Blog with interesting cases and/or problems related to anesthesia with discussion based on best evidence in the literature.

March 14, 2012

Cardiopulmonary Arrest-Teasing out a cause

A 47 YO F undergoes uncomplicated hysterectomy for fibroids, pain, and bleeding.  It is performed with a combined general epidural anesthetic.  A lumbar epidural is dosed towards the end of the case with 0.1% ropivacaine + 0.25 mcg/cc of sufenta for a total of 10 mLs.  The pt is found to be pain free in the pacu.

In the PACU, the patient is lucid, conversant, can wiggle toes, and not in pain.  She is discharged 45 min later to the floor.  Upon arrival to the floor, an initial BP is 85/50 mmHg with a HR of 60 bpm.  Saturations are normal.  30 min later, her BP has dropped to 79/45 mmHg with a HR of 59 bpm.  The RN notifies the anesthesiologist who did the case who orders her to d/c the epidural infusion (running at 12 cc/hr).  The RN complies, but about 50 minutes later is notified by a family member that the patient is unresponsive.  A code is called when the RN cannot detect a pulse.  CPR is begun, the initial rythm documented is PEA, and 1 mg of epinephrine is given.  The next blood pressure is 185/105 mmHg with tachycardia.  The patient is intubated within 5 minutes and saturations return to normal.  The initial blood gas drawn a few minutes after intubation shows a pH of 7.09, PaCO2 of 72 mmHg, PaO2 of 103 mmHg and HCO3 of 22.

Although, ROSC occurrs within 5 minutes of calling the code, the patient is unreponseive with minimal relfexes (pupillary and gag reflex are intact).   The patient remains comatose for 48 hours, after which she "wakes up" but with significant cognitive deficits requiring a transfer to rehab for intensive rehabilitation.  Work up done after the code inluded echocardiograms that note significant global hypokinesia with an EF of 30% improving to 35% over the course of a couple of weeks.  Cardiac enzymes drawn show a very mild bump, indicating some myocardial damage, however, an angiogram done two weeks post arrest fails to find any native coronary artery disease.  The diagnosis of takotsubo cardiomyopathy is rendered.

Tracking this patient's blood pressure from her sign in to her code, there is a clear progression to lower and lower blood pressures.  Her initial blood pressure upon admission is 150/80 mmHg.  When she arrives in the PACU, it is 115/50 mmHg and is checked q 15 min.  Her HR also decreases from a baseline of 80 bpm to a nadir of 54 bpm in the PACU.  Her blood pressure upon discharge from PACU is just under 100 mmHg systolic.  Neuraxial aneshesia produces this exact picture.  As a block  gradually develops an expected drop in blood pressure would occur.  If the block were below T5 or T6, the cardioaccelerator nerves will be free from blockade, and the heart rate will adjust upwards do to the baroreceptors increasing their stimulation to the brainstem sending messages to increase the HR to compensate for decreased blood pressure.  However, if the neuraxial blockade rises above these nerves, the heart rate decreases.  The most common clinical situation where this is seen is in spinals given to parturients in preparation for caesarian section.  Within 5 to 10 minutes after injection of IT bupivacaine, a dramatic decrease in blood pressure occurs with a concommitant decrease in HR.   In this particular case, it was noted that the epidural catheter had tested negative for IT placement by a test dose of lidocaine.  The catheter was removed the following day after negative aspiration for CSF. While a negative aspiration test does reduce the chances that the catheter was in the IT space, it certainly does not rule out this possibility.
The first lab drawn after the arrest was an ABG.  Her ABG demonstrated a severe respiratory acidosis. The rate of rise of CO2 during apnea has found a range of values.  In a general sense, during anesthesia, a good estimate is that CO2 will rise 6 mmHg during the first minute of apnea, and then 3 to 4 mmHg each minute thereafter.  In a study on health volunteers under deep GA, the ETT was clamped to reduce CO2 washout with ambient oxygen.  In this study, they showed that CO2 increased 12 mmHg after one minute and 3.4 mmHg each minute thereafter.  Presumably, GA will decrease overall metabolism, and therefore, in an awake patient who collapses into respiratory arrest, the CO2 will likely ramp up more quickly.  In this case, assuming 12 mmHg rise in the first minute, and then 3 mmHg/min after, it would estimate that she was apnea for about (33 mmHg- 12 mmHg=21 mmHg and then divide 21 by 3 equals 7).  Therefore, This indicates that she was likely completely apiece for at least 6 to 8 minutes.  A study of the predictive ability of the acid base status [6] found that if the arrest lasted longer than 10 minutes the average pH was 7.1 with a PaCO2 of 61 mmHg and Base Excess of -10.  The survival rate in this group was only 20% compared to 63% in a group where initial pH was7.39 PaCO2 of 37.6 mmHg and Base Excess of -3.2.  Clearly, the above situation represents an initial blood gas that should portend a very bad outcome and suggests that either the patient was down greater than 10 minutes prior to the start of CPR, or that the CPR provided was utterly ineffectual.  In the same study cited above, Carrasco et al. [6] noted that in their subjects a minority had either normal acid base status or were alkalotic, whereas the majority had either a metabolic acidosis or a combined acidosis.  They noted that the worst prognosis was associated with those who had a combined respiratory metabolic acidosis on the initial ABG (survivaly rate 11% vs. 63% for minimal or soley metabolic acidosis). In particular, this group showed that a severe respiratory acidosis was particularly predictive of a poor outcome.  A possible explanation of this relates to the rapid diffusibility of CO2 into cardiac myocytes, allowing more rapid and significant depression and maintenance of intracellular acidosis when the source of the acidosis is respiratory.  The above patient's severe hyercarbia may cause one to speculate that the patient had an inciting respiratory event given the mild degree of the metabolic acidosis.
 To properly understand the source of a metabolic acidosis, it is important to calculate the anion gap.  This gap is predominantly made up of albumin (negative charge) in healthy patients, so this must be taken into account in critically ill patients.    In our patient, the AGAP was calculated at 11.9 (16 if you add in K+).  The upper limits of normal is 12.  However, her albumin,  measured at the same time was 2.9 (normal is at least 3.4 to 5.0 g/dL).  The AGAP is reduced 2.5 for 1 g/dL that the albumin is below normal (so 3.4-2.9=0.6 g/dL and 0.6 g * 2.5=1.5 which is added to the AGAP of 11.9=13.4.  The key point, is that this patient had a significant respiratory acidosis combined with a mild GAP acidosis.  The obvious missing anion is lactate generated from hypoxia at the tissue level.  Unfortunatley, we do not know the patient's PaCO2 just prior to arrest.  However, in normal volunteers  holding their breaths, this level of PaCO2 would take just under 5 minutes to achieve.   Another important aspect of analyzing acid base status is to recognize the importance of the delta ratio.  This is simply analyzing whether the AGAP measured (assuming it is elevated) can be accounted for by a porportionate decrease in the HCO3-.  If this is not the case, then you may have either a combined GAP and non GAP acidosis or the patient has a GAP acidosis mixed with a metabolic alkalsosis (see aspirin overdose).  Pleaes review this important aspect of acid base physiology here.  Of course, Hgb is an effective buffer in acute respiratory acidosis, where CO2 binds to Hgb>>HHgb+ HCO3-. Therefore, acute respiratory acidosis will result in an increase in the HCO3- of 1 mEq/L for 10 mmHg increase in PaCO2.  In our case we would  expect HCO3- to be increased by 3 mEq/L (it was acutally measured at 18.9, so with out the provision of acid from the extra CO2, it would have measured only 15.9 mEq/L, a significant decrease from the normal of 24 mEq).  In this case, the patient was given an amp of bicarb shortly after ROSC presumably because of her severe acidosis.  Bicarbonate therapy for severe acidosis is considered somewhat controversial, but used quite often in cardiopulmonary arrest.  Nevertheless, in 1985, Graf et al. reported on the negative effects of bicarbonate therapy in lactic acidosis [1], where they showed that blood pressure and cardiac output decreased when bicarb was used to treat lactic acidosis in dogs. In 1990, Kette et al. was able to demonstrate that in cardiac failure bicarb treament resulted in worsening acidosis intracellularly despite an improvement of the extracelllular (arterial) acid base status and decreased the chance of successful rescusitation [2].   In 1998 Levy published a review article looking at more than 30 studies in animals to evaluate the efficacy of sodium bicarbonate in resuscitation [3]. Where survival was an outcome, 4 showed benefit, however, 7 others showed no benefit. In this same review, myocardial performance was found to be worsened by sodium bicarbonate in 12 studies, while none of the studies found benefit. Human studies have also looked to find a benefit with sodium bicarbonate with little success.  In fact, studies in humans have found that sodium bicarbonate produces hypernatremia, hyperosmolality, and metabolic alkalosis in many patients all of which are associated with worse outcome.  In humans, there is only 1 prospective randomized trial looking at the efficacy of sodium bicarbonate in patients after cardiac arrest where it was not able to improve ROSC or survival [4]. In neonates, another RCT was unable to find any benefit of sodium bicarbonate in treating acidosis following asphyxiation which required post delivery assisted ventilation. In general, although it is well known that acidosis decreases myocardial contractility and increases vasodilation, sodium bicarbonate does not reliably reverse this in studies, may cause a paradoxical intracellular acidosis, and is not recommended in the ACLS guidelines for pulmonary arrest (with hypercarbia), for short duration arrest, and is only suggested as potentially helpful in patients who have a significant lactic acidosis after a prolonged cardiac arrest. In the end, the best maneuver to improve an acidotic patient is to ensure adequate ventilation, volume status, and organ blood flow.

The diagnosis of Takotsubo cardiomyopathy in this case is interesting and important in that there were many charactersitics of this case that do not fit with the "typical" presentation of Takotsubo cardiomyopathy and, if this is Takotsubo cardiomyopathy, treatment options are unorthodox. Takotsubo cardiomyopathy was only recently described in Japan.  The name Tako-tsubo means octopus trap as the shape of the heart in these patients appears similar to these traps used in Japan. This diagnosis has only recently gained notariety in the US. with media reports of the so called "broken heart syndrome" diagnosed in patients who suffered significant cardiac events mimicking acute coronary syndrome (ACS), but not having any coronary disease when taken to the cath lab.  These patients were noted to have had a significant mental stressor that preceded the medical event, typically the loss of a loved one, thus the name "broken heart syndrome". The characterstic description of this syndrome on echo is one of apical ballooning with basal hyperkinesis.  Debate exists as to the etiology of this apical hypokinesis; however, the evidence seems to suggest that it is most likely a result of catecholamine toxicity resulting in a significant but reversible cardiomyopathy. Other important characteristics of this syndrome that aid in diagnosis are that it occurs primarily in post menopausal women (70%), EKG shows ST changes that begin with ST elevation in the anterior septal leads, followed by deep T wave inversions, and finally normalization over a period of several hours.  Finally, biomarkers, such as troponins, will elevate, but not to the same degree as in an acute MI.  This represents stunning of the myocardium, but not necrosis.  It is important to remember that this syndrome is one of exclusion: exludue MI from coranary plaques or spasm, myocarditis, aortic dissection, pericarditis, pheochromocytoma etc. Clinically, these patients often present with chest pain and shortness of breath such that clincians often beleive they are dealing with an acute MI. The above patient didn't have some of the characteristics of this syndrom: There was no acute changes in the EKG, the patient was not post menopausal, did not have a clear mental or psychological stressor to speak of, and finally, her cardiomyopathy did not return to normal within a short period of time.  Nor did she have chest pain.  However, the patient did have surgery, is female, and did have a mild increase in her troponins.

In treating someone with an acute onset of chest pain and SOB, who has EKG changes, it is common to give beta blockers.  This could prove dangerous if the patient is having an acute attack from Takotsubo cardiomyopathy.  Since the patient would presumably have elevated levels of catecholoamines, beta blockade would theoretically inhibit cardiac contractility, while leaving unopposed alpha adrenergic receptor  mediated arterial vasoconstriction.  This is mimicked clinically in the OR if the surgeon injects epinephrine containing LA into the field (i.e. ENT) which can lead to acute tachycardia with  hypertension.  If the anesthesia  provider is not astute, they may attempt to lower the heart rate and blood pressure with a pure beta adrenergic antagonist. This maneuver will leave epinephrine to bind exclusively to the alpha adrenergic receptors causing severe arterial vasoconstriction in the face of myocardial depression.  In some cases this combination can result in cardiac collapse. Therefore, if Takotsubo cardiomyopathy is suspected, phenylephrine can be used as a temporizing measure to treat low blood pressure, however, immediate placement of an IABP should be considered if the patient does not respond to initial measures as catecholamine infusions and betablockers are relatively contraindicated.

During cardiopulmonary arrest, down time is critical.  In particular, neuronal damage occurs rapidly.  In the Brain Resuscitation Research Trial [5], when the cardiopulmonary arrest lasted more than 6 minutes, there was a 0% chance of neurological recovery.  This was also true if CPR lasted longer than 15 minutes in a prehospital event.  Once neurons are deprived of oxygen, a cascade of events begins which untimately leads to loss of cellular integrity and cellular necrosis.  Brain injury after an arrest is caused by several pathways including excitotoxicity (EAA), disrupted calcium homeostasis, free radical formation, pathological protease cascades, as well as activation of cell death signaling pathways occurring from hours to days after ROSC. While there have been a multitude of interventions aimed at the above pathways, little has born fruit in reducing neuronal injury.  However, several items are known to directly worsen neuronal recovery after an ischemic event: low MAP during the first 2 hours after ROSC, pyrexia, hyperglycemia,  and seizures. Early signs that neuronal damage has occurred after ROSC include seizures, myoclonus and varying degrees of coma.  In this case, therapeutic hypothermia should be considered after ROSC. Furthermore, hyperglycemia is common in this patient population and should be treated aggressively. 

Therapuetic hypothermia:  The above case represents a patient who despite a ROSC in 5 minutes, remained comatose for more than 48hrs.  If the initial pH of 7.09 with a severe repiratory acidosis didn't provide a clue that this patient was likely to have a neurological insult, then her comatose state with onset of seizure acitivity would have made it very clear.  Two RCTs have been published showing an improvement in outcome when hypothermia was induced within minutes to hours in patients suffering VF out-of-hospital cardiac arrest who remained comatose after ROSC [7,8].  Patients in these studies were cooled to between 32 to 34 degrees C for 12 to 24 hours.  Other observational studies have found a similar benefit.  It has been proposed that while hypothermia does decrease CMRO2 by 6% for each 1C decrease in body temperature, this is only a minor portion of its therapuetic effect.  It is proposed that by decreasing harmful chemical reactions in the period of brain reperfusion, hypothermia prevents further neuronal cell death.  In particular, hypothermia is believed to decrease free radical production, excitatory amino acid release, and calcium shifts that lead to mitochondrial damage and programmed cell death.  Currently, most experts recommend that hypothermia be induced in any patient who has had a cardiac arrest unless they are suffering from severe cardiogenic shock, are pregnant, have or are at high risk for arrythmias, or have a significant coagulopathy [9].  In the case above, the patient did have a significant cardiac abnormality, but her blood pressure responded nicely to levophed.  This makes her case somewhat equivocal in terms of safety of inducing hypothermia.

1.  Graf H, Leach W, Arieff AI.  Evidence for detrimental effect of bicarbonate therapy in hypoxic lactic acidosis.  Science 1985;227(4688): 754.
2.  Kette F, Weil MH, von Planta M, Gazmuri RJ, Rackow EC.  Buffer agents do not reverse intramyocardial acidosis during cardiac resuscitation. Circulation. 1990;81(5):1660.
3. Levy MM. An evidence-based evaluation of the use of sodium bicarbonate during CPR. Crit Care Clin. 1998;14(3);457.
4. Dybvik T, Strand T, Steen PA.  Buffer therapy during out-of-hospital cardiopulmonary resuscitation. Resuscitation. 1995; 29(2):89-95.
5.  Abramson, NS, Safar, P, Detre KM: neurologic recovery after cardiac arrest: Effect of duration of ischemia.  Crit Care Med 1985;13:930-931.
6.  Carrasco HA, Oletta JF. Evolution of the Acid Base status in Cardiac Arrest.  Calif. Med. 1973; 118:7-12.
7. Hypothernia after cardiac arrest study group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Eng J Med 2002; 346:549.
8. Bernard SA et al. Treatment of comatose survivors of out of hospital cardiac arrest with induced hypothermia.  N Eng J Med 2002; 346: 557.
9.  ILCOR Advisory Statement: Therapeutic hypothermia after Cardiac Arrest.  Circ; 2003; 108:118.

No comments: