Transfusion with allogeneic blood may be diminished by predeposited autologous blood, intraoperative autotransfusion with a cell saver and hemodilution techniques (1), and many other mea-sures used as parts of an integrated blood conservation program (2).
Immediate preoperative isovolemic hemodilution (IVH) has been used extensively in all types of surgeries, and its safety has been demonstrated in all age groups (3–5). The concept of preoperative hypervolemic hemodilution (HVH) is relatively new (1,6). It involves in vivo dilution of erythrocytes by creating hypervolemia with asanguinous fluids in an attempt to reduce erythrocyte loss. There is very little literature regarding the evaluation of HVH and its comparison with IVH (1,6–8), and no study has used polygeline as an agent for HVH.
This study compared IVH and HVH, under a standardized protocol of predefined levels of hemodilution and hematocrit (Hct), with respect to their efficacy to reduce the requirement of allogeneic blood in the perioperative period. The hemodynamic effects and overall time and cost requirements of the two techniques were also compared.
After approval by the Institutional Ethics Committee, this randomized, prospective study was conducted on 30 ASA physical status I and II patients of either sex. The patients were posted for elective surgery for which the expected blood loss was >500 mL. The patients were randomly allocated to either an IVH group (n = 15) or an HVH group (n = 15).
All patients underwent routine examinations as determined by age, sex, and type of surgery. Specific examinations included Hct, prothrombin time with control, chest radiograph, electrocardiogram, and liver and kidney function tests. Patients having Hct <30%, coagulation disorder, or evidence of uncontrolled cardiac, hepatic, and renal disease were excluded from the study. Written, informed consent was obtained.
Patients were premedicated with IV meperidine 0.5 mg/kg and promethazine 0.25 mg/kg, 15 min before the induction of anesthesia with IV thiopental 5 mg/kg. Tracheal intubation was facilitated by IV pancuronium 0.1 mg/kg. Anesthesia was maintained with N2O and halothane in oxygen, and muscle paralysis was maintained with 0.02 mg/kg top-up doses of pancuronium. Supplemental doses of IV meperidine (10 mg) were given every hour. At the end of anesthesia, the residual muscle paralysis was reversed with IV neostigmine 60 μg/kg and atropine 20 μg/kg.
Monitoring consisted of heart rate (HR), noninvasive arterial blood pressure (systolic [SAP], diastolic [DAP], and mean [MAP]), Spo2, electrocardiogram, ETco2, central venous pressure (CVP), and urine output. The reading after the induction of anesthesia, tracheal intubation, placement of the CVP catheter (in internal jugular vein except in two cases of neck dissection, in which arm veins were cannulated) and urinary catheter and just before starting hemodilution was taken as the baseline. The variables were recorded every 5 min during hemodilution and then every 10 min. Urine output was monitored every half hour. Blood loss estimation was based on measurement of suction bottle collection and blood in the suction tubing, visual assessment of loss around the surgical site, and weighing of sponges and wash fluid.
Hct was measured by centrifugation method at the time of establishing the preoperative IV catheter (H1), at the end of the hemodilution procedure (H2), and when the allowable volume of blood had been lost, surgery had ended, or bleeding had stopped (whichever was earlier) (H3). H4 was measured in the IVH group when the whole volume of withdrawn blood had been transfused back and 3 h after stopping the polygeline infusion in the HVH group (polygeline has an average intravascular residence of 3 h). H5 was measured 24 h after surgery.
In the IVH group, autologous blood was collected in citrate-phosphate-dextrose-adenine bags, numbered in order of withdrawal, labeled with the patient’s name and hospital record number, and stored in the operation room (OR). The blood volume to be withdrawn to the aimed postdilution Hct (H2) of 25% was calculated by using the formula suggested by Gross (9). Simultaneously, polygeline was infused to match the volume of blood withdrawn, i.e., approximately 100 mL of polygeline for every 80 mL of blood withdrawn. The total time taken for the procedure, from wheeling the patient into the OR to handing him or her over to the surgeons, was noted.
In the HVH group, polygeline was infused at a constant rate of 100 mL/min to create a temporary state of hypervolemia and decrease the Hct (H2) to 25%. The volume of colloid required was calculated by the following formula:MATH where EBV is the estimated blood volume, Ho is the original (preoperative) Hct, and Hf is the final (postdilution) Hct—H2 (25%) in this study. The procedure was timed and the patient handed over to the surgeons.
In both groups, the allowable blood loss (ABL) during surgery was calculated to the final Hct of 20% by using Gross’s formula (9). The blood loss was replaced with polygeline to maintain the postdilution blood volume and the rest of the requirements were replaced with crystalloids as usual.
In the IVH group, if bleeding stopped before or at the ABL, autologous blood was started 3 h after stopping the polygeline infusion being given to replace the ABL. If bleeding continued beyond the ABL during surgery, autologous blood (in reverse order) was started during surgery. If the Hct after the completion of autologous blood transfusion (H4) was <25%, it was corrected up to 25% by infusing allogeneic blood.
In the HVH group, if bleeding stopped before or at the ABL, the polygeline infusion being given to replace ABL was stopped. H4 was determined after 3 h, and allogeneic blood was given if H4 was <25%. In case bleeding continued beyond ABL, the blood loss beyond ABL was replaced with allogeneic blood. Three hours after the polygeline infusion had been discontinued, Hct (H4) was measured and more allogeneic blood given if necessary to achieve a postoperative Hct of 25%. In both groups, Hct (H5) was repeated the morning after surgery. Postoperative blood loss and any significant events were noted.
The cost incurred was determined by taking into account the amount of colloid used, IV cannulae used, blood collection bags used to collect autologous blood, infusion sets required, three-way connectors used, and allogeneic blood needed. The cost of allogeneic blood was as per the assessment by the blood bank of the hospital, and it included the total expenditure from the donor to the recipient at the hospital rate.
The sample size of this study was decided on the basis of our previous unpublished study with an IVH group and a control group of 15 patients each, with similar standardizations. We found that IVH could help save approximately 600 mL (±200 mL) of allogeneic blood per patient during surgery. We hypothesized that HVH should also save at least 350 mL (±200 mL) per patient under similar circumstances. With these assumptions, an α of 0.05, and a power of 90% to detect the assumed differences, we needed 14 patients per treatment arm.
All data are expressed as median (interquartile range; IQR) or mean ± sd. To study the pattern of changes in individual hemodynamic variables in each group during various phases of study, an analysis of variance for repeated measures was performed. When the F value showed a significant difference (P < 0.05), a Student-Newman-Keuls test for all possible comparisons was performed to detect differences among phases (baseline, hemodilution, surgery, and postoperative). The hemodynamic data at various phases of study were compared between the two groups by using two-tailed unpaired Student’s t-tests. Blood loss and allogeneic blood given/saved were compared by using the nonparametric Mann-Whitney U-test and by 95% confidence intervals (CIs) for between-group differences.
To compare the volume of polygeline used, Hct at various phases of measurement, time taken to perform the procedure, and cost incurred in the two techniques and to compare the Hct actually reached after hemodilution (H2) and the target Hct of 25% in the two groups, two-tailed unpaired Student’s t-tests and 95% CIs for between-group differences were used. Where appropriate, P < 0.05 was considered significant.
The two groups were comparable in age, sex, and weight of the patients (P > 0.05). The age ranged from 18 to 60 yr (Table 1). These patients underwent orthopedic, general, or ear-nose-throat surgery (Table 2).
The SAP, DAP, and MAP did not alter significantly in either group throughout all the phases of study (i.e., baseline, hemodilution, surgery, and postoperative) (P > 0.05). Within the IVH group, the HR was significantly slower in the postoperative period as compared with the other 3 phases of the study (P < 0.05). Within the HVH group, the HR was also significantly slower in the postoperative period compared with the baseline and hemodilution phases (P < 0.05). Although there were no significant alterations in the CVP during any phase in the IVH group (P > 0.05), CVP was significantly higher during the HVH and surgery phases (P < 0.05) in the HVH group (Table 3).
The baseline hemodynamic variables in the two groups (HR; SAP, DAP, and MAP; and CVP) were comparable (P > 0.05). During hemodilution, although SAP and CVP were significantly higher in the HVH group than the IVH group (P = 0.0107 for SAP;P = 0.0281 for CVP), other hemodynamic variables were comparable (P > 0.05). All the hemodynamic variables were comparable between the two groups during the rest of the study periods (P > 0.05) (Table 3).
The preoperative Hct (H1) (mean difference, −0.1%; 95% CI, −1.7 to 1.5) and postdilution Hct (H2) (mean difference, −0.6%; 95% CI, −1.5 to 0.3) were comparable in the two groups. The lowest Hct reached (H3) was significantly lower in the IVH group than the HVH group, with a mean difference of −3.2% (95% CI, −5.8 to −0.6). The postoperative nearly steady-state Hct, H4 (mean difference, −1.1%; 95% CI, −3.5 to 1.3), and 24-h postoperative Hct, H5 (mean difference, −0.1%; 95% CI, −1.8 to 1.6), were comparable in the 2 groups. There was no significant difference between the postdilution Hct (H2) achieved and the target Hct of 25% in either the IVH (mean difference, 0%; 95% CI, −0.7 to 0.7) or the HVH (mean difference, −0.6%; 95% CI, −0.1 to 1.3) group.
Blood loss was similar in the two groups, with a mean difference of 290 mL (95% CI, −325 to 905 mL) (Table 1). To achieve the target minimum postoperative Hct of 25%, 6 of 15 (40%) patients in each group required allogeneic blood. The actual amount of allogeneic blood used in the two groups (median volume [IQR], 0 mL [0–475 ml] in the IVH group and 0 mL [0–525 mL] in the HVH group) was also similar, with a mean difference of −7 mL (95% CI, −326 to 312 mL) (Table 1).
The amount of allogeneic blood that the patient would have required (projected) to achieve the observed final Hct the day after surgery (H5) without use of the hemodilution technique but with same transfusion triggers was calculated (Appendix 1). This was comparable in the two groups (median [IQR], 781 mL [355–1174 mL] in the IVH group and 622 mL [394–976 mL] in the HVH group), with a mean difference of 199 mL (95% CI, −161 to 559 mL), and was significantly more than the actual amount of allogeneic blood required in either group (mean difference [95% CI] of −581 mL [−753 to −409] for IVH and −376 mL [−531 to −221 mL] for HVH). Allogeneic blood saved (projected − actual) was similar in the two groups, with a mean difference of 206 mL (95% CI, −2 to 414 mL) (Table 1).
The time taken to perform IVH and HVH was not significantly different (mean difference, 7 min; 95% CI, −0.5 to 14.5 min). Costs of performing IVH and HVH were comparable (mean difference, $1.70; 95% CI, −$4.10 to $7.50). The volumes of polygeline used in IVH and HVH patients were not significantly different (mean difference, −6 mL/kg body weight; 95% CI, −16 to 4 mL/kg body weight) (Table 1). No serious adverse effects were noted in any of the patients.
Acute preoperative IVH has been in use for a long time to reduce the requirement of banked homologous blood, but concerns have been raised regarding its efficacy (5,10). These concerns are mainly due to the lack of large, well controlled, randomized studies with clearly predefined transfusion triggers and control of perioperative transfusion regimens (11). The technique of acute preoperative HVH is relatively new (1,6,7), and its efficacy in avoiding perioperative allogeneic blood transfusion also needs proper evaluation.
Our study, although conducted on small number of patients, standardized Hct after hemodilution (25%) at the lowest (trigger) level (20%) and in the postoperative period (≥25%), thereby standardizing the perioperative regimen. Similar standardized reductions in Hct have been used recently during IVH (12), but there is no study on HVH in which standardized perioperative Hct has been attempted. Mielke et al. (6) did not correlate the procedure with the Hct values of the patients.
We hypothesized that IVH has a significant potential for reducing the need for allogeneic blood in the perioperative period without any significant hemodynamic effects and that HVH would be as effective as IVH, with minimal hemodynamic effects in ASA I/II patients, and would require less time and money. This study revealed that the two techniques were comparable, not only in the amount of allogeneic blood saved, but also in the time and cost requirements (Table 1). The significant transient hemodynamic alterations at the beginning of acute hypervolemia were well tolerated by the patient population studied.
The hemodilution technique is most effective when the surgical blood loss exceeds 1000 mL (6). Our cutoff limit of >500 mL is similar in terms of blood loss per kilogram body weight (10 mL/kg body weight), although eventually most (73.33%) of our patients bled more than 1000 mL.
To calculate the amount of volume expansion required to achieve a particular postdilution Hct during HVH, we devised a formula and hypothesized that the blood volume would have to expand by a factor of Ho/ Hf if the final Hct of Hf has to be reached from the original Hct of Ho. Thus, the volume to be added must be:MATH
To induce the volume expansion with polygeline (an expansion factor of nearly 80%), the volume required would be 100/80 (1.25) times the intended volume expansion:MATH
The postdilution Hct achieved with this formula was comparable to the target Hct of 25% (mean difference, 0.6%; 95% CI, −0.1% to 1.3%).
The lowest Hct reached in the IVH group was significantly lower than in the HVH group (Table 4). This difference was there because most (9 of 15) patients in the HVH group never bled to the maximum ABL during surgery, and their lowest Hct reached (H3) was significantly more than the minimum allowed of 20% (P = 0.0017). In the IVH group, H3 was near the lowest allowed (P = 0.93).
It was not the difference in the blood loss in the two groups (blood loss was similar;Table 1) that caused this difference in H3, but rather the increase in the ABL in the HVH group after hypervolemia. In the HVH group, Vo increased to Vo ×Ho/ Hf after hypervolemia, and because this hypervolemia was sustained throughout surgery, ABL was calculated on this increased volume (Table 5). It is obvious that the patients likely to escape allogeneic blood in the HVH group are those who either do not reach or just reach the lowest Hct during surgery. Of the latter, those who achieve an Hct of 25% three hours after stopping polygeline (when hypervolemia has settled) also escape exposure to allogeneic blood. Thus, the increased ABL in the HVH group and the availability of autologous blood in the IVH group finally lead to a similar number of patients (9 of 15) in either group who escape exposure to allogeneic blood and a similar amount of allogeneic blood used in the two groups (Table 1).
Several studies on IVH (3–5) and HVH (1,6–8) have commented on hemodynamic stability during the procedures. Our study also demonstrated hemodynamic stability in both groups, except for neostigmine-induced bradycardia in the immediate postoperative period in both groups, and hypervolemia induced a significant increase in CVP and SAP in the HVH group. As noted in a previous study, even these changes reverted quickly, reflecting adjustments due to increased ventricular end-diastolic volume, decreased end-systolic volume, decreased viscosity, and decreased systemic vascular resistance (8). Unlike the present study, a study (6) comparing IVH and HVH found a significant decrease in the SAP, DAP, and MAP after hemodilution in the IVH group. These authors considered the preinduction values of hemodynamic variables as baseline, whereas we, like some others (8), took the postinduction values as baseline. In addition, our previous experience with IVH made us wary of possible hypovolemia during blood withdrawal because of withdrawal-infusion mismatch.
(Appendix 1) shows the calculations for determining the amount of blood saved by either procedure. One may question the concept of calculation to the level of actual H5 rather than a predecided minimum level of postoperative Hct, say, 25%. Our reason was based on the simple premise that whatever H5 was attained, it was due to the hemodilution technique and the allogeneic blood transfused. Thus, if the value of H5 was more than our target level of 25%, the credit must go to the hemodilution technique, and the discredit must go to our method of replacing blood loss beyond the ABL (giving the volume of allogeneic blood [Hct = 40%] equal to the dilute blood [Hct ∼20%] lost by the patient at that stage). Some researchers have suggested and used (12,13) mathematical models to estimate blood saved. In agreement with these authors, we also divided the study period into phases for the purpose of these calculations and made corrections for the differences in the target and actual postoperative Hct. But, unlike them, we did away with the exponential-based equations and used the equation used by Gross (9) and derivations thereof and made calculations for the whole blood, taking care to make corrections for the changing Hct of the patient and the constant average Hct of allogeneic blood (taken as 40% for our blood bank) (Appendix 1). By this method, we found the two techniques comparable in allogeneic blood conservation, as well as in terms of allogeneic blood actually given in the two groups (6) (Table 1). Including a control group in the study going through the same transfusion trigger of 20% without any hemodilution technique would have been more appropriate in assessing allogeneic blood conservation by either technique.
We could not detect any significant time advantage of HVH over IVH (Table 1), as observed by some authors (6), because unlike them we did not include the reinfusion (of withdrawn blood) time as part of the IVH technique, and, in most surgical procedures, even IVH, like HVH, can be performed along with preliminary surgical preparation. The familiarity of our team with IVH was a big help.
Although HVH has been described as a less expensive option in the past (6), it was not so in our case (Table 1), probably because of surprisingly low rates of the material used. The processing costs of the allogeneic bank blood also did not affect the comparative total cost, because the amounts of allogeneic blood required were comparable. The cost of OR time was not included, because we did not have any such data available.
Maintenance of target blood volumes (normovolemia or hypervolemia) is considered vital during IVH and HVH (3,8). We chose polygeline because we considered its intravascular residence of three hours to be optimal for either procedure. For IVH, in most cases it starts leaving the circulation when the reinfusion of withdrawn blood begins, thus avoiding periods of hypervolemia. During HVH, it maintains hypervolemia and hemodilution for the time when most of the bleeding occurs, and the effect does not spill over to the postoperative period. In addition, it has a minimal effect on hemostasis and no cellular dehydrating property, and of the gelatin preparations, polygeline has a relatively less frequent incidence of anaphylactoid reaction (14). Although rapid hypervolemia with hydroxyethylstarch (6) and dextran-40/lactated Ringer’s solution (8) in the past and polygeline in this study have been found to be safe in healthy patients, that may not be so in patients with cardiac and autonomic nervous system disorders who cannot make compensatory adjustments in the face of acute hypervolemia. It must also be mentioned that in patients with impaired renal function, the excretion of polygeline might be delayed, thereby prolonging its intravascular residence time. The volumes of polygeline used (Table 1) were much larger than the earlier recommendation of 20 mL/kg body weight, but more recent literature has increased this limit (15,16). We must admit that ours was a relatively small study on healthy adults and that it was not supported by adequate testing for coagulation status.
The power analysis at the beginning of the study suggested 14 patients per treatment arm for 90% power, but at the end, the power of the study was 41%. This reduced power was because of less difference in the mean (206 mL instead of the 250 mL presumed) and wide variability in the volume of allogeneic blood saved. In retrospect, a study with 40 patients per treatment arm would give a power of 80%.
We did not apply a standardized method to calculate blood loss (12,13), which presumes that the blood volume of the patient remains constant and that the Hct changes are proportional to the erythrocyte volume change. We believe that, for our study, involving both isovolemia and hypervolemia, it would have been more appropriate to determine the blood volumes at two time points between which the blood loss needed to be calculated, as performed by some authors (16). Because we lacked the facilities to measure blood volume, we based our calculations on the observed and measured blood loss (6), although this added an element of subjective estimation. We quantified the visual assessment for the losses around the surgical site not accounted for by objective methods, on the basis of 1) the severity of capillary oozing as jointly decided by the anesthesiologist and the surgeon (16), 2) the area of the bleeding surface, and 3) repeated observations by the same person every 5–10 minutes, depending on the severity of the bleeding. We did not feel the need to use an observer blinded to the hemodilution method, because we were cross-checking these estimations by Hct measurements.
The Hct estimation was slow and took an average of 15 minutes. Thus, interventions based on the Hct level during surgery sometimes had to start on “guessed” Hct until the report became available. We tried to reduce this by drawing samples a few minutes ahead of the anticipated time of reaching the target Hct and then repeating the measurements at the deemed actual points of reaching the target Hct.
This relatively small study shows that both IVH and HVH are viable, safe, and comparable options for significantly reducing the perioperative requirement of allogeneic blood in ASA I/II adults and deserve validation by larger studies. IVH works by holding erythrocytes outside the patient’s body while he or she bleeds diluted blood, and HVH works by diluting the patient’s erythrocytes within his or her body by temporarily expanding the blood volume and increasing the ABL. The HVH formula we proposed was valid in this investigation and warrants further study in other populations.
We thank VK Bhatia, PhD for statistical help, CK Dua, MD for departmental support, and Manoj Sharma, MD for participating in our previous study that paved the way for the present study.