Perfusion

Following a succession of changes in circuitry and priming additives between 1993 and 1998, a comprehensive re-evaluation of neonatal cardiopulmonary bypass (CPB) practice was undertaken. Samples from 10 infants (Group 1) undergoing CPB were evaluated for osmolality, oncotic pressure, total protein, hematocrit, glucose, and electrolytes (Na⁺, K⁺, iCa²⁺). These samples were tested at six measurement points: (1) after priming, (2) patient pre-CPB, (3) CPB-start, (4) CPB-mid, (5) CPB-end, and (6) post-modified ultrafiltration (MUF). Prime volumes were also carefully measured as well as the type and amount of volume given during CPB. After evaluating the initial data, changes in protocol regarding mannitol, calcium correction, and oncotic strength on CPB were made. Following implementation of these protocol changes, a second set (Group 2) of 10 infants was identically evaluated.

Group 1 prime osmolality was 379 ± 44 mOsm/kg, while Group 2 prime osmolality was 324 ± 14 mOsm/kg (p = 0.003). There were no differences in osmolality between groups during bypass and osmolality was unaffected by modified ultrafiltration. Ionized calcium levels were significantly different at the end of bypass between Group 1, 0.6 ± 0.1 mmol/l; and Group 2, 1.17 ± 0.24 mmol/l (p < 0.001). In Group 1, there was a 40% drop (p = 0.001) in colloid osmotic pressure (COP) levels from pre-CPB (13.3 ± 3.4 mmHg) to CPB-end (8.8 ± 1.2 mmHg). In Group 2, there were no differences in COP during CPB. COP levels of Group 1 and Group 2 at CPB-end were 8.8 ± 1.2 mmHg and 14 ± 1.9, respectively (p < 0.0001). Total volume addition during bypass for Group 1 was 363.5 ± 148.7 ml and for Group 2 was 245.1 ± 92.2 ml (p < 0.05). In conclusion, progressive changes in neonatal circuits and techniques can have potentially wide-ranging effects on electrolyte and osmotic/oncotic homeostasis. An audit of perfusion management through expanded laboratory tests is recommended, especially in periods of change.

Heparin-bonded perfusion circuits have been reported to reduce the thrombus formation during various levels of systemic heparinization. The goal of this study was to compare the efficacy of thrombo-resistance of the Sarns 3M heparin-bonded circuit to Baxter Duraflo II and untreated control in a porcine model. Fifteen Yorkshire pigs (60-65 kg) were anesthetized, heparinized with 3000 IU, intravenously (i.v.) and surgically cannulated with an internal jugular outflow and a femoral vein inflow. All circuits consisted of a 22-Fr venous cannula, centrifugal pump, arterial filter, an 18-Fr cannula for return and connected with equal lengths of 3/8″ polyvinyl chloride tubing. The flows were maintained at 2.0 l/min for 4 h. Thrombus formation in filter samples were morphometrically analyzed through macro-densitometry, light microscopy, and scanning electron microscopy (SEM). Our findings revealed that the 3M circuit had significantly less gross thrombus (p < 0.001), 66% and 84% less microscopic thrombi and fivefold less SEM-measured aggregates (p = 0.03) compared to the Duraflo II and uncoated groups. This study demonstrated that the 3M heparin-bonded circuit had significantly reduced the formation of micro- and macro-thrombi in the minimally heparinized pig model compared to the Duraflo II and untreated control circuits.

Sedative agents are routinely administered to critically ill patients, both on and off extracorporeal membrane oxygenation (ECMO), to enable patients to be comfortable and facilitate patient management. It has been observed empirically in our paediatric intensive care unit that doses of sedative drugs required to achieve desired levels of sedation in ECMO patients are far greater than those used in non-ECMO patients. These differences could not simply be accounted for by differences in patient types, clinical status or sedation levels. We therefore undertook an in vitro evaluation of drug binding in ECMO circuits.

This study investigated how the polyvinyl chloride (PVC) and silicone rubber components of neonatal ECMO circuits affect drug delivery in patients through drug sorption. Phase 1 investigated drug uptake by the two polymers in static solutions of known concentrations of four commonly used sedative drugs: lorazepam, midazolam, diazepam and propofol. Phase 2 involved the setting up of a complete neonatal ECMO circuit, injecting the drug solutions pre-reservoir at a flow rate of 350 ml/min and collecting samples post-oxygenator for analysis.

Phase 1 results revealed significant uptake of drugs with losses in the range 40-98% and in the order propofol • diazepam • midazolam • lorazepam. Phase 2 results were similar and in the first 40 min of running an ECMO circuit only 10% of propofol passed through the circuit. These results may help to explain observed clinical phenomena and raise important issues regarding drug dosing in ECMO patients.

The bronchoconstrictive effects of alveolar hypocapnia during weaning from cardiopulmonary bypass (CPB) were investigated in patients undergoing elective coronary artery revascularization. Thirty patients were randomly assigned into two equal groups. In both groups, mechanical ventilation was initiated for 3 min prior to weaning from CPB with the venous pressure low. This kept the pulmonary vascular bed empty, resulting in alveolar hypocapnia (ETCO₂ < 2 kPa). Peak airway pressure (P _peak) and plateau pressures (P _plateau) were recorded. In group 1, 5% CO₂ was added to the inspiratory gas mixture and the ETCO₂ allowed to rise (ETCO₂ > 3.3 kPa). The ventilation pressure measurements were recorded again after 3 min stabilization. In group 2, the venous pressure was increased to allow the pulmonary venous bed to fill and the ventilation pressures recorded after a 3 min period of stabilization.

In group 1, the ventilatory pressures dropped significantly (p < 0.001) when the alveolar hypocapnia was reversed with added CO₂ (P _peak 19.71 ± 5.7 to 12.31 ± 2.8 cmH₂O and P _plateau 13.15 ± 3.28 to 9.15 ± 2.23 cmH₂O). In group 2, a similar effect was achieved by allowing filling of the pulmonary vascular bed (P _peak 17.46 ± 4.72 to 11.92 ± 3.03 cmH₂O and P _plateau 13.93 ± 4.10 to 9.37 ± 3.00 cmH₂O).

These results suggest that filling the pulmonary vascular bed prior to initiating ventilation, when weaning from CPB, prevents the otherwise deleterious effects of alveolar hypocapnia, resulting in raised bronchomotor tonus and raised airway pressures.

Ten patients admitted for coronary artery bypass grafting were investigated with respect to the influence of cardiopulmonary bypass (CPB) on red blood cell (RBC) trauma. Blood samples were collected prior to, at the start of, and at 30 and 60 min of CPB. RBC deformability was assessed by filtering re-suspended RBCs through a polycarbonate membrane using a computer-controlled filtrometer. Multiple regression analysis was employed to evaluate RBC flow-curve characteristics denoted by the initial filtration rate (IFR) and clogging slope (CS). Release of free haemoglobin was determined concomitantly. IFR was estimated at 90.39 μl/s and CS at -5.32 μl/s² prior to CPB. During 60 min of CPB, neither IFR nor CS deviated significantly (p > 0.05) from these reference values. However, release of free haemoglobin increased significantly (p < 0.018) from the start of CPB to the 60-min determination. In conclusion, 60 min of CPB seems not to alter significantly RBC deformability in a 5 μm pore filtration model, despite a significant release of haemoglobin.

In our early work in developing activated clotting time (ACT) assays, it became apparent that changes occurred in coagulation times as a whole blood sample aged (0-6 h). Subsequent studies showed that the coagulation parameters of plasma obtained from the samples remained stable during this time frame. These changes in whole blood clotting times during storage were eventually traced to the platelets. Several years of work demonstrated that this change was due to the removal of the blood from the vascular lining. This recalled a mechanism that was originally put forth in the 1970s with the discovery of prostacyclin. In this postulated mechanism, platelets are ‘time-bombs’. They are kept under control by prostacyclin (PGI₂) secreted by the vascular lining. Without this prostacyclin, platelets ‘preactivate’. Since that time, additional substances secreted by the vascular endothelium have been identified, such as nitric oxide, that also influence platelet activity.

The ‘preactivation’ of platelets in a blood sample can be followed using an ACT. In the same donor, the preactivation is uniform and reproducible over an extended period (months). There is, however, considerable variability between donors. Some donors’ platelets preactivate dramatically, while other donors show hardly any change. Prostacyclin, added to the blood sample when it is collected, prevents this preactivation. The clinical significance of these observations has yet to be clearly established, but these observations raise a number of questions with respect to methods for improving platelet function during bypass and in evaluating the risk of platelet-mediated cardiovascular disease.

Despite the acceptance of extracorporeal circulation as an effective modality to facilitate cardiac surgery, patient outcomes can be negatively influenced by the occurrence of perfusion incidents. A perfusion survey was conducted to identify safety techniques and incidents related to cardiopulmonary bypass (CPB).

An 80-question survey was mailed to chief perfusionists of all 1030 USA cardiac surgical centers using CPB. The survey was designed to examine practices and incidents that occurred during a 2-year period (July 1996 to July 1998). Five-hundred-and-fifty-two (54% response rate) surveys were returned, which accounted for 797 hospitals (79% of all cardiac centers) and 653 621 surgical procedures. Of the 27 identified CPB safety devices, the highest utilization was arterial line filters (98.5%) and the lowest arterial line bubble traps (3.4%). Of the reported cases, a CPB incident occurred once every 138 cases. The most common occurring incidents were protamine reactions (1:783), coagulation problems (1:771), and heater/cooler failures (1:1809). The rate of occurrence of an incident resulting in a serious injury or death was one for every 1453 procedures.

Although techniques and safety devices create a relatively secure environment for CPB, lower incident rates may be achieved with further improvements in coagulation monitoring and incident reporting.

Vacuum-assisted venous return (VAVR) has been reported to offer benefits for adults undergoing cardiopulmonary bypass (CPB), such as improved venous return, lowering priming volume (by eliminating the need to prime the venous line), and the use of smaller venous cannulae. All these benefits would be of particular value in pediatric perfusion because of the unique challenges of these smaller patients and the relatively large components of the CPB circuit. We have been using VAVR in children since the early summer of 1998 after we became comfortable with the technique and convinced of its efficacy in adults.

Ours is a medium-sized pediatric caseload of slightly more than 100 CPB cases per year. With that caseload, it is most effective for us to minimize the inventory of different sizes of disposables used. We have opted for an oxygenator/reservoir that has a maximum flow of 4 liters with a priming volume of about 1 liter. We have been unhappy with the large prime volume in infants and earlier, in 1997-1998, were using a smaller prime oxygenator/reservoir until it was recalled. Faced again with a larger priming volume in the infants, we decided to try vacuum to decrease hemodilution and to evaluate other possible benefits. Through the use of VAVR, we have been able to decrease our priming volume, use smaller venous cannulae, and have more consistent return while experiencing no adverse effects of VAVR in our pediatric cardiac surgery patients.

We investigated in vitro aortic arch vessel perfusion characteristics of single and multiple jet-stream cannulae and a new dispersion stream tip aortic cannula. Pressures and flows of all arch vessels were measured while directing cannulae jets at the different arch vessels using 6 l/min pump flow. The highest increase in pressure above the set systemic level of 80 mmHg and increase in flow above the set normal flow distribution in the arch vessels occurred in the jet-streamed arch vessels with the single stream cannula. The values were as follows: 29 mmHg and 118 ml/min for the innominate artery, 28 mmHg and 42 ml/min for the left common carotid artery, and 25 mmHg and 54 ml/min for the left subclavian artery. The dispersion stream cannula showed increases in pressure and flow, followed by the multiple stream cannula. Aortic cannula tips and the orientation of jets are potential sources of imbalances of arch vessel perfusion with possible clinical implications regarding perfusion of arch vessels during extracorporeal circulation.

An in vitro study has been carried out to assess the pump performance of a new peristaltic, extracorporeal displacement pump (Affinity) for cardiopulmonary bypass. The pump system consists of a pump rotor (0-110 rpm), a pump chamber, a venous reservoir with a 5/8″ connecting tube and the Affinity console. The polyurethane chamber is connected to the venous reservoir by a 5/8″ tube and fills passively due to the hydrostatic pressure exhibited by the fluid height in the venous reservoir. The implementation of an occlusive segment in the pump chamber, which collapses in low filling states, should prevent significant negative pressures. An in vitro circuit was filled with bovine blood (37°C, hematocrit 35%) and the pump flow was measured by an ultrasonic transit time flow probe with respect to pre-load, diameter and length of attached tubing in the venous line, pump speed (rpm) and size of the connecting tube (3/8″ and 5/8″).

At 108 rpm and a preload equal to 10 mmHg, the flow was 8.6 ± 0.42 l/min for an afterload of 80 mmHg. The reduction of the inlet connector to 3/8″ diminished the pump flow significantly to 5.2 ± 0.31 l/min (p < 0.0001). The pump flow decreased linearly with respect to the length of the attached tube in the venous line and for a 2 m long 5/8″ silicon tube, the rpm-optimized flow was still 6.0 ± 0.28 l/min at a preload of 10 mmHg. In case of low filling state or too high rpm, the occlusive segment collapsed and no cavitation bubbles could be detected.

Our in vitro measurements yield a nomogram for rpm-optimized blood flow with respect to the pre-load in the venous reservoir. The delivered 5/8″ connecting tube facilitates optimum filling of the pump chamber for high blood flow, but limits the use of venous reservoirs to Affinity products. The pump yields a high blood flow even when long tubing in the venous line is used. This makes the pump a candidate for a ventricular assist device. In hypovolemia or high rpm, the occlusive segment collapses and no negative pressure is generated at the inflow site of the pump chamber.