Critical Hematocrit in Intestinal Tissue Oxygenation during Severe Normovolemic Hemodilution

The protocol of the present study was approved by the Animal Research Committee of the Academic Medical Center at the University of Amsterdam. Animal care and handling were performed in accordance with the national guidelines for care of laboratory animals. The experiments were performed in 18 male Wistar rats with a mean (± SD) body weight of 337 ± 23 g.

The rats were anesthetized with an intraperitoneal injection of a mixture of 90 mg/kg ketamine, 0.5 mg/kg medetomidine, and 0.05 mg/kg atropine. Anesthesia was maintained with 50 mg·kg⁻¹·h⁻¹ketamine administered intravenously. To compensate for fluid loss, crystalloid solution was administered continuously at a rate of 15 ml·kg⁻¹·h⁻¹. Body temperature was measured with a thermocouple placed in the rectum and was maintained at 37 ± 0.5°C with a heating pad below and a warming lamp above the animal. Tracheotomy was performed, and a polyvinylchloride tube (charrière 6) was inserted into the trachea to enable mechanical ventilation with a mixture of 30% oxygen and 70% nitrogen. A heat and moisture exchanger (Humid-Vent Micro, Gibeck, Sweden) was placed between the tracheal tube and the ventilator to diminish loss of fluid through the mechanical ventilation. A capnometer (Capstar-100, CWE, Inc., Ardmore, PA) was used to measure end-tidal carbon dioxide partial pressure (PCo²). This information was used to adjust ventilator settings to maintain an arterial PCo²between 35 and 40 mmHg.

A polyethylene catheter (OD, 0.9 mm) was inserted into the right jugular vein for intravenous administration of drugs and fluids. The tip of the catheter was advanced near the right atrium for central venous blood sampling. A similar catheter was placed into the right carotid artery and connected to a pressure transducer for continuous monitoring of arterial blood pressure and heart rate. Catheters of the same size were placed into both the right femoral artery and vein for withdrawal of blood and administration of fluid. After midline laparotomy, a polyvinylchloride catheter (OD, 0.9 mm) was placed in the urinary bladder to prevent distension of the bladder wall and to monitor the urine production during the experiment.

The gut was exteriorized carefully and a 0.5-mm perivascular flow probe (Transonic Systems Inc., Ithaca, NY) was placed around the superior mesenteric artery and connected to a flow meter (T206, Transonic Systems Inc.). For mesenteric venous blood sampling, an ileocecal vein was isolated under the microscope, ligated distally, and cannulated with a polyethylene catheter (OD, 0.8 mm). The tip of this catheter was advanced into the vessel in a proximal direction to have an estimate of the mesenteric venous blood gas values. A thermocouple was placed on the intestinal surface, and all exposed organ surfaces were covered with plastic foil to prevent evaporative fluid loss.

Arterial pressure was measured in the carotid artery. Mean arterial pressure (MAP; mmHg) was calculated as MAP = diastolic pressure + (systolic pressure − diastolic pressure)/3. The amplitude of the arterial blood pressure was calculated as pulse pressure (P^puls) = systolic pressure − diastolic pressure. Blood flow in the superior mesenteric artery, Q̇^sma(ml·kg⁻¹·min⁻¹) was measured continuously and indexed according to body weight. Urine output through the urinary bladder catheter was monitored. To avoid bias by differences in the time needed for preparation of the animal, the volume of produced urine at each measurement point was normalized to the urine volume at baseline 1. Blood samples of 0.2 ml each were collected and replaced with the same volume of pasteurized protein solution (PPS; CLB, Amsterdam, the Netherlands). At each measurement point, an arterial sample was taken from the femoral artery, a central venous sample was taken from the jugular venous catheter, and a mesenteric venous sample was taken from the ileocecal venous catheter. The samples were used to determine blood gas values (ABL505, Radiometer, Copenhagen, Denmark), as well as hematocrit, hemoglobin concentration, and hemoglobin oxygen saturation (OSM 3, Radiometer).

Intestinal Do²was calculated as Do^2,int(ml·kg⁻¹·min⁻¹) = Q^sma· arterial oxygen content, which was calculated as (1.31 ×[hemoglobin]× Sao²) + (0.003 × Pao²). Intestinal V̇o²was calculated as V̇o^2,int(ml·kg⁻¹·min⁻¹) = Q^sma× (arterial − mesenteric venous oxygen content difference). Mesenteric venous V̇o²was calculated as (1.31 ×[hemoglobin]× S^mvo²) + (0.003 × P^mvo²). The intestinal oxygen extraction ratio was calculated as O²ER^int(%) =V̇o^2,int/Do^2,int. Because values of mesenteric or portal venous pressure were not available, an estimation of the vascular resistance of the superior mesenteric flow region was made: MAP:Q^smaratio (U) =[MAP/Q^sma]× 100.

The intestinal μPo²was measured using the oxygen-dependent quenching of Pd-porphyrin phosphorescence. 12–14Excitation of Pd-porphyrin by a pulse of light causes emission of phosphorescence with a decay in time, which is quantitatively related to the oxygen concentration. Pd-meso-tetra(4 carboxy-phenyl)porphine (Porphyrin Products, Logan, UT) is coupled to human serum albumin to form a large molecular complex that, when injected intravenously, is confined mainly to the vascular compartment. 13,15One milliliter of a 4-mm Pd-porphyrin solution was administered, corresponding with a dosage of 12 mg/kg body weight. The μPo²measurements were made with an optical fiber for transmission of excitation and emission light, attached to a phosphorimeter. To determine which microvascular compartment is measured by fiber phosphorimetry, the Pd-porphyrin phosphorescence fiber technique has been compared with a microscopic phosphorimeter. 16Simultaneous Po²measurements with the fiberoptic technique showed excellent correlation with microscopically measured Po²in capillaries and first-order venules, but not with arteriolar and venous Po²values, at different Fio²levels. 16This study allowed us to term the fiberoptic measurement of Po²as the measurement of μPo². The fiber was placed near the serosal surface of the terminal ileum. Fiberoptic measurements of μPo²incorporate blood vessels under the fiber over an area of approximately 1 cm²to a penetration depth of approximately 0.5 mm. 13,17Because the calibration constants in the calculation of μPo²from the phosphorescence decay time are temperature-dependent, temperature measurements from the intestinal surface were used for correction of these constants.

After surgery and stabilization, two baseline measurements were made during a 1-h period. At this point, the animals were randomized between the hemodilution (n = 12) and the control group (n = 6). Normovolemic hemodilution was accomplished by withdrawal of blood from the femoral artery and simultaneous administration through the femoral vein of PPS at the same rate. The oncotic pressures of this protein solution and rat blood were determined, and it was found that the protein solution is slightly hyperoncotic compared with rat blood (oncotic pressures of 14.5 ± 0.4 and 12.7 ± 0.6 mmHg for PPS and rat blood, respectively) and therefore suitable for isovolemic hemodilution. Infusion–withdrawal occurred at a rate of 20 ml/h, using a double syringe pump (Harvard 33 syringe pump, Harvard Apparatus, South Natick, MA), and did not cause any undesired hemodynamic reactions in our model. Four dilution steps were made: from baseline down to a hematocrit of approximately 25% (H1), then to 15% (H2), to 10% (H3), and finally to 5–10% (H4). A 15-min stabilization period followed each dilution step before measurements were made. Because hematocrit values lower than 5% were found to be incompatible with life in this model, the experiments were terminated after measurement H4; an overdose of pentobarbital (60 mg intravenously) was administered to the animals. In the control group, no hemodilution was performed, but measurements were made at similar time intervals as in the hemodilution group.

To provide more information on the effect of the hemodilution procedure on the animals’ volume status, in four additional animals the arterial and central venous pressures were measured during hemodilution in combination with measurement of the intestinal μPo². Hemodilution was performed in the same way as in the other animals, and subsequently a fluid challenge of 2.5 ml PPS was administered, which could be expected to cause an increase in MAP if the animal would be hypovolemic.

Values are reported as mean ± SD. Data within each group were analyzed using analysis of variance for repeated measurements. When appropriate, post hoc  analyses were performed with the Student-Newman-Keuls test. P^mvo²and μPo²were compared at each measurement point with the Student paired t  test. The hemodilution and the control group were compared with an unpaired t  test. P  values less than 0.05 were considered significant. The critical level of hemodilution was determined from plots of hematocrit and Do^2,intagainst both μPo²and V̇o^2,int. The critical points were defined as the points at which μPo²and V̇o^2,intbecame dependent on hematocrit and Do^2,intwith further hemodilution. These points were determined for each animal separately by the intersection of the two best-fit regression lines with a least sum of squares technique. 1The correlation between μPo²and V̇o^2,intwas calculated with the Pearson correlation coefficient.

On completion of the experiment, an average of 26.3 ± 5.1 ml blood had been exchanged for an identical volume of PPS. Baseline values in the hemodilution and the control group were not different. As shown in table 1, hemodilution decreased hematocrit from 44.9 ± 3.8% at baseline to 6.2 ± 0.8% at H4 and [hemoglobin] from 14.6 ± 1.3 g/dl to 1.8 ± 0.3 g/dl. Although hematocrit and [hemoglobin] decreased in the control group as well, to 35.1 ± 1.3% and 11.6 ± 0.5 g/dl at H4, respectively, the values in the hemodilution group were significantly lower from control from H1 to H4. MAP, Q̇^sma, heart rate, and MAP:Q̇^smaratio did not change in the control group. In the experimental group, MAP (106 ± 13 mmHg at baseline) decreased significantly at H2 (corresponding hematocrit, 14.8 ± 2.8%) to 85 ± 23 mmHg and reached a minimum of 46 ± 11 mmHg at H4. Heart rate (267 ± 20 beats/min at baseline) decreased significantly to 241 ± 35 beats/min at H3 and to 231 ± 35 beats/min at H4. During hemodilution, P^pulsincreased significantly compared with the control group, from 13 ± 5 mmHg at baseline 1 to 30 ± 13 mmHg at H3. P^pulsdid not change significantly in the control group. Q̇^sma(14.0 ± 2.3 ml·kg⁻¹·min⁻¹at baseline) increased at H1 to 17.5 ± 2.0 ml·kg⁻¹·min⁻¹(P < 0.05), reached a maximum value of 21.9 ± 3.3 ml·kg⁻¹·min⁻¹at H3, and returned to baseline values at H4. The MAP:Q̇^smaratio (2403 ± 466 U at baseline) decreased progressively to 861 ± 324 U at H4. Again, significance was reached at H1 (1810 ± 431 U). The urine production was normalized to the amount at the start of the experiment and was similar in the hemodilution and the control group. A summary of these data is given in table 1.

Results of the oxygen measurements and related parameters are shown in tables 2–4. In the hemodilution group, mesenteric venous Po²(P^mvo²) (60 ± 8 mmHg at baseline) decreased significantly at H1 (42 ± 4 mmHg) to 34 ± 4 mmHg at H4. Intestinal μPo²in the same group, however, remained stable around 60 mmHg until H2, then it decreased to 25 ± 3 mmHg at H3 to a final value of 14 ± 4 mmHg at H4. Comparing μPo²to P^mvo²(fig. 1), it is shown that the P^mvo²became significantly lower than the μPo²at H1 (P < 0.05). However, in contrast to this, at H3 and H4 (when hematocrit was decreased below 10%), μPo²had fallen below P^mvo²(P < 0.05). The figure clearly shows the difference with the control group, in which both parameters did not change.

Mesenteric venous oxygen saturation (S^mvo²) was 77 ± 9% at baseline, started to decrease at H1 to 67 ± 8% (P < 0.05), with a final value of 33 ± 8% at H4. A similar pattern was observed for the central venous oxygen saturation (S^cvo²), with values of 61 ± 12 at baseline to 45 ± 13 at H1 (P < 0.05) and 24 ± 7 at H4. In the control group, S^mvo²did not change, and S^cvo²decreased significantly at H3 and H4.

Despite an increased Q̇^sma, Do^2,intdecreased immediately when hematocrit was diminished by hemodilution, from 2.6 ± 0.3 ml·kg⁻¹·min⁻¹at baseline to 0.4 ± 0.1 ml·kg⁻¹·min⁻¹at H4. Initially, V̇o^2,intremained constant at 0.6 ml·kg⁻¹·min⁻¹but decreased significantly at H3 (0.5 ± 0.1 ml·kg⁻¹·min⁻¹) and reached 0.3 ± 0.1 ml·kg⁻¹·min⁻¹at H4. At H1, the O²ER^intwas significantly increased compared with baseline (from 25 ± 8% at baseline to 34 ± 8% at H1) and continued to increase to 70 ± 6% at H4. In the control group, V̇o^2,intremained constant around 0.6 ml·kg⁻¹·min⁻¹throughout the experiment. However, Do^2,int(2.6 ± 0.2 ml·kg⁻¹·min⁻¹at baseline) decreased slightly but significantly to 2.1 ± 0.1 at H4, which was still significantly higher than in the hemodilution group. As a result, the O²ER^intin the control group (20 ± 2% at baseline) was increased at H2 (P < 0.05) to 29 ± 3% at H4.

Arterial p  H and Sao²did not change in either group. Pao²did not change in the control group (171 ± 18 mmHg at baseline) but increased during hemodilution from 165 ± 17 mmHg at baseline to 194 ± 15 mmHg at H4. Arterial PCo²in the control group remained constant around 37 mmHg; in the hemodilution group (38 ± 4 mmHg at baseline) PCo²decreased at H3 to 23 ± 7 mmHg at H4.

In the control group, mesenteric venous p  H and PCo²were stable around 7.34 and 43 mmHg, respectively. After hemodilution, P^mvCo²did not change from baseline or control. However, p  H^mvdecreased significantly at H3 (7.26 ± 0.07) and H4 (7.19 ± 0.06).

The critical Do^2,intand hematocrit values are shown in figures 2 and 3. V̇o^2,intbecame dependent on supply at a Do^2,intof 1.5 ± 0.5 ml·kg⁻¹·min⁻¹. For the μPo², critical Do^2,intwas 1.6 ± 0.5 ml·kg⁻¹·min⁻¹. A similar critical value was calculated for the hematocrit:V̇o^2,intdecreased with hematocrit at 15.8 ± 4.6%. The critical hematocrit for μPo²was 16.0 ± 3.5%. The critical Do²values and hematocrit values were not significantly different for μPo²and V̇o^2,int, indicating that V̇o^2,intand μPo²became dependent on supply at the same point during hemodilution. This is illustrated in figure 4: during intact oxygenation, little correlation could be found between μPo²and V̇o^2,int. As soon as intestinal oxygenation was impaired by progressive hemodilution, a significant correlation between intestinal μPo²and V̇o^2,intcould be demonstrated (Pearson r²= 0.86;P < 0.0001).

The results of the combined measurement of the central venous pressure, MAP, and intestinal μPo²during hemodilution in four animals are shown in figure 5. The central venous pressure increased slightly but not significantly during hemodilution. The MAP and the intestinal μPo²showed a similar response to hemodilution as in the other animals. After the fluid challenge, hematocrit decreased from 8% to 6%, but the central venous pressure, MAP, and μPo²did not change significantly.

The main result of this study is that a critical point in the intestinal microvascular oxygenation could be identified during extreme normovolemic hemodilution. The critical point for μPo²corresponded with the critical point for V̇o^2,int: both V̇o^2,intand μPo²were limited by supply at the same stage during severe normovolemic hemodilution. Not only did we find similar critical Do^2,intvalues for μPo²and V̇o^2,int, but critical hematocrit values were of the same order as well. When the critical point in intestinal Do²had been reached and μPo²and V̇o^2,intwere limited by supply, a significant correlation between the V̇o^2,intand the μPo²could be demonstrated, but not between V̇o^2,intand P^mvo². At this time, O²ER^inthad not reached a maximum yet, but continued to increase until the end of the experiment, indicating that regulation of the intestinal tissue oxygenation was still functioning, although insufficiently.

Until a hematocrit of ± 15% at H2, μPo²remained constant, resulting in a significant divergence between μPo2 and P^mvo²(fig. 1). This might be explained by the notion from other investigators that hemodilution can increase microvascular perfusion by increasing the amount of perfused capillaries (capillary recruitment) or by vasodilatation of microvessels already perfused. 18–20As a result, the absolute amount of oxygen transported to the capillaries can be maintained and a more homogeneous distribution of the flow in the microcirculation is provided as well. 21,22This shortens the oxygen diffusion distances and facilitates the oxygen uptake (increased extracting bed). In the regulation of the intestinal tissue oxygenation, capillary recruitment is of great importance. 23In fact, it has been shown to be of greater quantitative significance than blood flow autoregulation in preventing cellular hypoxia when intestinal perfusion pressure is reduced. 24In addition, changes that have been observed in skeletal muscle microcirculation after the decrease in blood viscosity during hemodilution can be assumed to contribute to the preservation of the intestinal tissue oxygenation as well: a decrease in the precapillary diffusional oxygen loss, resulting in more oxygen being delivered to the capillaries and a diminished oxygen diffusion path, 25–29allowing a more efficient utilization of the remaining circulating erythrocyte volume. 30,31 

With more progressive hemodilution, Do^2,intand hematocrit reached the critical point where μPo²became limited by the decreasing oxygen supply and decreased progressively below P^mvo². This resulted in a Po²gap between mesenteric venous and microvascular blood, demonstrating that at this stage, oxygen was being shunted from the intestinal microcirculation. Principal mechanisms that can cause functional shunting within the microcirculation are: convective shunting through anatomical anastomoses, direct diffusion of oxygen from arterioles to venules, altered heterogeneity of the microvascular architecture leading to “vascular steal,” and inability of hemoglobin to off-load oxygen fast enough to the tissues as it passes through the microcirculation. 9Which of these mechanisms can be held responsible for the Po²gap during hemodilution has yet to be determined. Whether it is the origin or merely an effect of the inability of intestinal regulatory mechanisms to maintain tissue oxygenation is not clear.

The decrease in both S^mvo²and S^cvo²reflects the increase in O²ER during the experiment. O²ER^intwas increased to 70% at H4, indicating that regulation of the intestinal oxygenation was capable of increasing O²ER^int, although not sufficient to maintain the intestinal oxygenation at baseline levels. The similar critical values and the significant correlation between μPo²and V̇o^2,intin the state of oxygen supply dependency support the hypothesis that the amount of physically dissolved oxygen within the microcirculation can be considered as the major driving pressure for diffusion of oxygen into the tissue cells, and as such, can be regarded as an important determinant of tissue V̇o².

Q^smais an important determinant of the Do^2,intand as such influences the critical hematocrit and Do^2,intvalues during hemodilution. During hemodilution, Q^smahas been reported to increase sufficiently to maintain oxygen flux to the organs, including the small intestine, until a hematocrit of ± 10% is reached. 32Although Q^smaincreased significantly in this study, Do^2,intdecreased; the increase in Q^smawas never sufficient to fully compensate the decreased oxygen-carrying capacity of the blood. This observation has been reported in prior studies as well 2,33and indicates that an increase in O²ER is more important in the preservation of intestinal V̇o².

Despite an increase after hemodilution, Q^smafell back to baseline values in the final phase of the experiment, when hematocrit was only 6%. Simultaneously, MAP decreased to 46 mmHg, and the heart rate was significantly decreased as well. Most likely, hemodynamic stability could no longer be maintained at this point. This is supported by the decrease in arterial PCo²at H4. It can be hypothesized that the substantial hypotension and decrease in Q^smaat the lowest hematocrit were caused by inadequate cardiac function.

The arterial p  H did not change during the experiment, unlike the mesenteric venous p  H; the latter decreased significantly at H3. This was the same point where both the intestinal μPo²and V̇o^2,intdecreased significantly from baseline values, which not only supports the notion that now the intestinal tissue oxygenation had become impaired but also indicates that in this study, p  H^mvwas a more reliable indicator of intestinal tissue oxygenation than P^mvo².

One might argue that failure to maintain isovolemia throughout the experiment might explain these observations as well. In the absence of blood volume measurements, only supporting evidence could be provided regarding the maintenance of isovolemia during hemodilution, i.e. , the constant central venous pressure during hemodilution and the lack of effect of the subsequent fluid challenge on central venous pressure, MAP, and μPo²in four additional animals. In addition, a significant increase in P^pulsin the hemodilution group, the comparable urine production in the hemodilution and the control groups, the use of a slightly hyperoncotic protein solution, and a slight increase in total body weight of the animals during the experiment suggest that the animals in the hemodilution group were not hypovolemic throughout the experiment. Therefore, the hemodynamic parameters and the determination of a critical level of Do²for the intestinal V̇o²and μPo²in this model were not likely to be biased by an insufficient circulating volume.

Because of blood sampling and subsequent volume correction, a small but significant degree of hemodilution took place in the control group, resulting in a decrease in Do^2,int. V̇o^2,intand μPo²were preserved by an increase in O²ER from 20% at baseline to 29% at H4. This was only reflected in the decrease in S^cvo²; any change in the other parameters did not reach significance. The results from the control group demonstrated that the hemodynamic parameters were stable throughout the entire experiment.

In contrast with our results are the observations of Haisjackl et al. , 34who observed no significant change in pig V̇o^2,intuntil a systemic hematocrit of ± 6% was reached, as a result of a doubling in intestinal O²ER. This suggested an uncompromised microcirculatory oxygenation at an extremely low hematocrit. However, surface oxygen electrode measurements in their study demonstrated that mucosal Po²was preserved until a hematocrit of 6% was reached and that serosal Po²decreased below a hematocrit of 15%. The measurement depth of the Pd-porphyrin phosphorescence is in the order of 0.5 mm. 17Therefore, the size of the rat small intestine did not allow us to distinguish between a mucosal and a serosal μPo², as would be possible when using Pd-porphyrin phosphorescence in the pig. 17,35Thus, the intestinal μPo²values in our study, although in the same range as pig serosal Po²values measured with either oxygen electrode or Pd-porphyrin phosphorescence, must be considered as a resultant of the μPo²values in the different layers of the gut wall. As such, they are hardly comparable to μPo²values found in the pig. Furthermore, it has been shown that oxygen electrodes are sensitive to changes in arterial oxygen content. 9,17The use of the oxygen-dependent quenching of Pd-porphyrin phosphorescence might be more advantageous as this technique has a deeper penetration depth and, used in combination with an optical fiber, selectively measures the Po²in the capillaries and the venules. 16,17The discrepancy in intestinal V̇o²might be explained by the interspecies difference (pig vs.  rat), or by the fact that Haisjackl et al.  34used an isolated gut segment, which was perfused at a constant pressure, whereas in our experiments the intestines were not isolated and were perfused at the MAP, which decreased significantly below hematocrit 15%. This is supported by the results from Nöldge et al. , 36who found a decrease in both MAP and liver and small intestine surface Po²values at and below a hematocrit of 15% during normovolemic hemodilution. In this context, it must be realized that for an adequate interpretation of the results of hemodilution experiments, the use of anesthesia must be taken into account: the compensatory mechanisms during hemodilution are influenced by the use of anesthesia and a critical level of Do²could not be demonstrated during hemodilution in conscious dogs and humans. 37–39 

In conclusion, a critical point for the intestinal tissue oxygenation could be identified during normovolemic hemodilution. It was demonstrated that μPo²and V̇o^2,intwere limited by oxygen supply at the same time when the intestinal Do²decreased to a critical level. As soon as the intestinal oxygenation became dependent on oxygen supply, p  H^mvdecreased and an increasing Po²gap between the venous and microvascular blood could be demonstrated, reflecting shunting of oxygen from the microcirculation. At this point, a significant correlation between μPo²and V̇o^2,intcould be demonstrated, but not between P^mvo²and V̇o^2,int. These data indicate that as soon as the μPo²is impaired during hemodilution, the tissue V̇o²is affected as well. Unlike the regional venous p  H, venous Po²cannot be regarded as a reliable parameter for the judgment of the intestinal tissue oxygenation.