Friday, September 23, 2011

Hyperbaric Oxygen Therapy for Carbon Monoxide

Another mini-review where I agree with the Cochrane Review that essentially concludes: too much cost/risk, not enough proven benefit.


Hyperbaric oxygen therapy is of proven value in rapidly clearing carboxyhemoglobin from the serum - half-life approaches 20 minutes at 3 atmospheres vs. 1 hour on 100% face mask and several hours at lower concentrations of oxygen.  In addition, HBO has all sorts of beneficial effects in terms of preventing the damaging intracellular effects of carbon monoxide, including impairment of cytochrome oxidase a3 and of lipid peroxidation.  Lipid peroxidation, as you might imagine, in a brain full of lipids, is where you really end up in trouble with permanent neurologic sequelae.


Unfortunately, the first animal models that prove how well HBO works go directly from CO poisoning into multiple atmospheres of therapy.  As few HBO centers there are in the U.S., this is absolutely not a clinically relevant model because of the delays to therapy - and that's why the human literature is less conclusive.


There are essentially three large studies that contribute most of the weight to the 2011 Cochrane Review - one from 1989 that demonstrates no benefit, a second from 2002 in the New England Journal of Medicine that has a broad following, and, finally, a 2011 publication in Intensive Care Medicine.  Most toxicologists who are pro-HBO base their opinions on the 2002 Weaver article.


The good news about the Weaver article - it's a great study.  It enrolls a lot of patients, it enrolls all kinds of patients, it dives them to three atmospheres, it does three dives in a 24 hour period, and it has excellent objective testing and subjective evaluation follow-up.  This study was well-designed to give us the answers.  And, in the end, it finds a significant difference in objective neurologic function in the HBO group and a subjective difference in memory ability in the HBO group.  What is odd, however, is that there were many objective tests of cognitive function.  Most trended towards favoring the HBO group, but only one of them reached statistical significance - a trail marking test in which a line was drawn between similar numbers.  The absolute change in cognitive function between treatment and follow-up wasn't that much different between the two groups.  And, unfortunately, the larger issue for me is the baseline difference between the two groups in amount of time exposed to CO which trended towards much higher in the NBO group - 22 +/- 64 hours in the NBO group and 13 +/- 41 hours in the HBO group.


This difference is much more important because animal studies have demonstrated it's not the carboxyhemoglobin concentration that matters as much as the dissolved CO that diffuses intracellularly.  There's a fabulous study in dogs demonstrating this difference.  In one group, they exposed dogs to CO to get their carboxyhemoglobin concentration to 68% - and they all died.  In a second group, they bled dogs until they had lost 68% of their hemoglobin - in theory, the same level of deoxygenation as the CO group - and they all lived.  Third, they bled dogs down until they had lost 68% of their hemoglobin, then exposed that hemoglobin to CO in vitro and re-transfused it back into the dogs - in theory, the same 68% carboxyhemoglobin level as group 1 - and those dogs lived.  The difference between group 1 and 3 was the amount of dissolved CO in the blood and intracelluarly, and it was demonstrated that their cytochrome oxidase a3 activity was normal in group 3 despite the carboxyhemoglobin levels.


So, that's where I can't take the Weaver evidence as strong enough as the sole large study favoring HBO, even though it was well-designed.  The 2011 evidence, as mentioned before, is a study by Annane in Intensive Care Medicine.  This is the more recent study showing no benefit.  However, if you thought the Weaver study had flaws, this one is even worse.  They enrolled patients between 1989 and 2000 - but didn't publish until 2011, which is a massive red flag.  Their endpoint is fuzzier, as they simply have a questionnaire and a physical examination as their follow-up without a lot of details.  But, for what it's worth, they found HBO was futile in non-comatose patients and harmful in comatose patients.  They also do not dive as low or as long as the patients in the Weaver study.


Each of these studies makes it in the Cochrane Review which finds a cumulative nonsignificant trend towards minimal improvement in the HBO group.   The problem is, HBO therapy is expensive, causes hyperoxic seizures, barotrauma, anxiety, oxidative stress and both hyperthermia and hypothermia.  Weak evidence for mild improvement in delayed neurologic sequelae at 6 weeks is not a strong enough motivator for me to be enthusiastic about subjecting someone to the risks of HBO therapy.  I'd love to see more data.


"Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning."
www.ncbi.nlm.nih.gov/pubmed/12362006


"Hyperbaric Therapy for Acute Domestic Carbon Monoxide Poisoning: Two Randomized Controlled Trials."
www.ncbi.nlm.nih.gov/pubmed/21125215


"Hyperbaric Oxygen for Carbon Monoxide Poisoning (Review)."
www.ncbi.nlm.nih.gov/pubmed/21491385

Thursday, September 22, 2011

We're Covered in Filth

This is not the first study showing physician white coats and nursing uniforms are colonized with bacteria, nor that may of those bacteria are pathogenic and multi-drug resistant.  In the past, this has been used as a call for the abolishment of physician white coats, ties, and all long-sleeved apparel.

This study, however, shows that short-sleeved nursing uniforms were just as likely to be coated with bacteria - 49% to 54%.  Interestingly, even "changing uniform daily" still resulted in colonization with pathogenic bacteria.  The authors speculate the main issues are that all textiles easily transmit bacteria, and that hand hygiene might be more critical than uniforms in prevent transmission from patients to physician clothing.

This study also, like the many before it, doesn't demonstrate anything but colonization - not documented patient-to-patient transmission via healthcare worker clothing or any specific outcome measures.  However, I am a believer that white coats are fomites and medical relics that should go the way of bloodletting and golden elixirs. The studies in support of white coats cite patient satisfaction and ease of identification of roles - which, while important, could be mitigated by new interventions for identification of healthcare providers.  Even though we yet have no evidence of harms from this colonization with pathogenic bacteria, it's essentially a zero-cost intervention to stop wearing white coats and ties - so even if the number needed to treat to prevent a transmissible infection is immense, it's a free way to protect our patients as best we can.

"Nursing and Physician Attire as Possible Source of Nosocomial Infections."
www.ncbi.nlm.nih.gov/pubmed/21864762

Tuesday, September 20, 2011

CMS ED Quality Measures Are Coming

When the government is the largest healthcare payor, you pay attention when they start measuring "quality" - because it's usually not long after that payments are tied to "quality".

This is an Annals study looking at the pilot reporting program from CMS, which includes seven metrics that hospitals are going to have report in the next two years.  These metrics are:
 - Throughput time for admitted patients.
 - Throughput time for discharged patients.
 - Admit order to bed placement for admitted patients.
 - Time to pain management for long bone fractures (discharged and admitted)
 - Time to chest x-ray after order placed (discharged and admitted)

I love the pain management metric.  The literature that supports how poorly ED physicians manage pain is extensive.  Not a day goes by where have a resident trying to give a 2 milligram dose of morphine to treat acute pain.  The other metrics - well, so, faster is more and more is better, so therefore more, faster is "quality", right?  Well, there are some studies that show improved outcomes when patients reach the floor more rapidly, so, perhaps that's what they're getting at.  I don't know if we'll ever really know how improving these measures affects quality, because there are so many other confounding variables affecting these issues.

What is good about these metrics is that it should make a lot of EDs re-examine their processes and see how they can maximize the resources they have.  But, after the low-hanging fruit, you run up against issues of physical and financial resources - many of which are hospital-wide issues.  It will be interesting to see how our workplace changes in response to this.  Unsurprisingly, academic centers and busy EDs had the worst throughput statistics; not sure about magical solutions there.

"A Field-Test of Time-Based Emergency Department Quality Measures."
www.ncbi.nlm.nih.gov/pubmed/21868129

Monday, September 19, 2011

How Do We Miss Aortic Dissection?

This is a retrospective look at 109 cases eventually diagnosed with aortic dissection - with a focus on the differentiating clinical features present in the 17 cases where the diagnosis was initially missed in the Emergency Department.

Confounding clinical attributes that were associated with a missed diagnosis were walk-in vs. ambulance arrival and presence of anterior chest pain.  Chest x-rays lacking a wide mediastinum were nonsignificantly associated with missed aortic dissection, and with only 55% of their diagnosis cohort having a wide mediastinum vs. 25% of their misdiagnosis cohort.  Interestingly, over 75% of each group received a d-Dimer and they were all positive in the misdiagnosis group as well as all but one in the diagnosis group.  It would seem that they order so many d-Dimers that they've become fatigued to its clinical usefulness due to its poor specificity.

The good news is the patients who were initially misdiagnosed had similar mortality (18% vs. 15%) - despite 7 of the 17 being treated with antithrombotic agents.  Most of the missed diagnoses were classified as undifferentiated possible ischemic chest pain, but two were diagnosed with renal colic.

As always, the main problem with missed-diagnosis literature is there's no guarantee the authors didn't miss another set of cases themselves.

"Factors leading to failure to diagnose acute aortic dissection in the emergency room."
www.ncbi.nlm.nih.gov/pubmed/21889877