Physiology of Cardiopulmonary Bypass

Pressure & Flow

Modern CPB machines pump blood into the arterial system using a centrifugal pump. The flow of a centrifugal pump increases as its angular velocity (RPMs) increases. Pressures are therefore a function of the set flow rate and the resistance to flow, and can thus be modified by changing either the flow rates or the vascular resistance (ex. phenylephrine, vasopressin, phentolamine). That said, CPB machines cannot increase flows ad infinitum – 1) centrifugal pumps are afterload-sensitive and plateau as SVR increases 2) CPB flows are limited by venous return, as past a certain point collapse of veins limits VR, lowering CPB reservoirs, and potentiating entrainment of air and 3) depending on the size of the arterial cannula, excessive flow may require excessive line pressure, leading to tissue (blood) trauma

Average flows for a normal adult (assuming a hemoglobin of 8 mg/dL and a temperature of 30C, i.e. moderate hypothermia) might be 2.4 L/min/m2, although this will be adjusted based on age, hemoglobin, temperature, and depth of anesthesia.

Pressure vs. Flow

Fundamental to CPB is the delivery of oxygen to organs that need it. Because flow is a function of pressure gradient and vascular resistance, from a purely teleological standpoint, it would appear that maximizing blood flow (as opposed to maximizing pressure, which may or may not maximize flow, depending on vascular resistance) would be the most rational approach to CPB management. That said, despite over 50 years of experience and research, the ideal blood flow and/or perfusion pressure is not known. Nor is it known whether flow or pressure is more important

There are several problems with maximizing flows. First, higher flow rates are associated with significant trauma (to hematologic elements) and increase the inflammatory response to CPB. Second, increased flows carry an increased embolic burden. Third, an increase in total flows does not assure increased DO2 to organs of interest, because the CPB machine cannot alter regional flows. Thus, if CPB flow rates are increased by perfusing relatively unimportant tissue beds, the increase in trauma/inflammation cannot be justified and might be harmful. With regards to the latter, Slater et al studied regional flow rates in pigs on CPB, and found that regional flows to the brain, kidney, and pancreas all decreased when flows were increased from1.6 to 1.9 L/min/m2 [Slater JM et al. Ann Thorac Surg 72: 542, 2001]. Indeed, at these high levels, brain and kidney flows were at 65% and 55% of pre-CPB flow rates, respectively. This data is similar to Rogers et al.’s human data which showed that increasing CPB from 1.75 to 2.25 L/min/m2 has no effect on cerebral blood flow or cerebral metabolic rate consumption in humans [Rogers AT et al. J Thorac Cardiovasc Surg 103: 363, 1992], and consistent with Govier et al.’s data which showed that changes in flow rates from 1.0 to 2.0 L/min/m2 had no effect on CBF during hypothermic CPB in humans [Govier AV et al. Ann Thorac Surg 38: 592, 1984]

Arguments can be made in favor of optimizing pressure at the expense of flows – first, cerebral autoregulation maintains constant cerebral blood flow (CBF) from MAPs of 50-150 mm Hg (and possibly lower during hypothermic CPB). Hypertensive patients are thought to have a right-shifted autoregulation curve, and diabetics may have altered autoregulation, thus increased pressures (or flows) can potentially be justified in these patients. Second, and related, while the CPB machine can manipulate total flows, it cannot individually affect regional flows – this ability is left to the organs themselves. Under CPB, end organs have only one mechanism by which they can modulate regional flows – alterations in regional vascular resistance (RVR). The higher the pressure, the broader the range of regional flows available to each individual organ. Note that this concept has not been proven (neither have concepts in support of a flow-based approach to CPB) and is offered only as a possible explanation for the failure of the flow-based approach to prove itself clinically superior

High flow rates are also disadvantageous in that they increase suture line strain, increase bronchial flow to the lungs, and increase collateral flow to the heart (which diminishes the duration of effective cardioplegia)

Considerations against CPB Flow Rates Increased hematologic trauma Increased stress response Strain on suture lines Increased pulmonary shunting Accelerated washout of cardioplegia May not affect regional flow advantageously [Slater JM et al. Ann Thorac Surg 72: 542, 2001] No data to support it

“Low Flow” Cardiopulmonary Bypass

The major putative advantages of “low flow” CPB are reductions in hematologic stress and reduced embolic load. Note, however, that there are no prospective, randomized, controlled trials that adequately address the lower limit of acceptability

Alterations in Afterload

Decreased SVR is the predominant cause of hypotension following initiation of CPB (secondary to reduced blood viscosity, dilution of endogenous catecholamines in priming solution, and differences in pO2, pH, and electrolyte concentrations between the priming solution and native blood). As CPB progresses, SVR gradually increases, eventually to supranormal levels. This is presumably due to hypothermia (leading to vasoconstriction and catecholamine release), stress response (also leading to vasoconstriction and catecholamine release), and vessel closure (i.e. maldistribution of flow).

Microperfusion and Pulsatility

Clearly altered by temperature changes, edema, loss of pulsatility, RBC injury, emboli, and the inflammatory response. Some perfusionists will try to counteract these changes by using mannitol, vasodilators, hemodilution, microfiltration, and pulsatile perfusion techniques. Pulsatile flow is difficult to achieve in the aorta, as the arterial lines significantly dampen the oscillatory component of pulsatile pressure. In the absence of any convincing data to support its use, pulsatile pumps, despite being more “physiologic,” cannot be justified



During alpha-stat acid-base management, the ionization state of histidine is maintained by managing a standardized pH (measured at 37C). Alpha-stat pH management is not temperature-corrected – as the patient’s temperature falls, the partial pressure of CO2 decreases (and solubility increases), thus a hypothermic patient with a pH of 7.40 and a pCO2 of 40 (measured at 37C) will, in reality, have a lower pCO2 (because partial pressure of CO2 is lower), and this will manifest as a relative respiratory alkalosis coupled with decreased cerebral blood flow. During alpha-stat management you have no idea what the patient’s pCO2 is, your goal is to maintain a constant dissociation state of histidine


During pH-stat acid-base management, the patient’s pH is maintained at a constant level by managing pH at the patient’s temperature. pH-stat pH management is temperature-corrected. Compared to alpha-stat, pH stat (which aims for a pCO2 of 40 and pH of 7.40 at the patient’s actual temperature) leads to higher pCO2 (respiratory acidosis), and increased cerebral blood flow. Often CO2 is deliberately added to maintain a pCO2 of 40 mm Hg during hypothermia.

Alpha Stat vs. pH Stat

A study by Kiziltan et al., in which 52 patients were randomized to alpha-stat versus pH stat management, showed that pH stat management led to increased jugular venous oxygen concentrations, implying increased CBF [Kiziltan HT et al. Anesth Analg 96: 644, 2003; FREE Full-text at Anesthesia & Analgesia]. A study by Sakamoto et al., comparing pH stat to alpha stat during repair of cyanotic neonatal congenital heart disease, demonstrated that pH stat management led to less pulmonary collateral circulation as well as higher oxyhemoglobin and lower deoxyhemoglobin levels on cerebral near-infrared spectroscopy, suggesting greater cerebral oxygenation through improved oxygen delivery with pH stat [Sakamoto T et al. J Thorac Cardiovasc Surg 127:12, 2004]. A prior study by Murkin et al. comparing pH stat to alpha stat showed that during pH stat, CBF and CMRO2 become uncoupled (CBF is pressure-dependent), whereas during alpha-stat CBF is related to metabolic needs (CMRO2) and not to cerebral perfusion pressure [Murkin JM et al. Anesth Analg 66: 825, 1987]. The major concern with pH stat is the potential for increasing the cerebral embolic load


Temperature is related to both the delivery and consumption of oxygen and is thus critical to effective CPB. The Q10 for metabolic reactions in humans is ~ 2, meaning that a 10 C decrease in temperature slows the rate by 1/2. Hypothermia increases blood viscosity and shifts the oxygen-hemoglobin dissociation curve left, both of which decrease oxygen delivery/release

Effects of Temperature on DO2 and VO2 Hypothermia Increases Viscosity (decr. DO2) Hypothermia Decreases Oxygen Release (decr. DO2) Hypothermia Decreases Metabolic Needs (decr. VO2)

Initial attempts at temperature management focused on the relationship between temperature and metabolic requirements. Since that time, there has been a trend towards increasing CPB temperatures [Cook DJ. Anesth Analg 88: 1254, 1999; FREE Full-text at Anesthesia & Analgesia]. Cited advantages include lower pump time, decreased coagulopathy, and avoidance of overwarming. As DO2 is increased, the lower limit of flow/pressure may be higher (unproven), and anesthetic requirements may be increased (probably true). With regards to neuroprotection, Grigore et al. found no difference in neurocognitive outcomes when comparing hypothermic (29C) to normothermic (36C) CPB in 270 patients [Grigore AM et al. Anesthesiology 95: 1110, 2001; FREE Full-text at Anesthesiology]. Mora CT et al compared 28C to 35C in 138 patients and found an increased risk of immediate neurocognitive dysfunction in the normothermic group, but no differences by 4-6 weeks postoperatively [Mora CT et al. J Thorac Cardiovasc Surg 112: 514, 1996]. Importantly, at least from a conceptual (but not practical) standpoint, it was previously thought that the benefits of hypothermia were secondary to reduced metabolic needs, however now it is now thought that the possible benefits of hypothermia are due to a reduction in the cytotoxic cascade


CPB leads to significant changes in RBC morphology (stiffer, more susceptible to hemolysis). CPB leads to release of PMNs, particularly during rewarming. Protein function is also altered, chylomicrons are released, the coagulation and fibrinolytic systems are activated, as is complement and the kallikrein-bradykinin cascade. Clearly these perterbations can significantly affect the patient, but their individual contribution to morbidity is difficult to tease out (as is the potential to affect change by modifying the response)

The optimal hemoglobin concentration during CPB is not known. Most patients are diluted to a Hgb of 7-10 during CPB, based on the belief that lower blood viscosity improves regional DO2 despite the decreased carrying capacity of blood. Animal work (piglet model) by Duebener et al. has cast some doubt onto the assumption that hemodilution improves viscosity [Duebener LF et al. Circulation 104 (12S1): I260, 2001; FREE Full-text at Circulation]