Methods

The study was approved by the Institutional Review Board of The Children's Hospital of Philadelphia. The cardiac surgery database at The Children's Hospital of Philadelphia was reviewed. We identified all patients with hypoplasia of the left heart or a variant undergoing the first stage of Norwood reconstruction over a period of two years beginning September 1, 2000 and ending August 31, 2002. Demographic and anatomic data were recorded from the medical record.

During the period, reconstruction was performed by four surgeons. A dedicated team of cardiac anaesthesiologists managed all patients during the procedure. Alpha-stat blood gas management was utilized. Deep hypothermic circulatory arrest was used in all patients.

Haemodynamic parameters, including heart rate, systolic and diastolic blood pressure, were recorded after separation from bypass immediately prior to the institution of arterio-venous modified ultrafiltration and immediately after completion of modified ultrafiltration. Use of inotropic drugs was recorded, and any change in usage was noted during the period of modified ultrafiltration. If a patient underwent more than one episode of cardiopulmonary bypass and modified ultrafiltration during the operation, we analysed the haemodynamic data for the second period.

Data are presented as mean plus or minus standard deviations. We compared values obtained before and after ultrafiltration using the paired t-test, considering p values of less than 0.05 to be significant.

Results

During the period, we performed reconstruction in 99 neonates, with 78 having classic hypoplasia of the left heart, and variations on this anatomy in the remaining 21. Mean weight at surgery was 3.1 plus or minus 0.7 kilograms. We defined a group of patient to be at high risk based on previously described criterions, including the presence of additional cardiac defects or genetic syndromes, weight at time of surgery less than or equal to 2.5 kilograms, and/or the presence of major additional noncardiac anomalies.¹ These criterions were satisfied by 44 patients.

Mean duration of deep hypothermic circulatory arrest was 45 plus or minus 15 minutes. Total support time, representing cardiopulmonary bypass plus the period of deep hypothermic circulatory arrest, was 100 plus or minus 26 minutes. A modified Blalock– Taussig shunt was constructed in 95 patients, with a conduit placed from the right ventricle to the pulmonary arteries in four. Arterio-venous modified ultrafiltration was performed in all patients after separation from cardiopulmonary bypass. A second period of cardiopulmonary bypass was necessary in 12 patients, and a second period of modified ultrafiltration was performed in 9 patients.

Standard management from the stance of inotropic support consisted of administration of 3 micrograms per minute of Dopamine instituted prior to separation from cardiopulmonary bypass. In most patients, a loading dose of 100–250 micrograms per kilogram of Milrinone was administered prior to modified ultrafiltration. Only 8 patients were receiving additional inotropic support at the time of modified ultrafiltration. Inotropic support remained constant for 95 patients during the period of modified ultrafiltration, but needed to be increased in 4 patients.

Modified ultrafiltration was performed successfully in all 99 patients. There were no complications related to ultrafiltration, and we did not need to discontinue the procedure because of haemodynamic instability in any patient. The mean duration was 10 plus or minus 2 minutes, and the mean volume of filtrate removed was 104 plus or minus 29 millilitres per kilogram. The haematocrit increased from 31 plus or minus 4 percent prior to modified ultrafiltration, to 46 plus or minus 6 percent at the end of modified ultrafiltration, signifying significant haemoconcentration (p value less than 0.0001). During the period of modified ultrafiltration, the mean heart rate decreased from 170 plus or minus 17 beats per minute to 158 plus or minus 16, systolic blood pressure increased from 57 millimetres of mercury plus or minus 12 millimetres of mercury to 63 millimetres of mercury plus or minus 13 millimetres of mercury, and diastolic blood pressure increased from 35 plus or minus 8 to 30 plus or minus 7 millimetres of mercury (all significant at values less than or equal to 0.0001). Of the 55 patients deemed to be at low-risk, 3 died in hospital (5.5 percent), but 12 of the 44 (27.3 percent) considered to be at high-risk died prior to discharge from hospital following the attempted first stage of reconstruction.

Discussion

Our current study demonstrates that the use of arterio-venous modified ultrafiltration is well tolerated following the first stage of the Norwood sequence of reconstructions used for palliation of hypoplasia of the left heart and its anatomic variants. The immediate haemodynamic changes demonstrate a decrease in heart rate, and an increase in both systolic and diastolic blood pressures, consistent with previous reports of modified ultrafiltration. In addition, there is significant haemoconcentration. There were no complications related to the procedure, and we did not need to discontinue ultrafiltration in any patient because of haemodynamic instability.

We used the arterio-venous form of modified ultrafiltration in our study. No changes are made in the protocol for those undergoing the first stage of the Norwood sequence compared to patients undergoing repair of other defects. After separation from cardiopulmonary bypass, blood is withdrawn from the aortic cannula at a rate of 10 to 15 millilitres per kilogram per minute, and passed through the haemofilter. The concentrated blood is returned to the right atrium. As fluid is removed from the patient, intravascular volume, and thus filling pressures decrease. Blood from the bypass circuit is concentrated with the haemofilter, and re-infused to maintain intravascular volume. Thus this form of modified ultrafiltration provides the capability to remove excess water from the patient, and also to salvage blood from the bypass circuit.^{2, 3} There was initial concern that withdrawal of blood from the aorta in the period immediately subsequent to cardiopulmonary bypass would result in haemodynamic instability and hypotension. This worry proved unfounded, however, as use of modified ultrafiltration corresponded with an increase in blood pressure.⁴

During the period of our study, flow of blood to the lungs after first stage reconstruction was usually provided by an aorto-pulmonary shunt arising, in most cases, from either the brachiocephalic or subclavian arteries. A conduit placed from the right ventricle to the pulmonary arteries was used as the source of pulmonary flow in only 4 patients. The number of patients in our series undergoing reconstruction with such a conduit is too small to determine if the outcome is different compared to an aorto-pulmonary shunt. Coronary arterial flow following the first stage of reconstruction is partially or wholly dependent on retrograde flow through the native ascending aorta, which may be very diminutive, sometimes measuring less than 2 millimetres in diameter. There has been concern that arterio-venous modified ultrafiltration incorporating withdrawal of blood from the aortic cannula might compromise either coronary arterial or pulmonary flow, and hence produce haemodynamic instability. The current study demonstrates that, as for patients with other cardiac defects, this concern is unfounded for patients with hypoplasia of the left heart. Indeed, there is a significant improvement in heart rate and blood pressure during the period of modified ultrafiltration. The increase in diastolic blood pressure may have a beneficial effect on coronary arterial flow. This improvement in haemodynamics cannot be attributed to alterations in inotropic drug therapy, as our use of drugs remained constant in 95 of 99 patients during the period of modified ultrafiltration.

There are several limitations to our study. Most importantly, it is retrospective, and there is no control group. Its purpose, however, was to determine if the use of modified ultrafiltration is safe, and if it produces haemodynamic benefits similar to those reported in patients with other cardiac defects. Demonstration of improved haemodynamics, of course, does not necessarily correlate with improvement in outcome. Further studies will be necessary to determine if use of modified ultrafiltration results in a significant outcome benefit for patients with hypoplasia of the left heart.

In summary, therefore, we have demonstrated that use of modified ultrafiltration is well tolerated in patients following the first stage of the Norwood sequence of reconstructions, and produces haemodynamic benefits similar to its use in patients with other cardiac defects. We have shown that modified ultrafiltration in these patients should not be avoided because of fear of haemodynamic instability. The early outcomes following surgical palliation in these patients continue to improve.¹ We submit that the use of modified ultrafiltration is an additional incremental strategy that may contribute to further improvements in outcome.

Presented at the 2nd Joint Meeting of the European Association of Cardiothoracic Surgeons and the Society of Thoracic Surgeons, October 12–15, 2003, Vienna, Austria.

Dr. Kuypers and Dr. van Rossem were supported by grants from the Dutch Heart Foundation and The Karel Frederik Stichting Foundation.