ICU Rounds Report - Dec 16th, 2010

Auto-Anticoagulation of Cirrhosis? Patients with liver failure commonly present with elevated INRs and low platelet counts. Liver dysfunction alters synthesis and production of both the pro- and anti-coagulants produced by the liver. In addition, numerous other effects alter normal coagulation in the cirrhotic patient: low platelets, platelet dysfunction, altered endothelial function (including elevated levels of prothrombotic vWF and tFVIII), elevated levels of nitric oxide, etc. Further confusing things: liver failure can dramatically increase rates of fibrinolysis. What does it all mean?

In other words: Does a liver failure patient with an INR of 2.3 needs DVT prophylaxis?

It depends on who you ask. The recent medical literature suggests that cirrhosis offers little, or no protection.  Now, in the first study of its kind, researchers at the University of Missouri looked at the relationship between pathologic INR elevations and DVT. They retrospectively evaluated 190 hospitalized patients with varying degrees of chronic liver disease and found rates of DVT of 6.3% overall (by comparison, rates of DVT in high risk surgical inpatients without prophylaxis is between 4 - 8%).  Interestingly, higher elevations in INR were associated with greater DVT risk. Lastly, the study found that receiving DVT prophylaxis didn't protect against DVT, although the study was underpowered to detect a statistical difference and only 25% of liver failure patients received some kind of prophylaxis. Sadly, there is no data available for surgical patients.

What everyone can agree on is that current panel of test used to diagnosis coagulation abnormalities in liver failure (plt count, PT, aPTT) are worthless because the balance of pro/anticoagulant factors is not considered by these tests (which tend to measure simple deficiencies of anticoagulants). Abnormal hemostatic tests have never been shown to correlate to bleeding tendencies in liver failure. Further, it is clear from liver transplantation surgery that preemptive "correction" of the abnormal values does not reduce, and probably promotes, bleeding. With this shift in thinking, many centers are now reporting the ability to keep liver transplantation "bloodless" (no transfusions needed) greater than half the time. To accurately diagnosis whether a liver failure patient is pro- anti- or euthrombotic requires much more sensitive testing, such as TEG,  which is not widely available.

So we're left where we started, which is were every good paper ends: more research is needed! Do liver patients needs SCDs/heparin? Anyone's guess. How can we even measure coagulation in liver failure? Not really possible. Does prophylaxis even work in the setting of liver failure? No one has a clue. Is working in an ICU awesome anyway? Yes, very! 
Dabbagh, O, et al. Coagulopathy does not protect against venous thromboembolism in hospitalized patients with chronic liver disease. Chest 2010; 137:1145.
Lisman, T, Porte, RJ. Rebalanced hemostasis in patients with liver disease: evidence and clinical consequences. Blood 2010; 116:878.
Gulley D et Al.  Deep vein thrombosis and pulmonary embolism in cirrhosis patients. Dig Dis Sci . 2008; 53( 11): 3012- 3017.
Northup PG, et al. Coagulopathy does not fully protect hospitalized cirrhosis patients from peripheral venous thromboembolism. Am J Gastroenterol 2006 ;101(7):1524-1528.

Gotta have more? Old Issues HERE.

ICU Rounds Report - Dec 9th, 2010

Creatinine, an oldie but a goodie? Creatinine is a breakdown product of creatine kinase in skeletal muscle. First found in blood in 1896, it was noticed that levels go way up with chronic kidney disease in the 1910's. We've been fixated on it every since as the biomarker for kidney function. It's been routinely measured on hospital inpatients for at least 60 years.  It's worth reviewing it strengths and weaknesses.

In the strength column, we can place that we have lots of experience with it. It's widely available, cheap to measure and tracks chronic disease quite nicely. Downsides? First, I've never liked saying it. It's a funny word. Second, the production of creatinine is highly variable, dependent on multiple factors including age, sex, race, muscle mass, nutritional state and others. Second, up to 40% of urinary creatinine is secreted by the tubules not filtered. In other words, it will never directly correlate with GFR as well as something that is not secreted at all, such as cysatinC (CysC). This leads to delays in the diagnosis of acute drops in GFR. Which brings us to the third major problem: It's slow. When the kidney filtering slows or stops, it takes days the serum creatinine to rise because basal production rates are so slow. Currently, we have to wait 2 days to learn the kidneys have been injured, beyond the time when interventions are thought to possibly improve outcomes. Its like trying to diagnosis an acute MI without troponins and just waiting to see how it will all settle out.

These limitations have lead to the intense development of serum and urinary biomarkers that better reflect GFR, give real time information about kidney injury and will, in all likelihood, make our old friend creatinine obsolete for ICU use with in 10 years . We've got several great candidates which work quicker, correlate better with disease severity and give more prognostic information coming down the pipeline, some of which are already being used in POC testing around the world (NGAL, KIM-1, L-FAB).
Rabinowitch IM. The prognostic value of the study of the blood creatinine in nephritis: based upon the study of fourteen cases with complete postmortem examination. Can Med Assoc J 1921; 11:320–322.
Moore, E, et al. Novel biomarkers of acute kidney injury: ready for clinical application? Current Opinion in Critical Care. Dec 2010 16(6): 523–525


Ultrasound for ICP. Occasionally, (meningitis, trauma, liver failure, strokes) we wonder what the pressure inside the head is. CTs of course, let us peek inside but only allow inferences. Sometimes, we wonder so much about ICP that we have the brain surgeons put an invasive monitor thru the skull. That's nice because it can also be therapeutic (in the case of EVDs). But is there a quicker way? Yes! Does it work? Maybe!

In 2010, when someone asks the question, 'is there a quicker way' the correct answer is either Epic or Ultrasound. Ultrasound? Sure. Recall that the optic nerve is sheathed within the dura. It appears that elevated pressures allow CSF buildup in this semi-compliant sheath and widen the diameter. (That's also the basis for blathering on about papilloedema). Multiple studies have measured the diameter of the optic sheath with ultrasound (at a standardized location, 3mm past the globe) and found a nice concordance with ICP. On study found a diameter greater than 5 (normal=3mm) was strongly predictive of an of ICP >20 mm Hg (sensitivity and specificity 94% and 76%, respectively).

Further work with MRI looking at the optic nerve sheath has shown that it is even more sensitive. With MRI, a cut-off value of 5.30 mm yields a sensitivity of 100% for diagnosis an ICP greater than 20.
Soldatos T. Optic nerve sonography: a new window for the non-invasive evaluation of intracranial pressure in brain injury. Emerg Med J  2009;26:630-634
Geeraerts T, et al. Use of T2-weighted magnetic resonance imaging of the optic nerve sheath to detect raised intracranial pressure. Crit Care 2008;12(5):R114. Epub 2008 Sep 11.

ICU Rounds Report - Dec 7th, 2010

It's so cold out there! Survival after out of hospital cardiac arrest with a good outcome is a rare event. Despite advances across other fronts in medicine, survival with an intact brain hadn't improved much in the last 50 years until someone tried post-arrest cooling. Previously, post-resuscitation efforts have been entirely supportive. Now, for comatose patients after vfib arrest, cooling the body to 34 degrees for 24 hours can nearly double the chance of recovery. Currently at UVa, we're cooling 3-4 patients a month (last month we did 7!).

But how do we cool patients? Everything has been tried. From ice bags, cold saline, irrigating bladder/stomach with cold fluids to cardiopulmonary bypass. All present technical and clinical challenges, not the least of which is that body vigorously defends its tightly regulated core temperature by increasing metabolic rate and shivering, usually requiring paralysis. Surface cooling is cumbersome and ineffective. Intravascular cooling (which we use) is expensive and risky. What if there was another way?

Researchers in toasty warm LA have published a trial using cannabanoid agonists in rats after resuscitation in vfib arrest. The results are very good news if you are a rat in a coma after vfib arrest: the drug appears to safely lower body temperature to ~34 degrees with no other major noted side effects. More importantly, the neurologic and functional outcome was significantly improved relative to controls.  The idea may represent a new therapeutic class for a common and lethal problem for thousands of Americans.

Finally, a quote from recent editorial on the new findings by Samuel Tisherman, "From the campaign 20 yrs ago (showing a frying pan representing marijuana and an egg representing your brain), to current controversies regarding legal medical uses, marijuana has a long history of capturing our interest. Wouldn't it be cool if a derivative actually improves brain function rather than frying it?"
Sun S, Tang W, Song F, et al: Pharmacologically induced hypothermia with cannabinoid receptor agonist WIN55, 212-2 after cardiopulmonary resuscitation. Crit Care Med 2010; 38:2282–2286
Bernard SA, Gray TW, Buist MD, et al: Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002; 346:557–563

ICU Rounds Report - November 24th, 2010

Metformin Acidosis - The ICU Yeti? You've probably heard someone call oxygen toxicity the sasquatch of the ICU - frequently talked about but never seen. Speaking of mythical creatures - how many of you have actually seen acidosis from metformin? In 1950, metformin's cousin phenformin was introduced and was widely used as an oral hypoglycemic. Twenty-five years and over 300 cases of frequently-fatal metabolic acidosis later, the FDA removed the drug. When metformin (Glucophage), a chemically similar biguanide drug, was introduced in in 1995 there was wide spread concern about metabolic acidosis. Merck, the maker, and the government sponsored multiple large trials which, in aggregate, evaluated over 36,000 patient-years without a single incidence of acidosis due to metformin.

By contrast, some case reports have shown up from time to time. The FDA reports that among the first one million treated with metformin, 47 cases (20 fatal) were reported. Forty three of these occurred in the setting of acute renal failure and the rest occurred during sepsis, overdose or heart failure. In other words, when used according to the label, there is zero chance (or close enough to zero as to be clinically irrelevant) of the drug causing acidosis.
Salpeter S, Greyber E, Pasternak G, Salpeter E: Risk of fatal and non-fatal lactic acidosis with metformin in type 2 diabetes. Cochrane Database Syst Rev 2:CD002967, 2003
Stades AME, Heikens JT, Erkelens DW, Holleman F, Hoekstra JBL: Metformin and lactic acidosis: cause or coincidence? A review of case reports. J Int Med 255:179–187, 2004

Neurogenic Pulmonary Edema.  After major head injury (status, trauma, hemorrhage) the aveloli and interstitial spaces in the lungs will occasionally fill with fluid causing patients to cough blood, breath quickly and desaturate. On autopsy studies, up to half of patients with those diseases will have evidence of it. We're not quite sure how it happens, but based on animal studies it is likely related to large, sudden increases in sympathetic outflow which causes 1)pulmonary microvascular damage 2)intense pulmonary vasocontriction raising transcapillary pressures and/or 3) acute, transient RV/LV dysfunction. The effect can be blocked with phentolamine in animal models. No human trials have been done to see if alpha blockade is safe or reasonable.

Onset is typically quite rapid after the head insult. Diagnosis is one of exclusion. Treatment is supportive and should be directed at the cause and sequelea of the head injury. Like neurogenic heart dysfunction, the huge majority of cases are transient and improve with time making prognosis more dependent on the extent of the head injury. This disease is sometimes confused with ARDS and may look radiographically similar. The treatment and prognosis are very different, however. Unlike ARDS, high PEEP levels and permissive hypercapnia should be avoided because of their effect on ICP.
Baumann, A, Audibert, G, McDonnell, J, Mertes, PM. Neurogenic pulmonary edema. Acta Anaesthesiol Scand 2007; 51:447.

ICU Rounds Report - November 19th, 2010

From Lis
How bad is my hemorrhage, Doctor? Drs. Hunt, Hess and Fisher are often employed to help. Since the 1930's physicians have attempted to compile an adequate grading scale for sub-arachnoid hemorrhage. Ideally this scale would guide management decisions affected by severity, provide a working prognosis, allow for treatment research among the various grades, and monitor changes in severity in an individual patient.

Hunt and Hess (used at UVa):
The Hunt and Hess scale is a 1968 modification of an older scale (Botterell, 1956). The scale was designed to assess surgical risk and if patient is a surgical candidate when is the ideal time to operate. Today it is used to evaluate clinical outcome, which according to one study Hunt and Hess was the strongest predictor of clinical outcome at 6 months when compared to GCS/WFNS (Aulmann, 1998). The scale is built on three axes:
a. Intensity of meningeal inflammation reaction
b. Severity of neurological deficit
c. Level of arousal
d. Presence of significant associated systemic disease à increases grade by one level

The benefit of Hunt and Hess lie in its easy administration. Disadvantages include the low inter-rater reliability (κ = 0.42), vague terminology, and requiring the clinician to evaluate all three axes on a single scale. This forces us to choose which axis is most important and contributes to lax category boundaries and relatively arbitrary grading assignments. Finally, data indicates that there are only differences in outcome associated with some of the grades - specifically Grade 2 and 3 and Grade 3 and 4.

1  Asymptomatic or mild headache and slight nuchal rigidity
2  Moderate to severe headache, nuchal rigidity, no neurologic deficit except cranial nerve palsy
3  Drowsy or confused, mild focal neurological deficit
4  Stuporous, moderate or severe hemiparesis, possible early decerebrate rigidity and vegetative  disturbances
5  Coma, decerebrate rigidity, moribund appearance

Fisher:
The Fisher Scale is a 1980 scale designed to predict cerebral vasospasm based on the blood pattern seen on the first post bleed CT scan. The main advantage of the Fisher scale is the high level of inter- rater reliability (κ = 0.90). Since its development, clinicians have started using it as a clinical outcome predictor as well. However the newer scales are combining patient age, Hunt and Hess, and Fisher in order to improve the predictive strength of the scale. One example is the Olgilvy and Carter scale.

1 No blood detected
2 Diffuse deposition or thin layer with all vertical layers (in interhemispheric fissure, insular
cistern, ambient cistern) less than 1 mm thick.
3 Localized clot and/or vertical layers 1 mm or more in thickness
4 Intracerebral or intraventricular clot with diffuse or no subarachnoid blood
Aulmann C, Steudl WI, Feldmann U. [Validation of the prognostic accuracy of neurosurgical admission scales after rupture of
cerebral aneurysms]. Zentralbl Neurochir 1998;59:171–180.
Hunt W, Hess R. Surgical risk as related to time of intervention in the repair of intracranial aneurysms. J Neurosurg 1968;
28:14.
Rosen DS, Macdonald RL. Subarachnoid hemorrhage grading scales: a systematic review. Neurocrit Care. 2005;2(2):110-8.
Review.

STATs note: Kappa scores are generally thought to be a more robust measure than simple percent agreement calculation since κ takes into account that the agreement may be occurring by chance. Recall that if the raters are in complete agreement then κ = 1. If there is no agreement among the raters (other than what would be expected by chance) then κ ≤ 0.

κ	Interpretation
< 0	No agreement
0.0 — 0.20	Slight agreement
0.21 — 0.40	Fair agreement
0.41 — 0.60	Moderate agreement
0.61 — 0.80	Substantial agreement
0.81 — 1.00	Almost perfect agreement

-

ICU Rounds Report - November 18th 2010

Sticky Platelets. Back in the days when the only thing to do with an MI was to talk about it (i.e. the 60 and 70's), physicians noticed that heart attacks tended to happen most commonly in the morning. Same for strokes and TIAs. Why? Partly because platelets tend to have diurnal variation and aggregate more avidly in the AM, particularly when people first assume the upright position after lying recumbent all night. These events also to the corresponds to the typically highest blood pressures of the day.

The platelet part of this theory was confirmed in the 1980's in a series of studies that measured platelet function in healthy men around the clock. Reliably, platelets are more likely to aggregate between 6-10am. The implications of this are two: first, don't get out of bed unless you really have to. Second, if you're treating patients at risk for arterial thrombosis with anti-platelet agents, make sure you're covering the morning.
Tofler GH, Concurrent morning increase in platelet aggregability and the risk of myocardial infarction and sudden cardiac death. N Engl J Med. 1987 Jun 11;316(24):1514-8.

Teachable Moments. Patients undergoing major surgery or trauma are significantly more likely to pay attention to healthful suggestions, like to drink less or stop smoking. Multiple randomized trials of smoking cessation prior to major surgery have shown big benefits, particularly with cardiac and wound complications. Other trials have shown that repeatedly reminding patients to stop smoking and drink less makes an real, measurable impact. This is especially true at stressful times (i.e. new cancer diagnosis, major surgery, trauma).

Alcohol abuse is common problem in our patients and of course a significant risk factor for trauma. Up to 50% of trauma patients will screen positive to alcohol abuse. To decrease the trauma burden, researcher at Harborview Medical Center randomized over 700 trauma patients who screened positive for abuse symptoms to a brief (30 minute) intervention from an addiction counselor or routine care. Patients were told they were in a study regarding trauma outcomes and were not aware that the addiction intervention was part of the study.  The results were striking.

At one year, alcohol use was reduced (self-reported) by 22 drinks per week (P<0.03). There was a 47% reduction in injuries requiring either emergency department or trauma center admission (hazard ratio 0.53, 95% confidence interval 0.26 to 1.07, p = 0.07) and a 48% reduction in injuries requiring hospital admission (3 years follow-up).

Given the risks, benefits and costs of this intervention, it seems like alcohol counseling should be routinely given to at risk trauma patients. While that happens from an institutional stand point, it seems reasonable for all of us to take some time and address alcohol and smoking in our patients in these "teachable moments."
Warner, DO. Surgery as a Teachable Moment Lost Opportunities to Improve Public Health Arch Surg. 2009;144(12):1106-1107.
Shi, Y et Al. Surgery as a Teachable Moment for Smoking Cessation. Anesthesiology.  2010 112(1): 102-107

ICU Rounds Report - November 17th 2010

The TRACS trial is here! More than 10 years after the publication of the TRICC trial, we're finally getting some more proper, randomized data about blood transfusion in critical care. JAMA last month published the long awaited results of a single center RCT in Brazil involving over 500 patients randomized to liberal (transfuse to >30% hematocrit) vs conservative (>24%) transfusion strategy in patients undergoing routine CABG and/or valve surgery requiring bypass. The intervention significantly reduced PRBCs transfusion (78% and 47% receiving, respectively). Morbidity and mortality was the same in both groups.

Unlike our practice at UVa, all the blood given was young (median age 3 days) and not leukodepleted. Older blood, as is frequently transfused here, has been associated with worse outcome while leukodepletion (standard here) appears to improve outcomes.

In the same issue of JAMA, researches using the STS database found transfusion practices among 700 US hospitals for 100,000 patients undergoing CBP showed wide variance in the rates of product transfusion (RBC (7.8%-92.8%), plasma (0%-97.5%), and platelet (0.4%-90.4%)). These differences persisted after adjusting for hospital and patients factors. Like the TRACS trial, there was no apparent difference in mortality or morbidity.

Taken together, these results point to a continued epidemic of costly, unnecessary and potentially dangerous over-utilization of blood products. Previously, this was, in part, due to a lack of RCT-based evidence in this patient group. Now, this is no longer true. We can say with assurance that transfusion of even a single PRBC unit has real risks that must be carefully balanced against benefits which, while real, are increasingly harder to find.
Hajjar LA, Vincent J-L, Galas FRBG; et al. Transfusion requirements after cardiac surgery: the TRACS randomized controlled trial. JAMA. 2010;304(14):1559-1567
Bennett-Guerrero E, Zhao Y, O'Brien SM; et al. Variation in use of blood transfusion in coronary artery bypass graft surgery. JAMA. 2010;304(14):1568-1575.

A FAST HUG Every 5.7 weeks on rounds, someone mentions the mnemonic developed by critical care god Jean Louis Vincent, FAST HUGS. Here it is written down so you have a nice reference to check in case you fancy yet forget it. You can add (although Jean Louis doesn't think you should) a BID, to emphasize this should be done at least twice a day.

F Feeding
A Analgesia
S Sedation
T Thromboembolic prophylaxis

H Head of bed elevation
U Ulcer (stress) prophylaxis
G Glycemic control
S Spontaneous breathing trial

B Bowel regimen
I Indwelling catheter removal
D De-escalation of antibiotics
Vincent JL: Give your patient a FAST HUG (at least) once a day. Crit Care Med 2005; 33:1225–1229