Acute Hemodilution in a Chronic Polycythemic Patient May Be Deleterious

An 80-yr-old woman (height, 1.60 m; weight, 59 kg; body surface area, 1.64 m²) was scheduled for surgery to remove an erythropoietin-producing renal tumor. Physical examination and laboratory tests on admission revealed an arterial blood pressure 160/75 mmHg, an erythropoietin concentration of 116 U/l (normal range, 1–10 U/l), and a hematocrit of 62%. The electrocardiograph showed signs of left ventricular hypertrophy. The patient was treated with a calcium-blocker (nifedipine) and a β blocker (atenolol) for chronic hypertension. It was decided to perform acute isovolemic hemodilution just before surgery, which would result in a quantity of autologous blood for transfusion during surgery if acute bleeding should occur.

After having obtained informed consent, and before anesthesia was induced, venous and arterial cannulae were inserted, including a pulmonary artery thermodilution catheter (Edwards 7-French; Baxter Healthcare Corp., Irvine, California) through the right internal jugular vein. Baseline measurements of the mean arterial blood pressure, pulmonary artery pressure, pulmonary artery wedge pressure, central venous pressure, and cardiac index were made before anesthesia. Arterial and mixed venous blood samples were obtained for the determination of hemoglobin, hemoglobin oxygen saturation (OSM 3; Radiometer, Copenhagen, Denmark), hematocrit, and partial pressure of oxygen (Po²), partial pressure of carbon dioxide (Pco²), pH, and base excess (ABL 505; Radiometer, Copenhagen, Denmark). From these data, systemic vascular resistance index, pulmonary vascular resistance index, left ventricle stroke work index, systemic oxygen delivery (Ḋo²), oxygen consumption (V̇o²), and oxygen extraction ratio (O²ER) were calculated, using the standard formulas. In addition, whole blood viscosity was measured using a Contraves LS 30 viscometer (Basel, Switzerland), measuring the blood viscosity at high (70 s⁻¹), medium (0.5 s⁻¹), and low (0.05 s⁻¹) shear rates. The microcirculation (sublingual mucosa) was visualized using orthogonal polarization spectral (OPS) imaging (Cytoscan; Cytometrics, Philadelphia, PA). 4This device, which is commonly used in our hospital with approval of the Medical Ethical Committee, allows on-line microscopic observation of the microcirculation by use of a small endoscopic-like light guide placed on tissue surfaces. By placing this handheld light guide, which is attached to a video camera, on the sublingual mucosa of a patient, the human microcirculation can be visualized in a completely uninvasive way.

Preanesthetic medication consisted only of an oral dose of 50 mg atenolol, 6 h before baseline. Anesthesia was induced with 5 mg/kg thiopental, 2 μg/kg fentanyl, and 0.5 mg/kg rocuronium. After tracheal intubation, the lungs were ventilated with a mixture of oxygen in air (fractional inspired oxygen tension [Fio²], 0.4), maintaining normocapnia. Anesthesia was maintained with isoflurane (end-tidal concentration, 0.7%) and fentanyl (100 μg/h); muscle relaxation was maintained with rocuronium. After induction of anesthesia, isovolemic hemodilution was performed by withdrawal of blood and simultaneous infusion of a 3.5% modified gelatin solution (Gelofusine^®, molecular weight 35,000 Da; B.Braun Melsungen, Melsungen, Germany) in a ratio of 2:3. All measurements were repeated after exchange of 500, 1,000, 1,500, 2,000, and 2,500 ml, and after retransfusion of 500 ml autologous blood. Except for postoperative data, all measurements were performed before surgery was started, over a period of approximately 4 h. A selection of the parameters is given in table 1.

At baseline, a cardiac index above the normal range and a low systemic vascular resistance index were found. Exchange of up to 2,000 ml blood resulted in a decrease in hematocrit from 59% to 34%. Simultaneously the whole blood viscosity, which was 7.6, 64, and 121 mPa · s (at high, medium, and low shear rates, respectively) at baseline, decreased to 4.6, 18, and 41 mPa · s. During hemodilution, cardiac index increased only slightly despite a further decrease in systemic vascular resistance index. Pulmonary artery wedge pressure and central venous pressure increased, Ḋo²gradually decreased, and O²ER increased from 10% at baseline to 15%.

On further hemodilution (to hematocrit 30%), V̇o²suddenly decreased to a value of 40% of baseline, although the O²ER remained at a low level (12%). At this last step of hemodilution, mean arterial blood pressure decreased to 77 mmHg, which was accompanied by acute ST depression and inverse T waves on the electrocardiograph. Simultaneously, the mixed venous base excess, which had been within the normal range of −3 to 3 mm, markedly decreased to −10.5 mm. Arterial base excess remained within the normal range. Because of these sudden changes, it was decided to retransfuse 500 ml autologous blood to a hematocrit of 37%. On retransfusion of this volume of autologous blood, the electrocardiographic abnormalities disappeared and V̇o²increased above postanesthetic baseline values. Mixed venous base excess returned to baseline values. No additional blood was retransfused.

Orthogonal polarization spectral imaging showed an increased number of microcirculatory networks with significantly dilated venules as compared to normal (fig. 1). During the whole procedure, this image did not change.

After retransfusion, the surgical procedure was started. Because of metastases, surgery was discontinued after obtaining some tissue samples from the tumor. The postoperative course was uneventful, and treatment with chemotherapy was started.

Alterations in hematocrit lead to corresponding changes in blood viscosity, which tend to cause blood flow to change in an inverse relation; isovolemic hemodilution is accompanied by an increase in carbon monoxide, whereas hemoconcentration (e.g. , chronic polycythemia) results in a decrease in carbon monoxide, due mostly to changes in systemic vascular resistance. 2,5However, in the current chronically polycythemic patient, systemic baseline parameters were not as expected: the cardiac index was not decreased, but increased, and the systemic vascular resistance index was below normal levels. Together with the increased number of microcirculatory networks with dilated venules on the OPS images, these findings suggest shunting of blood at tissue level, which may explain the fixed low O²ER, decreased systemic vascular resistance index despite high viscosity, and consequently the high cardiac index. Although to our knowledge OPS imaging has not been validated during polycythemic conditions, the OPS images, as well as the O²ER, remained unchanged during hemodilution, suggesting that the possible shunting persisted. As blood is bypassing the capillary beds, more oxygen ends up at the venous side of the vascular bed without being utilized by the tissues, as is reflected in the current case by the mixed venous partial pressure of oxygen (PT̄vo²) and mixed venous saturation (ST̄vo²). Even when the V̇o²started to decrease at hematocrit 34% (hemoglobin, 9.5 g/dl), suggesting that oxygen uptake was reaching a state of oxygen supply dependency, PT̄vo²and ST̄vo²showed little change.

The critical level of hemodilution at a hematocrit of 30–34% is in contrast with previous reports, where in anesthetized animals 1,3,6,7and humans 8a critical hematocrit of 9%–12% could be demonstrated, and in conscious humans no critical level of hemodilution could be determined. 9At further hemodilution to a hematocrit of 30%, V̇o²decreased severely, and cardiac ischemia became apparent on the electrocardiographic registration. This indicates that, in our patient at a hematocrit of 30% (hemoglobin, 8.0 g/dl), myocardial oxygenation was already compromised. Although this report encompasses only a single patient, such a severe decrease in V̇o²is not likely to be caused by intermeasurement variations. In a recent study of hemodilution down to a hemoglobin concentration of 5.2 g/dl in awake volunteers, 3 of 55 subjects showed electrocardiographic changes, but not before a hemoglobin concentration between 5.0 and 6.7 g/dl. 10Administration of 500 ml autologous blood not only restored the hemodynamics in our patient, but also led to what can be interpreted as an overshoot in V̇o², which might have served to meet a possible oxygen debt.

Based on the above data, it may be hypothesized that chronic polycythemia can be compensated for by peripheral shunting of the blood. This adaptive response to conditions of increased blood viscosity does not seem to change during acute hemodilution, thereby decreasing the tolerance for acute isovolemic hemodilution. One might argue that advanced age or the chronic use of a β-blocking agent may have influenced the critical point of hemodilution. However, in clinical studies it has been demonstrated that these factors do not result in a functional impairment of the compensatory mechanisms at work during hemodilution. 11,12 

In conclusion, from the results of the presented case it may be hypothesized that chronic polycythemia can be compensated by peripheral shunting, of which the exact nature is subject for further investigation. Although extrapolation beyond this case will be difficult, one should be cautious with acute hemodilution to subnormal levels in chronically polycythemic patients in the meantime, because this may be deleterious for the patient.