Enhanced formation of rapidly labelled bilirubin by phenobarbital: hepatic microsomal cytochromes as a possible source

Reversibility of acute intermediate phase bilirubin encephalopathy

AIM

To show the potential for reversing acute intermediate to advanced phase bilirubin encephalopathy.

METHODS

Case studies.

RESULTS

Six extremely jaundiced infants had symptoms of intermediate to advanced phase acute bilirubin encephalopathy. The infants were treated aggressively. Two patients had brain magnetic resonance imaging showing increased signals in the globus pallidus. On follow-up, all infants are neurologically normal.

CONCLUSIONS

Intermediate-to-advanced stage acute bilirubin encephalopathy may occasionally be reversible. These cases provide a strong argument in favour of rapid and aggressive intervention in infants presenting with extreme jaundice and neurological symptoms.

Induction of a dose-related increase in sulfobromophthalein uptake velocity in freshly isolated rat hepatocytes by phenobarbital

To determine whether phenobarbital affects hepatocellular bilirubin/sulfobromophthalein uptake mechanism, we administered it to male Sprague-Dawley rats, body weight 175 +/- 25 gm, at doses of 1 to 75 mg/kg body wt/day for 7 days. Control rats were given an equivalent volume of physiological saline solution. On day 8, hepatocytes were isolated by means of collagenase perfusion, suspended in Hanks' solution without albumin and incubated with high specific activity (3 Ci/mmol) [35S]sulfobromophthalein, which was synthesized in our laboratory and purified by means of a new reverse-phase high-pressure liquid chromatography procedure. The initial uptake rate of sulfobromophthalein was determined at sulfobromophthalein concentrations of 1 to 50 mumol/L with a rapid filtration technique. The maximum uptake velocity and Michaelis constant for sulfobromophthalein uptake at each phenobarbital dose were determined by means of a computer analysis. In control studies, maximum uptake and Michaelis constant were 48.0 +/- 16.7 (mean +/- S.D.) pmol/50,000 cells/min and 22 +/- 4 mumol/L, respectively. Maximum uptake velocity increased linearly with the log of the phenobarbital dose (r = 0.98, p < 0.01), the increase achieving statistical significance at a dose of 3 mg/kg/day. Michaelis constant, however, was essentially unchanged at phenobarbital doses of 50 mg/kg/day or less. The maximal observed increase in maximum uptake velocity of sulfobromophthalein (to 619% of control values) was appreciably greater than the maximal increase in UDP-glucuronyltransferase activity (200% of control) or immunoreactive ligandin concentrations (260% of control) seen in earlier studies, suggesting a direct effect on the plasma membrane transport mechanism.

Chemistry and biology of heme. Effect of metal salts, organometals, and metalloporphyrins on heme synthesis and catabolism, with special reference to clinical implications and interactions with cytochrome P-450

Although free porphyrins occur in nature in small quantities, no known function has been assigned to them. In contrast, heme and cobalamin, which are Fe and Co chelates of porphyrins or porphyrin derivatives, respectively, carry out crucial biological functions. Heme is the prosthetic group for a number of hemoproteins. These include myoglobin and hemoglobin, which carry out oxygen binding or transport; mitochondrial cytochromes aa3, b, c, and c3, which are important in transferring electrons; microsomal cytochrome P-450, which catalyzes mixed-function oxidations; catalase, which decomposes H2O2; peroxidase, which activates H2O2; and tryptophan pyrrolase, which catalyzes the oxidation of tryptophan. Recently, heme has also been shown to be the prosthetic group of prostaglandin and peroxide synthetase and indoleamine dioxygenase. The elegant studies of the biochemical pathway for the formation of heme demonstrated the arrangement in the porphyrin macrocycle of the carbon and nitrogen atoms originating from the eight glycine and the succinic acid molecule that are the precursors of porphyrins. There are eight enzymes involved in the synthesis of heme. The first and last three of these enzymes are localized in mitochondria, while the intermediate enzymes are localized in cytosol. The catalytic site of HMOX recognizes metalloporphyrins with central metal atoms other than iron; it favors some of these metalloporphyrins over heme as a potential substrate, sometimes by a large factor, permitting the synthetic heme analogue to serve as a potent competitive inhibitor of HMOX reaction. Since these synthetic metalloporphyrins do not bind molecular oxygen, they are not metabolically degraded by ring rupture and do not add to the body pool of bile pigment. One possible consequence of this competitive inhibition of heme degradation is suppression of bile pigment formation to such a degree that excessive plasma levels of bilirubin may be diminished. The studies of Drummond and Kappas (1981) and later studies in rats, mice, monkeys, and man, and also our studies have proved the latter phenomenon. The compound does not appear to affect the metabolic disposition of preformed bilirubin but inhibits biliary bilirubin excretion derived from the metabolism of endogenous or exogenous heme. Whether some of the effect of Sn-PP on naturally occurring or experimentally induced jaundice in animals reflects diversion of heme to nonheme to oxygenase-dependent pathways of heme metabolism, or whether a pathway which is normally latent becomes activated concurrent with HMOX inhibition is not known.(ABSTRACT TRUNCATED AT 400 WORDS)

Studies on the mechanism of Sn-protoporphyrin suppression of hyperbilirubinemia. Inhibition of heme oxidation and bilirubin production

The synthetic heme analogue Sn-protoporphyrin is a potent competitive inhibitor of heme oxygenase, the rate-limiting enzyme in heme degradation to bile pigment, and can entirely suppress hyperbilirubinemia in neonatal animals and significantly reduce plasma bilirubin levels in a variety of circumstances in experimental animals and man. To further explore the mechanism by which this metalloporphyrin reduces bilirubin levels in vivo, we have examined its effects on bilirubin production in bile duct-cannulated rats, in which bilirubin derived from heme catabolism is known to be rapidly excreted in bile. The administration of Sn-protoporphyrin (10-50 mumol/kg body weight) was followed by prompt (within approximately 1 h) and sustained (up to at least 18 h) decreases in bilirubin output, to levels 25-30 percent below the levels of bilirubin output in control bile fistula animals. The metalloporphyrin had no effect on bile flow or the biliary output of bile acids. Infusions of heme, which is taken up primarily in hepatocytes, or of heat-damaged erythrocytes, which are taken up in reticuloendothelial cells, resulted in marked increases in bilirubin output in bile in control animals; these increases were completely prevented or substantially diminished by Sn-protoporphyrin administration. By contrast, the metalloporphyrin did not alter the high levels of bilirubin in plasma and bile that were achieved in separate experiments by the constant (16 h) infusion of unconjugated bilirubin to bile duct-cannulated rats. Thus, Sn-protoporphyrin exerts no major effects on the metabolic disposition of preformed bilirubin. Heme oxygenase activities were markedly decreased in microsomal preparations from liver, spleen, and kidneys in these experiments, to a degree comparable to the decreases we have observed in the intact rat. We also demonstrated that a substantial proportion (19-35%) of a dose of Sn-protoporphyrin is promptly excreted in bile and that the time course of biliary excretion of this compound more closely reflects plasma concentrations of the metalloporphyrin, which decline rapidly, rather than concentrations in liver, which are considerably more persistent. These results indicate that Sn-protoporphyrin substantially reduces the in vivo production of bilirubin from the degradation of endogenous as well as exogenous heme in the rat. Moreover, this inhibitory effect of the synthetic metalloporphyrin on bilirubin production occurs in both hepatocytes and reticuloendothelial cells, which are the major tissue sites for bilirubin formation. In other studies, we have established that heme oxygenase blockade by Sn-protoporphyrin leads to a marked and rapid excretion of heme into bile presumably because the synthetic metabolism to bile pigment and making it available for excretion via the biliary system in to the gut, These studies strongly suggest that Sn-protoporphyrin diminishes hyperbilirubinemia in animals and man by inhibiting the production of the bile pigment in vivo, and that its principal mode of action involves a potent and sustained competitive inhibition of heme oxygenase.

Effect of low-dose phenobarbital on hepatic microsomal UDP-glucuronyl transferase activity

To determine whether hepatic microsomal enzyme induction occurs in rats following administration of phenobarbital at doses similar to those used in humans (0.5 to 7.5 mg/kg), UDP-glucuronyl transferase (UDPGT) and cytochrome P-450 activities were measured in liver homogenate and microsomal preparations from control rats and rats treated for 6 days with phenobarbital at 1 and 3 mg per kg per day. While no significant increases in liver weight and protein content of homogenate and microsomal preparations were observed with either dose of the drug, both UDPGT and P-450 activities were enhanced significantly following administration of phenobarbital at 3 mg per kg per day. The activity of P-450 was increased by approximately 30% and that of UDPGT by 15-24 and 45-66%, respectively, employing bilirubin and p-nitrophenol as the acceptor substrate. The extent of induction of bilirubin or p-nitrophenol UDPGT was similar when measured with "native" enzyme or with enzyme activated by UDP-N-acetyl glucosamine, digitonin or deoxycholate. These data suggest that the discordant effects of phenobarbital on UDPGT and cytochrome P-450 previously reported in humans and rats may not be attributable solely to differences in the drug doses employed.

Increases in cytochrome p-450 in cultured hepatocytes mediated by 3- and 4-carbon alcohols

The amount of cytochrome P-450 was increased to different extents after treatment of cultured chick embryo hepatocytes with n-propanol, isopropanol, n-butanol, or isobutanol. These increases were associated with increases in benzphetamine demethylase activity, a cytochrome P-450-catalyzed oxidation, and glucuronidation of phenol red, catalyzed by UDP-glucuronyl transferase. The responses were similar to those obtained with ethanol or propylisopropylacetamide, which the phenobarbital-like inducers. Pretreatment of cells with cycloheximide prevented the increases in both cytochrome P-450 and glucuronidation of phenol red, indicating that protein synthesis was required for these responses.

The mechanism of haem catabolism. Bilirubin formation in living rats by [18O]oxygen labelling

1. The pathway of haem breakdown in living rats was studied by using 18O in the oxygen that the animals consumed. By cannulation of the common bile duct and collection of bile, labelled bilirubin was isolated and its mass spectrum determined. One set of results was obtained for a rat to which haemoglobin had been intravenously administered and another set obtained for a rat that was not given exogenous haem. Isomerization of bilirubin IXalpha to the XIIIalpha and IIIalpha isomers did not occur to any significant extent. The 18O-labelling pattern obtained in the bilirubin was consistent with a Two-Molecule Mechanism, whereby the terminal lactam oxygen atoms of bilirubin are derived from different oxygen molecules. The consequences of this mechanism are discussed in terms of the possible intermediates of the catabolic pathway. 2. 18O-labelled bilirubin appeared in the bile in less than 10 min after exposure of the animals to labelled oxygen. This result suggests that all of the chemical transformations involving production of biliverdin, reduction to bilirubin and conjugation of the bilirubin are fast processes. 3. The quantitative recovery of label obtained in the experiments suggests that there is little or no exchange of newly synthesized bilirubin with existing bilirubin pools in the animal.

Hyperbilirubinemia and cholestasis

Although the morphologist continues to describe cholestasis on the basis of precipitated bile seen on light microscopic sections of the liver or dilated canaliculi with loss of microvilli seen by electron microscopy, the physiologist can distinguish clearly between hyperbilirubinemia and cholestasis. Both bilirubin and bile acids are specifically removed from sinusoidal plasma by the normal hepatocyte and appear in bile in high concentration. Bilirubin conjugation and excretion appear to be governed by hepatocellular mechanisms that are, for the most part, separate from the conjugation and excretion of bile acids. Disturbances in bilirubin transport are recognized by hyperbilirubinemia which represents a number of clinical syndromes that can be classified by the nature of the block in the transport system. Serum bile acids appear to remain normal in hyperbilirubinemic syndromes. By contrast, cholestatic syndromes are characterized by marked bile acidemia with normal to slightly elevated bilirubin levels. Severe cholestasis, because of the marked reduction in bile flow, can however, engender jaundice. Further exploration of these excretory pathways will provide interesting new insights on the numerous cholestatic and hyperbilirubinemic syndromes that occur in nature.

The synthesis and excretion of non-erythroid bilirubin in the dog

Between 10 and 20% of the bilirubin excreted in the bile is not derived from the breakdown of hemoglobin. When delta-aminolevulinic acid is given as a bilirubin precursor, 99% of the bilirubin formed is of this non-erythroid variety. The non-erythroid bilirubin has been supposed to be synthetized exclusively in the liver. Sequential samples of arterial, portal and liver vein blood, as well as of bile and thoracic duct lymph, were analyzed for bilirubin and non-bilirubin radioactivity following the injection of labeled delta-aminolevulinic acid in dogs. Radioactive bilirubin appeared in blood before it could be demonstrated in bile or lymph. The concentration differences between aortic, portal and liver vein blood displayed a considerable extrahepatic non-erythroid bilirubin synthesis. Most non-erythroid bilirubin synthetized in the liver is excreted directly into the biliary canaliculi.

Effect of hexachlorobenzene on haem synthesis

Several drugs are known to induce the liver microsomal mixed-function oxidase system when administered in vivo or even in vitro in cell culture. A sequence of events has been suggested in which the drug is visualized to induce delta-aminolaevulinate synthetase, the first and rate-limiting enzyme of the haem-biosynthetic pathway, which is followed by enhanced haem synthesis and cytochrome P-450 content, facilitating the increase in the drug-metabolizing activity of the liver microsomal fraction. The present studies show that the fungicide hexachlorobenzene, when administered to female rats, can lead to enhanced amounts and rate of synthesis of cytochrome P-450 under conditions when the rate of total haem synthesis has not appreciably altered. The subsequent increase in the rate of total haem synthesis as well as the initial increase in amounts of cytochrome P-450 are brought about under conditions when delta-aminolaevulinate synthetase activity remains constant. However, manifestation of porphyria due to prolonged drug administration is accompanied by a twofold increase in delta-aminolaevulinate synthetase activity. The increase in enzyme activity appears to be due to a decreased degradation rate of the enzyme.

The relationship between delta-aminolaevulinate synthetase induction and the concentration of cytochrome P-450 and catalase in rat liver

The porphyrinogenic drug 2-allyl-2-isopropylacetamide causes the degradation of microsomal cytochrome P-450 and inhibits the synthesis of catalase in rat liver. The inhibition of catalase synthesis follows the induction of delta-aminolaevulinate synthetase and the consequent overproduction of haem. The allylisopropylacetamide-mediated breakdown of cytochrome P-450 is a rapid event and has a reciprocal relationship to the pattern of delta-aminolaevulinate synthetase induction. Breakdown of cytochrome P-450 appears to be one of the conditions leading to the ;derepression' of delta-aminolaevulinate synthetase.

Metabolic regulation of heme catabolism and bilirubin production. I. Hormonal control of hepatic heme oxygenase activity

Heme oxygenase (HO), the enzyme system catalyzing the conversion of heme to bilirubin, was studied in the liver and spleen of fed, fasted, and refed rats. Fasting up to 72 hr resulted in a threefold increase in hepatic HO activity, while starvation beyond this period led to a gradual decline in enzyme activity. Refeeding of rats fasted for 48 hr depressed hepatic HO activity to basal values within 24 hr. Splenic HO was unaffected by fasting and refeeding. Hypoglycemia induced by injections of insulin or mannose was a powerful stimulator of hepatic HO. Glucose given together with the insulin abolished the stimulatory effect of the latter. Parenteral treatment with glucagon led to a twofold, and with epinephrine to a fivefold, increase of hepatic HO activity; arginine, which releases endogenous glucagon, stimulated the enzyme fivefold. These stimulatory effects of glucagon and epinephrine could be duplicated by administration of cyclic adenosine monophosphate (AMP), while thyroxine and hydroxortisone were ineffective. Nicotinic acid, which inhibits lipolysis, failed to modify the stimulatory effect of epinephrine. None of these hormones altered HO activity in the spleen. These findings demonstrate that the enzymatic mechanism involved in the formation of bilirubin from heme in the liver is stimulated by fasting, hypoglycemia, epinephrine, glucagon, and cyclic AMP. They further suggest that the enzyme stimulation produced by fasting may be mediated by glucagon released in response to hypoglycemia. The possibility is considered that the enhanced HO activity in the liver may increase hepatic heme turnover and hence, bilirubin production, which may explain the rise of unconjugated serum bilirubin observed in fasting or hypoglycemic individuals.

Bilirubin excretion in rats with normal and impaired bilirubin conjugation: effect of phenobarbital

The effect of phenobarbital on bilirubin excretion was studied in rats with different capacities for bilirubin conjugation. Drug treatment induced substantial increases in bilirubin UDP-glucuronyl transferase activity in the liver of both normal and heterozygous Gunn rats, but not homozygous Gunn rats in which enzyme activity is completely absent. However, enhancement of bilirubin excretion in vivo was observed only in heterozygous Gunn rats. In these animals the maximum capacity to excrete bilirubin into bile (T(max)), like the activity of the conjugating enzyme, was half normal; phenobarbital caused an increase in T(max) to levels characteristic of normal animals, with a twofold rise in the excretion of conjugated pigment. This appeared to be largely unrelated to enhancement of bile flow, and there was no stimulation of alternate pathways of bilirubin excretion. Conjugated bilirubin was consistently recovered from the plasma and urine of both untreated normal and heterozygous Gunn rats infused with unconjugated pigment. The quantities thus recovered comprised a similar fraction of the total pigment conjugated in both types of animal. Moreover, there were linear correlations between T(max) and both the rate of bile flow and the activity of the conjugating enzyme over the range of values represented by control rats of both types. These findings suggest that the process by which conjugated bilirubin is secreted into the bile is closely related to conjugation and limits the final excretory rate at different levels of pigment excretion. The phenobarbital effect uniquely observed in heterozygous Gunn rats appears to be mediated primarily by enhancement of the limited capacity for bilirubin conjugation with an associated rise in functional secretory capacity.

Inducible heme oxygenase in the kidney: a model for the homeostatic control of hemoglobin catabolism

We have recently identified and characterized NADPH-dependent microsomal heme oxygenase as the major enzymatic mechanism for the conversion of hemoglobin-heme to bilirubin-IXalpha in vivo. Enzyme activity is highest in tissues normally involved in red cell breakdown, that is, spleen, liver, and bone marrow, but it usually is negligible in the kidney. However, renal heme oxygenase activity may be transiently increased 30- to 100-fold following hemoglobinemia that exceeded the plasma haptoglobin-binding capacity and consequently resulted in hemoglobinuria. Maximal stimulation of enzyme activity in rats is reached 6-16 hr following a single intravenous injection of 30 mg of hemoglobin per 100 g body weight; activity returns to basal levels after about 48 hr. At peak level, total enzyme activity in the kidneys exceeds that of the spleen or liver. Cyclohexamide, puromycin, or actinomycin D, given just before, or within a few hours after, a single intravenous injection of hemoglobin minimizes or prevents the rise in renal enzyme activity; this suggests that the increase in enzyme activity is dependent on continued synthesis of ribonucleic acid and protein. The apparent biological half-life of renal heme oxygenase is about 6 hr. These observations indicate that functional adaptation of renal heme oxygenase activity reflects enzyme induction either directly or indirectly by the substrate, hemoglobin. Filtered rather than plasma hemoglobin appears to regulate renal heme oxygenase activity. Thus, stabilization of plasma hemoglobin in its tetrameric form with bis (N-maleimidomethyl) ether, which diminishes its glomerular filtration and retards it plasma clearance, results in reduced enzyme stimulation in the kidney, but enhances its activity in the liver. These findings suggest that the enzyme is localized in the tubular epithelial cells rather than in the glomeruli and is activated by luminal hemoglobin. Direct support for this concept was obtained by the demonstration of heme oxygenase activity in renal tubules isolated from rabbits that had been injected with hemoglobin.

Photocatabolism of labeled bilirubin in the congenitally jaundiced (Gunn) rat

To elucidate the mechanism by which phototherapy reduces serum bilirubin, studies were performed on the catabolism of labeled bilirubin in homozygous jaundiced Gunn rats before, during, and after a period of exposure to 1700 foot candles of daylight fluorescent light. Following equilibration with the body pool of an intravenously administered tracer dose of (3)H- or (14)C-bilirubin, radioactive and diazo reactive compounds were excreted in the bile at a slow, steady rate and plasma specific activity declined semilogarithmically. Subsequent exposure to light caused a marked increase in the biliary excretion of radioactive and diazoreactive compounds. Fecal and urinary radioactivity increased also but remained minor fractions of the total excreted radioactivity. After extinguishing the lights, these variables reverted gradually to control values. Spectral and chromotographic analysis of the excreted pigments and their azopigments demonstrated that the increased biliary radioactivity during phototherapy consisted of two roughly equal fractions: (a) unconjugated bilirubin, excreted at rates comparable to the output of conjugated bilirubin in the bile of normal nonjaundiced rats; and (b) water-soluble bilirubin derivatives, chromatographically identical with those found in Gunn rat bile under control lighting conditions but different from the products of photodecomposition of bilirubin in vitro. In some animals, phototherapy produced little decline in plasma bilirubin despite comparable acceleration of bilirubin catabolism. This was attributed tentatively to increased synthesis of early labeled bilirubin in these animals.

Hemolysis of "stress" reticulocytes: a source of erythropoietic bilirubin formation

The formation of bilirubin-(14)C was measured in rats given transfusions of red blood cells containing (14)C-labeled hemoglobin heme. Per cent conversion of hemoglobin-(14)C to bilirubin was 4 times greater with transfusion of "stress" reticulocytes from rats responding to hemorrhage than with normal reticulocytes from unstimulated donors. When the increased number of labeled reticulocytes produced by hemorrhaged donors was also considered, the total magnitude of labeled bilirubin formation was almost 20 times higher with stress as compared to normal reticulocytes. The findings were not influenced by splenectomy of either donor or recipient rats, iron loading of donors, or bleeding of recipients. However, bilirubin-(14)C formation fell off progressively as studies were performed at longer intervals after erythroid stimulation. Total bilirubin-(14)C formation in rats transfused with stress reticulocytes was compared to the production of early-labeled bilirubin from all potential sources in intact rats bled according to the same schedule used in the transfusion experiments. It is estimated that degradation of hemoglobin from sress reticulocytes accounts for virtually the entire rise in erythropoietic bilirubin formation from 24 to 96 hr after glycine-2-(14)C administration, but that additional sources make a major contribution before that time. These findings are consistent with the concept that destruction of immature erythroid cells in the peripheral blood, and probably in the bone marrow, accompanies the physiologic response to erythroid stimulation.

Catabolism of heme in vivo: comparison of the simultaneous production of bilirubin and carbon monoxide

The quantitative relationship between the catabolism of heme and the formation of bilirubin and carbon monoxide (CO) was studied in untreated rats and in animals treated with phenobarbital or the porphyrogenic drug, allylisopropylacetamide (AIA). A novel metabolic chamber permitting continuous collection of the bile and breath was utilized for quantitation of bilirubin-(14)C and (14)CO after the administration of hematin-(14)C or glycine-(14)C. After intravenous infusion of hematin-(14)C, control and phenobarbital-treated rats produced equimolar amounts of labeled bilirubin and CO; a minor fraction of the infused radioactivity appeared in the bile in other metabolites. The equimolar relationship in the formation of bilirubin and CO was also observed after pulse-labeling with glycine-2-(14)C; in phenobarbital-treated rats both metabolites were formed at an increased rate as compared to controls. By contrast, AIA treatment reduced the fractional conversion of hematin-(14)C to bilirubin and CO; a major fraction of the infused radioactivity appeared in the bile in metabolites other than bilirubin. In addition, in AIA-treated animals the molar CO/bilirubin recovery ratio was consistently greater than 1.0. Comparable results were obtained in AIA-treated rats after pulse-labeling with glycine-2-(14)C. These findings suggest that (a) in control and phenobarbital-treated rats infused hematin and heme formed in the liver are converted predominantly to bilirubin and CO, appearing in equimolar amounts; only a minor fraction of the hematin is degraded to other metabolites; (b) treatment with phenobarbital results in a proportional increase in the formation of both bilirubin and CO, reflecting increased heme synthesis and degradation in the liver; and (c) treatment with the porphyrogenic drug AIA shifts the CO/bilirubin ratio in favor of the gas, and enhances the formation of nonbilirubin metabolites.

Mechanically induced intravascular hemolysis in dogs

The endogenous production of carbon monoxide and the flow of hemoglobin to and from plasma were measured in 11 anesthetized dogs after pumping blood through an extracorporeal circuit for short periods. Two different pumps were used. In all animals the increase in CO production was greater than could be explained by catabolism of hemoglobin lost from plasma, an average of 11.4 times greater with one pump and 2.49 times greater with the other pump. Evidence is presented that this discrepancy could not be explained by catabolism of heme other than that of hemoglobin, and we therefore concluded that rates of hemoglobin catabolism were much greater than indicated by plasma hemoglobin kinetics and that extravascular hemolysis is a major cause of erythrocyte destruction during mechanically induced hemolysis. Extravascular hemolysis apparently caused an average of 72.9% and 37.2% (with the two pumps) of the total quantity of erythrocytes destroyed during pumping and for 3 hours after pumping.