Safe Use of Bicarbonate

Sodium bicarbonate (NaHCO3) is a medication often given during cardiopulmonary bypass (CPB) by perfusionists. I think this medication hurts patients more than most perfusionists realize.  Giving a lot of NaHCO3 just to restore the bicarbonate level and keep the pH in the normal range during CPB can increase the blood osmolarity to a very high level.  This is particularly true when mannitol is also used in the prime or given during CPB.  It is easy to estimate the blood osmolarity changes caused by the addition of NaHCO3 from point-of-care (POC) testing of whole blood electrolytes.  But if a lot of mannitol is given, that medication’s effect cannot be estimated and the blood must be sent to the lab for osmolarity measurement. Artificially correcting the bicarbonate balance with NaHCO3 can also mask under oxygenation as a result of inadequate capillary perfusion.

Inadequate cellular oxygenation (hypoxia) results in metabolic acidosis and HCO3 consumption.     The development of true metabolic acidosis during CPB is relatively rare. If metabolic acidosis develops on routine cases (hypoxia without hypoxemia) there is something wrong with the technique being used. I believe the solution to correcting acidosis during CPB is to increase the blood flow if practical, increase the sweep FiO2, change the CO2 level or change the CO2 strategy (alpha stat to pH stat or vice versa if hypothermia is being used).  In my experience, these seem to improve oxygenation at the cellular level and stop or reverse the development of a base deficit. However sometimes an acidosis may develop which is not the result of metabolic production of acid. The HCO3 level can unintentionally be reduced iatrogenically. This may occur on CPB in two ways that I know of.

Firstly, the infusion of HCO3-free (or HCO3 precursor-free) crystalloid or the entrainment of HCO3-free crystalloid irrigation into the pump will dilute the circulating HCO3 level. This will induce a dilutional acidosis which is not related to the adequacy of perfusion (1).  Infants and small children are particularly susceptible to this phenomenon. I usually only gave enough NaHCO3 during CPB to compensate for any bicarbonate free crystalloid irrigation sucked into the pump.

Secondly, blood bank red blood cells (RBCs), even when washed through autotransfusion equipment, carry a heavy lactic acid load (2).  When RBCs are infused into the CPB circuit, this acid will consume some HCO3 and cause an acidosis to develop.

NaHCO3 administration during CPB to compensate for all these problems may result in hyperosmolarity (> 300 mOsmols/L) which may cause renal failure (>320) and brain damage (>360) (3,4,5,6). Osmolarity is the number of osmoles of solute per liter of solution. A similar measurement is osmolality which is the number of osmoles of solute per kilogram of solvent.  The difference in value is miniscule when talking about blood, but I prefer to use osmolarity because blood is measured in liters rather than kilograms.

A balanced crystalloid solution containing acetate, gluconate or lactate can be used in adult primes and as supplemental fluids as these HCO3 precursors will convert to HCO3 within six minutes during CPB upon passing through the liver in adults (7,8).  In children, the efficiency of this conversion is much slower (in the 6 hour range in infants).  Consequently, waiting for this conversion is not practical in children. A crystalloid prime (i.e., Plasmalyte 148) with 25 mEq/L of NaHCO3 added will prevent dilutional acidosis at the initiation of CPB. But the pediatric prime osmolarity would increase from 294 (with just Plasmalyte 148) to 319 with NaHCO3 added; near the upper limit for renal function. Osmolarity (in the absence of mannitol) can be estimated from POC testing with this formula: (Na mEq/L X 2) + (glucose mg% /18 ) + 15 = estimated osmolarity.  When priming a CPB or ECMO pump I used a bicarbonate balanced (normal Na) crystalloid made up of 1 liter of balanced electrolyte solution like Plasmalyte or lactated Ringer’s solution mixed with one liter of ½ normal saline solution with 50 mEq/l of NaHCO3 added. This resulted in a crystalloid with a bicarbonate of 25 mEq/L, a sodium of about 135 mEq/L and an osmolarity of about 266 mosmoles/L. Adding mannitol to the prime will make the osmolarity unpredictably higher.

For metabolic acidosis, administer 1 mEq/L (combined pump and patient circulation volume) for each -1 mEq/L base deficit. Repeat as needed.  Alternate formula: (Desired HCO3 – Actual HCO3) x KG x 0.3 = NaHCO3.                 Each unit of banked RBC will require 5-10 mEq of NaHCO3 to neutralize the effect of the acid load in the RBCs. NaHCO3 should not be given until the base deficit is -4 or greater.  The entrainment of excessive irrigation into the CPB circuit will cause a dilutional acidosis requiring the administration of NaHCO3.  The excessive irrigation will need to be removed by diuresis or ultrafiltration (UF).  The amount of ultrafiltrate from irrigation can be estimated by the amount of NaHCO3 administered. For example, assuming a normal HCO3 level of 25 mEq/L, 400 mls of irrigation ultrafiltrate would require the administration of 10 mEq of NaHCO3 to maintain a normal HCO3 level and pH. On the other hand, the removal of fluid by UF which does not require the need for NaHCO3 supplementation to maintain normal HCO3 levels indicates that the fluid was removed from the patient’s own extracellular compartment.   Care should be taken to prevent the Na from increasing beyond 145 mEq/L due to NaHCO3 dosing. Small amounts of 0.45% NS w/ 50 mEq/L of NaHCO3 added (127 mEq/L Na w/ 254 mOsmols/L) may be administered to prevent hypernatremia followed by diuresis or UF to remove excess fluid volume

NaHCO3 administered too rapidly can form CO2 gas emboli and trigger a bubble alarm.     NaHCO3 given concurrently with calcium chloride in the prime may form calcium carbonate (chalk).        If Na is elevated too quickly with NaHCO3 in the prime or by its addition later, central poutine myelinolysis (aka osmotic demyelination syndrome) or other brain damage can occur, particularly in infants and patients with severe hyponatremia. This might be confused with post pump chorea.   NaHCO3 given immediately prior to weaning may result in systemic vasodilation, decreased cerebral blood flow and decreased cardiac function resulting in lower arterial pressure, lower NIRS values and low cardiac output after weaning.

Consider THAM (aka tris or tromethamine, 3.6 gm/100 mL = 30 mEq or 0.3 mol/L w/ 389 mOsmol/L.) for patients with elevated Na levels, chronic acidosis or patients with severe hyponatremia. THAM acetate dosage: ml of 0.3 mol/L = KG x Base Deficit (mEq/L) x 1.1; with a maximum dose of 500mg/Kg.

Over the past 40 years I have seen resuscitation patients with sodium cation concentrations well over 150 mEq/l as a result of excessive NaHCO3 administration during the active CPR phase.  Serum sodium values this high translate into excessively high blood osmolarity. Previous American Heart Association Advanced Cardiac Life Support (ACLS) guidelines recommended routine bicarbonate administration as part of the ACLS algorithm. But recent guidelines no longer recommend its use because there is little evidence that it is beneficial in a cardiac arrest situation. In fact, NaHCO3 is thought to contribute to a variety of problems: 1) It may reduce systemic vascular resistance resulting in hypotension. 2) It may cause RBC alkalosis that unfavorably shifts the oxyhemoglobin saturation curve. 3) It may cause hypernatremia and hyperosmolarity which results in organ damage; the kidneys and brain in particular. 4) It produces excess CO2, which may diffuse into myocardial and cerebral cells and paradoxically contribute to intracellular acidosis. 5) It can exacerbate central venous acidosis with excess CO2 and cause pulmonary hypertension. And 6), it may inactivate simultaneously administered catecholamines (9). I don’t know the cumulative effects these problems may have during CPB versus CPR, but caution is the better part of valor when alternative methods are available.

References Show/Hide