acutely harmful. Emerging data has put the brakes on “Tight glucose control” (JAMA 2008 Aug; 300(8):933).  Sure, if your patient is substantially hyperglycemic, treat it. That’s not what we’re talking about. The past evidence suggested that critically ill or injured patients should be kept in fairly tight glucose control during their illness. What we know now is that hyperglycemia in critical illness is most likely a self-protective mechanism in response to physiologic stress. So, if the body has this figured out, stay out of the way. If the glucose is only mildly or moderately elevated and that elevation is likely in response to other problems, resist the temptation to treat it.

Myth #7: Therapeutic hypothermia is too expensive & unsubstantiated

Although there are many critics of therapeutic hypothermia, this concept has been endorsed for years by the International Liaison Committee on Resuscitation (ILCOR). While the data is limited, it is still fairly compelling. (New England Journal of Medicine. 2002 Vol. 346, No. 8). In patients with a primary v-fib arrest who have spontaneous return of circulation, but do not regain consciousness, their chances for a better neurological outcome may be enhanced by doing this. Hey, if they survive in the ED, we’re sending them to the ICU anyway. Why don’t we give them every possible chance at survival? The only strong argument against this modality is cost. If you keep the medical device companies out of this and stick to the basics, this can be a bit messy, but not expensive. Ice is cheap!



Myth #8: Using Amiodarone increases survival rates

Everyone using amiodarone instead of lidocaine? If so, why? This move was made by the American Heart Association in the 2000 guidelines. Interestingly, this was based on the ARREST trial, (N Engl J Med. 1999; 341:871–878), comparing amiodarone to placebo. The ALIVE trial compared amiodarone to lidocaine (N Engl J Med. 2002; 326:884–890). In both studies, more treated with amiodarone survived to ICU admission. I think the AHA stopped reading the articles at that point. The numbers discharged from the ICU with a good neurological recovery were equal in both groups for both studies. All amiodarone accomplishes is taking up an ICU bed you’ll need for someone else and jacking up the costs for the family who’ll lose their loved one anyway.

Myth #9: Rectal exams mandatory in trauma patients

If your favorite procedure is the rectal examination, you’ll be sadly disappointed. The ACS now states that rectal examinations are not mandatory in all trauma patients. Further, they support the literature, which confirms that patients don’t conceal their bladder or urethral injuries. These patients will have physical findings, such as hematuria, trauma in the perineum or signs of a pelvic fracture that should raise your index of suspicion (J Trauma. 2005 Dec;59(6):1314-9). If these signs are not present, no need to glove up! Also of importance is that the practice of checking for poor rectal tone as a sign of spinal cord injury is nearly worthless. If this sign is relied upon, 63% of spinal cord injuries could be missed (Ann Emerg Med. 2007 Jul;50(1):25-33, 33.e1).

Conclusion: Learn to ask “Why”

Resuscitation is one of the most critical areas of emergency medicine. And although it’s rich with assumptions and dogma, it is short on evidence. Perhaps, one of the most critical resuscitation skills all EPs should develop is when to ask, “Why?” Being critical of the “Why” will prompt better answers for future generations of physicians and better care for our patients.

Kevin Klauer, DO is the editor-in-chief of Emergency Physicians Monthly



  1. Hmm… I’m not convinced that “CPR is not effective” is a conclusion that you can draw from the with vs without ventilation study.
    Perhaps ventilation is provided by the compression and expansion of the chest cavity when CPR is administered properly?

  2. Benjamin Abella on

    I am an expert in cardiac arrest resuscitation and serve on the American Heart Association guidelines committee that reviews all evidence for BLS and ACLS. As such, I feel it my duty to correct the writer of this piece regarding point #1. His statement is not based on any clinical data, and in fact the literature suggest the OPPOSITE is the case.


    The author may not be aware of the fact that there are at least 5 studies that come to mind IN THE LAST 5 YEARS (and there are plenty before that) that show that CPR doubles survival, holding all else equal, regardless of shockable rhythm/defibrillation. Some examples: Christenson et al, Circulation 2009 clearly showed in a very large clinical cohort that the more compressions that are delivered to patients, the more they are succesfully resuscitated. Or perahsp Bobrow et al, JAMA 2008, where in a large prospective interventional study, removing ventilations from initial CPR TRIPLED survival in the state of Arizona from cardiac arrest. Or perhaps Abella et al Circulation 2005 and Kramer-Johansen et al Resuscitation 2006, both showing that improving the quality of CPR had important statistically significant improvements in survival. The list could go on.

    So, I might suggest that POOR QUALITY CPR doesn’t improve survival, but it is very clear from the evidence based literature that CPR indeed works, matters, and should be delivered scrupulously.

    The author of this piece dramatically misinterprets the Sayre data, which show that as performed by bystanders, inclusion of breaths didn’t improve survival over not having breaths… this is very different from saying CPR doesn’t work.

    I invite the editors to make sure the author is aware of this, for his point #1 does not serve patients or physicians in their care of cardiac arrest patients.

    Benjamin Abella, MD MPhil
    Clinical Research Director
    Center for Resusciatation Science
    University of Pennsylvania

  3. Marvin Wayne, MD on

    We have established and implemented a community hospital therapeutic hypothermia program over the past 4.5 years. We have cooled about 152 patients (topical cooling with device) and have a go home survival from all rhythms of 15% with Mean Rankin Scores (CPC also) of 2 or less. Age range 20-82. 1/3 or these had PEA/Asystoli as presenting rhythm. So indeed this is inexpensive and works.

  4. CPR does improve survival. The lack of benefit from ventilation does not mean that CPR does not work.

    Effectiveness of bystander cardiopulmonary resuscitation and survival following out-of-hospital cardiac arrest.
    Gallagher EJ, Lombardi G, Gennis P.
    JAMA. 1995 Dec 27;274(24):1922-5.
    PMID: 8568985 [PubMed – indexed for MEDLINE]

    “CONCLUSION–The association between bystander CPR and survival in out-of-hospital cardiac arrest appears to be confounded by CPR quality. Effective CPR is independently associated with a quantitatively and statistically significant improvement in survival.”

    More recent is the research by Dr. Ewy on compression-only CPR or CCR (CardioCerebral Resuscitation):

    Cardiocerebral resuscitation improves neurologically intact survival of patients with out-of-hospital cardiac arrest.
    Kellum MJ, Kennedy KW, Barney R, Keilhauer FA, Bellino M, Zuercher M, Ewy GA.
    Ann Emerg Med. 2008 Sep;52(3):244-52. Epub 2008 Mar 28.
    PMID: 18374452 [PubMed – indexed for MEDLINE]

    “RESULTS: In the 3 years preceding the change in protocol, there were 92 witnessed arrests with an initially shockable rhythm. Eighteen patients survived (20%) and 14 (15%) were neurologically intact. During the 3 years after implementation of the new protocol, there were 89 such patients. Forty-two (47%) survived and 35 (39%) were neurologically intact.

    CONCLUSION: In adult patients with a witnessed cardiac arrest and an initially shockable rhythm, implementation of an out-of-hospital treatment protocol based on the principles of cardiocerebral resuscitation was associated with a dramatic improvement in neurologically intact survival.”

    When you discuss the medications for resuscitation – epinephrine, amiodarone, lidocaine, et cetera, I agree that there is no evidence of any effect other than improved return of a pulse, but at what cost? Why can’t we keep these patients alive to a discharge with good neurological function?

    Maybe the therapeutic hypothermia will help by addressing some of the apparent harm from the drugs we keep giving in spite of a lack of evidence of an improvement in any meaningful outcome. Maybe the drugs will have some specialized cases of cardiac arrest where they do improve outcomes, such a calcium for hyperkalemia. Until we have evidence of benefit, these drugs should be limited to controlled studies, since they are just experimental.

    You could say the same thing about amiodarone for termination of VT. The conversion rate half an hour after initiation of amiodarone is about 30%. Procainamide and sotalol do much better. For EMS, we should just consider a loading dose of sedation and being prepared to cardiovert if the patient does deteriorate prior to transfer to the ED.

    We need to make sure that all personnel have experience using the cardioverter, at least on a mannequin. As with pacing, cardioversion seems to be more often performed incorrectly. We have more problems with operator error, than with anything else.

    I completely agree with your conclusion. EMS could easily be described as Emergency Mythology Services.

  5. It is strange that the conclusion of your post dismisses EMS as “Emergency Mythology Services” when the beginning of your post quotes from a study showing that 15% of patients with out-of-hospital cardiac arrest (presumably treated by EMS) survived with good neurologic recovery. How many interventions in medicine confer a 15% survival advantage? Not many.

  6. RWH,

    Much of what we do is not based on evidence.

    Some of what we do is based on evidence.

    We need to eliminate the treatments, such as epinephrine for cardiac arrest, that do not have evidence to demonstrate improved meaningful outcomes. This is not just about getting a pulse back, but a person. There are plenty of other examples. Unfortunately, we do not do a good job of eliminating the witchcraft.

  7. About Myth #7, I looked over the study published in the New England Journal of Medicine, and I would like to add some thoughtful insights like Dr. Greg Henry did with CPR and ACLS. The data does not show compelling evidence that hypothermia is beneficial. If you look over the data critically, these are just a few of the short-comings: (1) small study, (2) not double-blinded, (3) insignificant changes in small number of patients (patients ended up in a vegetative state and died after discharge 6 of 8 for normothermia group and 4/6 in hypothermia group), no control for bystander CPR, time from witnessed arrest to defibrillation in v fib/v tach, and last but not least, there was a higher survival and lower comordities (diabetes and coronary artery disease) in the hypothermia group which leads me to believe that there were more healthy patients in the hypothermia group. Even the author stated that the “results were inconclusive.” Before we make hypothermia a standard of care and increase healthcare costs with this labor-intensive treatment, I ask all physicians to look at the data before incorporating hypothermia in your treatment algorithm.

  8. Gary Grist RN CCP on

    Most perfusionists will say that the failure of CPR does not necessarily mean that the patient is dead. Extracporporeal cardiopulmonary resuscitation (ECPR) can be effective in saving 30-40% of in-house cardiac arrest pediatric patients who do not achieve ROSC. Potentially, if cooling is instituted “in the field” during out-of-hospital CPR and the patient brought to the ED and placed on ECPR, there is a good chance for revival. The most important thing is that the physician knows how to approach the problem of reperfusion injury by using the pump for core cooling and revival. Here are a couple of articles that explain the concept:

    Grist G. Extracorporeal membrane oxygenation (ECMO) or extracorporeal cardiopulmonary resuscitation (ECPR): A critical life or death choice. Progress in Pediatric Cardiology; 2008 January; 24(2):113-116.

    Grist G. Perfusionists Need to Partner With A New Kind of Physician. AmSECT Today. May/June 2009, 12(3): 6,13.

  9. Kevin Klauer, DO on

    Dr. Abella,

    Certainly, your expertise in cardiac resuscitation is noted and truly appreciated. I do concede that the momentum in resuscitation research has been to support the effectiveness of CPR. However, every coin has two sides. Physicians may reasonably come to different conclusions by reviewing the same data.

    Interestingly, my conclusions are based on clinical data, much of the same data you have cited. The conclusions of these studies usually seem to be positive, irrespective of methodological questions and the plausibility of alternative conclusions. We all must review such data with a critical eye. Most of this evidence is at best, Level B, as performing blinded, randomized, controlled trials of CPR is nearly impossible. In Christenson, what about cause and effect? Are more compressions really the reason for better survival? Perhaps. compressions were done on patients responding better to other interventions, causing the providers to continue? Those in the 0%-20% compression time group didn’t do as well. Not a surprise. When you’re not viable, we stop doing CPR. Noting other plausible reasons for the differences identified is valuable.

    Did I misinterpret Sayre’s data or just see it differently? Compression-only CPR revealed similar survival rates to traditional CPR. Maybe that means ventilation isn’t that important. To me, it questions the effectiveness of CPR altogether.

    Thank you. Your work, and discussions like this, help us all to better apply the literature that we review.

Leave A Reply