Citrate Toxicity During Massive Blood Transfusion - Notify

Preparing to load PDF file. please wait...

0 of 0
Citrate Toxicity During Massive Blood Transfusion - Notify

Transcript Of Citrate Toxicity During Massive Blood Transfusion - Notify

Citrate Toxicity During Massive Blood Transfusion
Walter H. Dzik and Scott A. Kirkley

THE USE OF sodium citrate as the blood anticoagulant for transfusion science dates from 1914 to 1915 and the almost simultaneous publication of the work of four independent investigators. Articles from Hustin,l Agote,2 Weil,3 and Lewisohn,4 published between May 1914 and January 1915, each reported the successful use of citrated blood for human transfusion. Because citrate was known to be toxic to animals, Lewisohn carefully titrated the minimum concentration of citrate required to prevent clotting. Despite the demonstration by Weil that citrated blood could be stored for several days and still be effective,3 and the discovery by Rous and TurmneerrS5 that citrated blood supplemented by dextrose was capable of more prolonged storage, citrated blood was not quickly accepted by the general medical community. Though used with success in limited trialis on the battlefield in World War 1,6 transfusions with citrated blood were often associated with chills and fever which were incorrectly attributed to the citrate. Febrile reactions were so common with citrated blood that during the 1920s most transfusions consisted of the rapid transfers of nonanticoagulated whole blood, and the use of stored citrated blood did not become commonplace until the mid-1930s. 7
While improvements in transfusion technology and the establishment of blood banks made the administration of blood a standard procedure in the operating room, blood usage was generally limited to a few units for any given patient. Advances in knowledge of the biochemistry of citrate and calcium led to an improved understanding of their interaction as well as the relationship of serum
ionized Ca + + to total serum ca1lccium. The devel-
opment of citrate toxicity due to acute hypocaa1lcemia was demonstrated in dogs as early as 1944. 8 After Wordld War II surgical techniques of increas-
From the Blood Bank and Tissue Typing Laboratory, New England Deaconess Hospital, Department ofj Medicine, Harvard Medical School, Boston.
Address requests to Walter H. Dzik, MD, Blood Bank and Tissue Typing Laboratory, 185 Pilgrim Road, Boston, MA 02215.
© 1988 by Grune & Stratton, 1nc. 0887 796331/88/100220022--0004$03..00001/0

ing complexity required rapid transfusion of larger volumes of blood. In 1955 several cases of 'citric acid intoxication' following transfusion were reported by Bunker.9 In the years that followed, numerous investigations were published and considerable controversy developed regarding the management of citrate toxicity during massive transfusion. Recognition and therapy of citrate toxicity was enhanced by the widespread application in the 1970s of ion selective electrodes capable of accurate measurement of the level of ionized
Ca + +. With the development of advanced
trauma care, liver transplantation, and prolonged extensive surgical procedures in pediatrics, there has been a renewed interest in the role of citrate toxicity in the setting of ultramassive transfusion.
This review focuses on citrate toxicity during massive blood replacement in adults. The chemistry of the citrate calcium interaction; the dose, distribution, metabolism, and excretion of citrate; the toxic effects of citrate; and the treatment of citrate induced hypocaa1lcemia are discussed. Occasionally we have offered a personal view based on our own experience with massive transfusion during hepatic transplantation. Citrate toxicity is also discussed in most general reviews of massive transfusion. 10,1l
Citric acid (molecular weight 192 daltons) is a ubiquitous organic compound with three ionizable carboxyl groups. With three pKs (3.14,4.77, and 6.39) all <7.4, the majority of citrate present in the body has aUll three carboxyl groups ionized (Fig 1l)). These ionized carboxyl groups are responsible for the major pharmacologic action of citrate, the binding of divalent cations. This binding is accomplished by having two of the valences occupied by the divalent calcium ion. Because of the third ionized carboxy group, citrate is still highly soluble in aqueous media even when bound to a divalent cation. Citrate may bind to any of the metaUllic divalent cations and subsequently lower the concentration of the ionized form of that cation. While most reports dealing with the effects of citrate deal with its effects on ionized calcium, weUll documented depressions of magnesium have


Transfusion Medicine Reviews, Vol 2, No 2 (June), 1988: pp 76-94


C6H5 07
Fig 1. Chemical structure oft citrate.

also been reported. 12,13 Citrate binds slightly
stronger to Mg + + (formation constant 2.9 X
103) than it does to Ca + + (formation constant
1.88 x 103).12,14,15 Citrate is found in all human cells and is an intermediary in the Kreb's citric acid cycle. Because the citric acid cycle takes place within the mitochondria of a cell, tissues with a high number of mitochondria per cell (such as liver, skeletal muscle, and kidney) contain larger amounts of those enzymes responsible for the production and metabolism of citrate. Whbile neither a common nor routine laboratory test, plasma citrate levels can be measured by several methods. One common method involves incubation of plasma with bacterial citrate lyase in the presence of zinc ions, The reaction produces oxaloacetate whiceh is then acted upon by malate dehydrogenase resulting in the production of NADH from NAD. Production of NADH is measured spectrophotometrically after the serum proteins are precipitated. The normal adult plasma concentration of citrate is from 0.9 mg/dL to 2,5 mg/dL when measured with the citrate lyase method (Table 1l)). Slightly higher levels are found in children and in patients with hepatic or renal disease. 1166,17
Early in this century it was realized that calcium exists in biologice fluids in at least two forms, one diffusable across a dialysis membrane, the other

Table 1. Normali Adult Concentrations oft Citrate and Ilonized Calcium


Ilonized Calcium

mgg//ddLl mEqq//Ll tnmoll//Ll

0.9-2.5 0.14-0.39 0.047-0.130

4.5-5.4 2.3-2.7 1.1-1.4

nondiffusable. The nondiffusable form was found to be bound to serum proteins, especially albumin. By the 1930s most researchers in the field agreed that the diffusable portion of calcium existed in two states, bound to diffusable small lliiggands such as lactate or citrate, and in the free or ionized state. 18 These three states exist in equilibrium in the plasma and it is generally agreed that ionized
Ca + + is the physiologically active form. In the
healthy human, approximately 47% of calcium is in the ionized form, 40% bound to serum proteins (mosstt1ly albumin), and approximately 13% bound to smaller ligands. 19 These proportions differ with changes in pH, protein concentration, ligand concentration, and total calcium concentration.
While measurement of total serum calcium remains useful for gross or chronic disturbances of
serum calcium, acute changes in ionized Ca + +
are often missed with these measurements, Most laboratories use dye binding methods such as orthocresolphthalein or arsenazo III dye where the change in absorption by the dye is proportional to the total calcium concentration. 2o0 Atomice absorption specetrophotometry is also used and is an aceceurate and reproduceible method, but is rarely automated and requires ecareful maintenancee and standardization, While these methods have good acceuracey and precision, the difficeulty is that total calcium measurements may not aceceurately reflecet the econecentration of ionized ecalcium which is the physiologically aective form.
Early attempts to determine the ionized Ca + +
relied on nomograms for estimation of ionized
Ca+ + from measured total eca1lccium. In 1935,
MceLean and Hastings, in their extensive monograph on ecalcium in body fluids, developed a nomogram for estimating the econecentration of
ionized Ca + +. The nomogram was based on the
total ecalcium and total serum protein econecentration at a set pH and temperature. 18 With the advent of the more rapid ion selecetive elecetrode measure-
ment of ionized Ca + +, this nomogram and
derivations of it, have been shown to give poor
estimates of the ionized Ca + + levels. 21 ,22,23 The
inacecuracies of early nomograms likely result from the assumption that pH and small ligand
econecentrations had little effect on ionized Ca + +
econcentration. Another outdated method for
determining ionized Ca + + is the method of Soulier whiceh estimated the ionized Ca + + by its
effe'cet on the thromboplastin time of decealcified



p1laassma. This method is poor for ionized Ca + +

greater than 1.5 mmo1l/L (3 mEq/L) and can not

measure values <0.5 mmo1l/L (l mEq/L).19

The modern era of measurement of ionized
Ca + + began with the introduction of ion

se1leecctive electrodes: Initially, ion se1leecctive

e1leeccttrodes were limited by inference from other

b1lo0o0d and serum components. Present day ion

se1leecctive e1leeccttrodes operate by comparing the
binding of ionized Ca + + with one side of a non-

permeeaabb1le membrane which is specific for binding

ca1lccium ions and comparing it to a reference









20 20

If fewer Ca + + ions adhere to one side of the

membrane than the other, an electric potennttiiaa1l is

set up which can be measured via a second refer-

ence electrode, in contact with the serum sampp1le,

acting as a salt bridge. The membranes bind the

calcium by either an ion exchange mechanism,

where the binding moiety of the membrane forms

a ca1lccium salt, or by forming a neuttrraa1l but steri-

cally and electrostatically favorable binding pocket for the calcium ion. 2o0 Both these methods give

measurements which are rapid, highly reproduc-

ible, and can be adjusted to sample whole blood,

plasma, or serum. Measurement devices which

employ ultrafiltration or dialysis are not as useful

because ligand bound calcium, including citrate bound calcium is also measured. 20 One drawback

to the ion selective electrode is the lack of a

standard reference method, making results be-

tween different devices and different llaabborrattoorńies

difficult to compare. Work is presently being done

to come to an agreement for such a standard. Thus,

normal ranges depend on the individuuaalllaboratory.
Estimations of the normal range of ionized Ca + +
are given in Table 1.
Dosage and Distribution
The 1leevvel of citrate in the bloodstream durńing massive transfusion results from the net balance of citrate dose versus citrate removal. Due to the compp1lex interrelation of seveerraa1l factors affecting citrate delivery and removal, predicting the level of citrate and consequently its effect on the 1leevvel of
ionized Ca + + is extreemmee1ly diffiiccuu1lt. 24 The dose
of citrate is determined by several factors including the partiiccuu1lar blood component, the anticoagulant preservative formuu1laation, the rate of administration, the recipient b1lo0o0d volume, and the duration of administration. The citrate burden of various anticoagulant preservative formulations
is shown in Table 2. Citrate (MW = 189) is
present as both trisodium citrate dihydrate (MW = 291) and ciittrńic acid monohydrate (MW = 210) in
most formu1laattions. Given the moleccuu1lar weights, citrate represents 65% and 90% of ttrńissooddiiuummcitrate and citric acid respectively. The concentration of citrate can be estimated for whole blood, red b1lo0o0d cells, and fresh frozen plasma (PFFP) or platelets (Tabb1le 2). The highest concentration of citrate is naturally found in FFP. Although for many years acid-citrate-dextrose (ACD) solution presented the greatest citrate doseto the b1lo0o0d recipient, the newest additive formulations (AS-3) have the

Table 2. Citrate Content oft Various Anticoagulant-Perservative Formulations

Grams trisodium ceitrate (2H2O) Grams ceitriec acid (H2O) Grams citrate per unit Conecentration oft citrate per liter (mmoI/L) Conecentration of citrate (mg/dL)* in:
Whole Blood Packed RBC FFP Quantity of citrate (mg)* in: Whole Blood Paecked RBC FFP

1.485 .540
1.451 7.6
280 87
1451 200 976

1.656 .206 1.261 6.7
246 76
1261 176 843

1.656 .206 1.261 6.7
206 54
1261 176 843

2.244 .248
1.681 8.9
274 181 384
1681 596 843

* Calculated assuming a 450 mL donation; HCT, 41, and no movement of citrate into ceells. For ACOD, CPOD and CPODA-l assumes the production of paecked RBC with Hect, 80; FFP, 230 mL; and platelet ceonceentrate, 55 mL. For AS-l and AS-3 assumes produection of red blood ceells with final Hect, 56; FFP, 230 mL; and platelet conecentrate, 55 mL. ,

hbighest quantity of citrate. AS-3 red blood cells deliver three times the quantity of citrate as ACD or CPD packed cells.
For any given blood producet, the rate of administration and the size of the recipient are the key determinants of citrate administration. Thus, studies of citrate toxicity generally refer to mg citratelkg recipient/minute. For example, one unit of CPDAA--1l whole blood administered to a 70 kg man over five minutes corresponds to 1261 mg citrate/ 70 kg/5 min or 3.6 mg citratelkg/min. Citrate is rapidly removed by the liver and kidney. However, the distribution of citrate also plays an important role in determining the citrate level following transfusion. Although the relative importance of metabolism v redistribution of citrate is not fu1lly investigated, citrate can be considered as a first order approximation to be distributed throughout the extracellular fluid space. This distribution occurs within five minutes of infusions of mild-to-moderate quantities of citrate. For example, in one study 500 mL of citrated blood was infused over five minutes to adults (4 mg citrate/ kg/min x 5). Citrate levels were measured every minute. At the end of the infusion, the measured citrate level was 66% that which would have been predicted had citrate remained in the intravascular space. Within three minutes after stopping the infusion, the blood level was equal to that which would have been predicted had citrate distributed itself over the extracellular volume. 25 After infusion of 2.35 mg citrate/kg/min over ten minutes, the concentration of citrate at the end of the infusion was equal to that which would have been expected from redistribution over the extracellular volume. 9 However, rapid challenges of large quantities of citrate can exceed redistribution, metabolism, and excretion. 26 Five patients receiving a mean of 5.5 mg citrate/kg/min over 15.6 minutes developed average peak citrate levels of 62 mg/dL which was 1.5 times greater than the expected citrate level assuming complete extracellular redistribution. 27 As a result of metabolism and redistribution outside the vascular space, there is an initial rapid exponential decline in the concentration of citrate after cessation of rapid blood infusion. Further removal of citrate results from continued metabolism and renal excretion. Due to the permeability of cells to citrate and the apparent large volume of distribution, complete metabolism and excretion following prolonged

rapid transfusion would be expected to take several hours. Evidence for such delayed clearance has been found in studies of metabolism and renal excretion of citrate following massive transfusion. 2288.•2299
The duration of citrate administration is also a key determinant ofthe blood citrate level. Plateletpheresis studies with relatively constant low dose citrate infusion over approximately 100 minutes have shown that the citrate level attained is only 25% of that which would be predicted from redistribution throughout the extracellular space. Without metab0o1lism and excretion of citrate during the period of infusion, toxic levels of citrate would have developed. 30,31 In two independent studies very similar blood citrate levels developed in individuals receiving 4 mg citrate/kg/min over five minutes as were found in individuals receiving only 1.6 mg citrate/kg/min over 113 minutes. 30 Thus, prolonged rapid blood transfusion would be expected to result in higher citrate levels than equally rapid infusions of short duration. This is of clinical importance in settings such as hepatic transplantation where rapid transfusion may continue for hours. With prolonged duration of blood administration, citrate metab0o1lism and excretion (rather than distribution) become the most important defense against the development of toxic concentrations of citrate.
Another means for the body to deal with a low ionized calcium due to a citrate load is to mobilize stores of calcium. Parathyroid hormone (PTH) levels are measurably elevated within 2 to 4 minutes after administration of citrate32 and reach peak levels between 5 and 15 minutes after infusion. 13 Elevations in PTH have been found both experimentally and during surgical procedures requiring blood support in adults and during exchange transfusions in infants. 13 During plasmapheresis, one study found that the ionized calcium remained low with continuous citrate infusion but the total calcium dropped at first, then returmned to near normaalllleevvels. Ultrafilterablie cal-
cium, which equals the total ionized Ca + + plus ligand bound Ca + + (but not the protein bound
calcium) continued to rise during the apheresis procedure. 1l3 This is probably a result of calcium mobilization with most of the newly mobilized calcium being bound to the infused citrate. Protein bound citrate decreased by 35% during the procedures, representing a significant buffer of



calcium when ionized Ca + + levels drop. These changes may be affected by temperature, pH and other alterations in addition to the variability of various PTH assays among different methods and laboratories. 33
Citrate Metabolism
The metabolism of citrate involves multiple biochemical pathways (Fig 2). Citrate can directly enter the Kreb's tricarboxylic acid cycle to be completely metabolized to CO2 and H20, can participate in fatty acid and amino acid synthesis, and can be converted to glucose via gluconeogenesis. Each of these pathways will be reviewed in turmn. Overall, the complete oxidation of citrate results in the production of CO2 and H20. Complete metabolism of ionic citrate - 3 to nonionic endproducts consumes three H + ions according to:
C6Hj 0 7 - 3 + 4.5 O2 + 3 H+ ~ 4 H20
+ 6 CO2
Exogenous citrate can be actively transported through the mitochondrial membrane to participate in Kreb's cycle reactions. Tricarboxylic acids (citrate, isocitrate, aconitate) are actively transported across the mitochondrial membrane by carrier molecules. Movement of citrate into the mitochondria is linked to movement of malate out of the

mitochondria. Once inside the mitochondria, two carbons of citrate are removed and two molecules offCO02 formed. With additional turmns oftthe Kreb's cycle, the original carbons of citrate are converted to CO2 ,• Since turning of the Kreb's cycle regenerates cycle intermediates from the carbons introduced by acetylCoA, metabolism of exogenous citrate exclusively by Kreb's cycling would not reduce the concentration of citrate nor have a net effect on acid/base balance. Instead, the Kreb' s cycle intermediates are able to be metabolized by additional pathways. Within the mitochondria, oxaloacetate can be converted to pyruvate by the action of pyruvate carboxylase, which under normal conditions serves to convert pyruvate to oxaloacetate in an anapleurotic reaction. Conversion to pyruvate requires biotin as a cofactor, consumes a second H + ion, and yields CO2 ,• The pyruvate formed can be then converted to acetylCoA in a series of reactions that consume the third H + ion and yield an additional CO2,•
However, other Kreb's cycle intermediates are also able to be transported through the mitochondrial membrane. Alpha-ketoglutarate and malate are two such intermediates. As the concentration of cycle intermediates increases inside mitochondria following citrate administration, alpha-ketoglutarate and malate would be expected to leave


Exogenous Ctlrofe-3

SyY~f~S'S \

a.Ky 2 ASspPoorrtt]]ote.,


teł Acetyl

C oA


0,0100celol..-2 ~-------

Exogenous Ct~rote-3

Fig 2. Metabolism oft citrate.

the mitochondria. Evidence suggests that metabolism of exogenous citrate results in net transport of malate outside the mmiittochonnddńraia.. Because the metabolism of citrate to malate involves incomplete turmning of the Kreb's cycle, the normal balance of
H + ion production and consumption is not maintained and a net single H + ion is consumed.
aOnncce in the cytoplasm, malate is able to be converted to oxaloacetate which can participate in two biochemical pathways described below.
Oxaloacetate is a ketoacid and as such can undergo a transaminase reaction with glutamate to form aspartate and alpha-ketoglutarate. The reaction is reversible with no net loss of oxaloacetate carbons and no net production or consumption of
H + ion. A second pathway of cytoplasmic oxalo-
acetate metabolism involves conversion to phosphoenolpyruvate (PEP) by PEP carboxykinase. This enzyme is a key enzyme in gluconeogenesis and is found in celIls such as those of the liver, kidney, and skeletal muscle. The reaction requires energy in the form of GTP and results in the release of CO2., Thus, exogenously administered citrate can be converted to PEP through oxaloacetate. Oanncce formed, PEP can be converted to glucose via the enzymes of gluconeogenesis. However, during conditions of ATP depletion the metabolism of PEP to pyruvate is favored. The formation of pyruvate from PEP consumes an
additional H + ion. Pyruvate is freely permeable to
the mmiittochoonnddńraia.. aOnncce inside the mitochondria, it can be converted to acetylCoA in a series of
reactions that release CO2 and consume one H +
ion. The acetylCoA is then free to combine with mitochondrial oxaloacetate and be metabolized to CO2 and H20. Thus, the complete metabolic breakdown of the six carbons of citrate mayyaaIlso occur via pathways that include exttrraammiittoocchhoondńriaall
enzymes, produce CO2, and consume three H +
ions. Citrate conversion to glucose via gluconeogenesis yields fewer CO2 molecules but also
consumes three H + ions.
An additional biochemical pathway of citrate metabolism is that provided by citrate cleaving enzyme (Fig 2). This enzyme, ATP-citrate lyase is found in liver and adipose celIls and splits cytoplasmic citrate in the presence of CoASH to form ace-
tylCoA and oxaloacetate. One H + ion is
consumed in the reaction. The acetylCoA formed is not permeable to the mmiittochonnddńraia,, but is able to participate in the biosynthesis of fatty acids. Thus,

increased cytoplasmic citrate as a result of massive transfusion might be expected to tteemmppoorraańrillyy stimulate fatty acid synthesis. The cytoplasmic oxaloacetate formed as a result of citrate cleavage may then be further metabolized according to the reactions outlined above. The metabolism of cytoplasmic citrate via citrate cleaving enzyme to oxaloacetate and then to PEP, pyruvate, and mito-
chondrial acetylCoA also consumes three H +
ions. The rate limiting step of citrate metabolism fol-
lowing massive transfusion is unknown. Evidence from liver transplantation and from studies of renal excretion of citrate suggest that alkalemia slows the metabolism of citrate presumably by retarding movement of citrate and malate across the mitochondrial membrane. Whether factors which might influence the activity of citrate cleaving enzyme exert an overall effect on the metabolism of administered citrate is less welIl studied.
Citrate Clearance
Clearance of citrate is highest in those organs which receive a high proportion of the cardiac output and which are composed of celIls with numerous mitochondria.' CelIls which are dependent on glycolysis for their energy needs, such as red blood celIls, have low levels of citrate and do not remove citrate from the circulation. Liver, kidney, and skeletal muscle are responsible for most of the metabolism and excretion of citrate.
Since the 1930s, investigators have recognized that basal serum citrate levels are slightly higher in patients with chronic liver disease than in those without liver disease. 16 ,34 In normaIl fasting humans, about 20% of the endogenous citrate in the serum is removed with each pass through the liver. 1166,34 In one study when a perfusate containing 54 mg/dL citrate was presented to isolated calf livers, there was a rapid initial clearance of 50% of the citrate over the first thirty minutes folIlowed by a more gradual clearance of the remaining 50% over the next two and a half hours. 35 Convincing evidence of the importance of the liver in citrate clearance in humans is seen during liver transplantation when there is a marked rise in serum citrate with concomitant drop in ionized
Ca + + during the anhepatic phase of the surgery
with a rapid reversal of these changes when the new liver is perfused.22.36.37 Citrate clearance by the liver appears to be greatly reduced during hy-

pothennia, increasing as the patient is warmed. Hypotension with decreased hepatic perfusion also leads to a decrease in clearance of citrate by the liver. 28
The kidney is also essential for citrate clearance. Some investigators feel that the kidney is the most important organ for citrate clearance since it can metabolize large amounts of citrate and also can excrete nonmetabolized citrate in the urine. 38,39,4O Estimates of the amount of exogenous citrate handled by the kidney range from 20%28,36 to 68%40 of a given load in humans. Most of the uptake of citrate by the kidney is from reabsorption of filtered citrate in the proximal tubule but up to 30% of the total renal uptake is peritubular uptake from postglomerular blood. 41 The proximal tubular cells, which are rich in mitochondria, are responsible for the metabolism of citrate.
Normally < 1I% of the citrate filtered by the kidney
is excreted in the urine, but in the presence of high plasma concentration of citrate or alkalemia, there is a dramatic increase in the fractional excretion in

the urine,42 Up to 60% of filtered citrate may appear in the urine of humans during metabolic alkalosis,41 This increase in excretion results from inhibition of citrate transport into mitochondria by the tricearboxylic acid carrier. Other conditions leading to increased urinary excretion of citrate include high levels of other organic acids of the Krebs' cycle such as fumarate or malate and specifice inhibitors of the Krebs' cycle pathways such as fluorocitrate and malonate, Calcitonin, vitamin D, calcium, and magnesium have also been shown to increase citrate excretion. 4411 Conversely, the excretion of citrate is reduced during acidosis. 42
The effects of K + and bicarbonate levels on ci-
trate excretion appear to be due to their effects on acid base balance rather than a direct effect of the ion. 42 In the cliniceal setting, clearance of citrate by the kidney may also be limited by the decreased glomerular filtration rate and renal ischemia that frequently develops during massive transfusion. More work is needed to assess the relative contributions of the liver and kidney in citrate me-



Fig 3. Effect of rapid citrate infusion on the EKG. The top panel shows a prolonged GQoTc interval. Following a period of rapid transfusion, the ilonized calcium felli to 0.6 mEq/L and a widened GQRS complex developed I(middle panelli. The effect was reversed with intravenous administration of calcium (bottom panel).

tabolism and excretion during massive citrate loads.
Effects on the EKG
Depression of the ionized Ca + + has
characteristic effects on the EKG. The most commonly recognized effect is prolongation of the QT interval (Fig 3). The QT interval corresponds to the time from contraction to repolarization and varies with the heart rate. The corrected QT interval (QTc), which takes into account the effect of heart rate, is the time from the origin of the QRS complex to the end of the T wave divided by the square root of the RR interval. The corrected QT interval mayaIso be defined by substituting the time from the origin of the QRS complex to the origin of the T wave rather than the end of the T wave, and is then referred to as the QOTC. 43 The QTc in normal adults ranges from 350 msec to 440 msec44 and the QoTc from 180 msec to 240 msec. 45 The measured QT interval for a particular individual is influenced by numerous factors including age, sex, autonomic innervation, myocardial ischemia, antiarrhythmics, hypothyroidism, and severe hypothermia in addition to hypocalcemia. 46
The relationship between hypocalcemia and a prolonged QT interval has been recognized for many years. In early studies, sodium citrate was infused into normaI conscious volunteers or patients under anesthesia and the prolongation of the QT interval observed. 26,27,47 Similar studies in dogs documented similar effects on the EKG and demonstrated that the effects were abolished with administration of calcium.48,49,50 Despite the reproducible effect of citrate on the QT interval, studies in patients undergoing transfusion showed that the QT interval was an unreliable guide to administration of supplemental calcium.24 The re-
lationship between the level of ionized Ca + + and
the length of the QT interval during rapid transfusion was found to be nonlinear and has been described by both 10garithmic51 and hyperbolic45
curves. At mild depressions of the ionized Ca + +
(>1.75 mEq/L), little or no effect on the QoTc is seen. At more moderate depressions of ionized
Ca + + (1.0 mEq/L to 1.75 mEq/L), the average
QoTc interval for a group of individuals will become prolonged. 45 ,51 This prolongation

increases more sharply at severely depressed levels
of ionized Ca + + (0.25 mEq/L to 1.0 mEq/
L).45,51 A study of twenty adults undergoing intraoperative rapid transfusion correlated corrected
QT intervals and ionized Ca + + and found the
correlation coefficient was only r = .61.45 The
correlation was slightly higher using QoTc compared with QTc. The 95 percent confidence intervals indicated that a given QoTc was often associated with a twofold or greater range in
predicted ionized Ca + + levels.45 Because the range of ionized Ca + + associated with any given
QTc interval is broad, measurement of the QTc interval cannot be substituted for accurate mea-
surement of the ionized Ca + + in patients likely
to develop citrate toxicity during massive transfusion.
The effect of citrate infusion on the QT interval has also been studied in the setting of apheresis procedures uncomplicated by many of the confounding variabIes found during intraoperative massive transfusion. One study of 15 apheresis procedures measured a 32% average decrease in
the level of ionized Ca + + with a corresponding
prolongation of the QT interval by an average of 80 msec (range 50 msec to 120 msec).30 These effects occurred during an average citrate infusion of 1.58 mg/kg/min for 113 minutes resulting in a mean postprocedure citrate level of 26.7 mg/dL. Although the QT interval correlated with the rate and dose of citrate administered, the level of
ionized Ca + + could not be predicted from the
QT interval. Following cessation of the procedure, the QT interval returned to normal in 2 to 15 minutes which was significant1y less than the time
required for levels of ionized Ca + + and citrate to
return to normal. In another study of 12 plateletpheresis procedures, prolongations in the QT, QTc, QoT, and QoTc were correlated with the
level of ionized Ca + + during procedures that
infused citrate at an average rate of 1.38 mg/ kg/min. 52 Although the best correlation was found
with the QoTc (r = .59), the QoTc could not predict the level of ionized Ca + +. Among
individuals with depressed levels of ionized Ca + +, only 9 of 30 QoTc measurements were longer than the longest baseline reading. In addition, the authors found that the level of ionized
Ca + + could not be predicted by the degree of
prolongation of the QT interval over baseline for a
given individual. In a third study of 79 platelet-

pheresis procedures levels of ionized Ca + + ranging from 2.25 to 1.35 mEq/L were found to be in 1liinnear relation to levels of serum citrate ranging from 2 to 40 mg/dL. 31 However, the average QTc prolonged by only 13 msec following the procedure.
When ionized Ca + + levels become severree1ly depressed, other effects on the EKG are observed. Earr1ly reports of ventricular fibrillation and cardiac arrest during massive transfusion lack complete documentation of electrolyte and acid/base disturbances. 53,54,55 Progressive severe hypocalcemrnia would be expected to result in prolongation of the QRS complex. Fig 3 shows an intraoperative EKG from an adult patient in our hepatic transplantation program. Before the onset of rapid transfusion, the tracing showed annoorrmmaaIl QRS morphoo1lo0gy but a pro1lo0nged QoTc interval. In the middle panel is shown the EKG during a period of rapid blood infusion with a corresponding pH =
7.26, K+ = 5.2, Temperature = 92°F and ionized Ca + + = 0.6 mEq/L. Following two
grams of intravenous (IV) calcium chloride and no change in pH, K +, or temperature, the EKG returmned to normal (bottom panel). Such dramatic effects on the EKG are rare outside the setting of liver transplantation or rapid blood exchange in neonates.
The mechanism of prolongation of the QT interval during citrate toxicity and hypocalcermnia is likely related to the plateau phase of the myocardial action potential. 56 The action potential of ventricular depolarization/repolarization is divided into discrete phases which are regu1lated by selective movement of ions across the myocardial cell membrane. The initial rapid depolarization results from a closure of K + channels and an opening of fast moving Na + channels. This corresponds temporally to the QRS tracing on the EKG. Depolarization is then sustained by a balanced slow inward movement of Ca + + and Na + and a slow outward movement of K +. Gradually the potential across the membrane becomes more negative as K + ions leave the cell. aOnncce a threshold is reached, fast K + channels open and a sudden movement of K + outside the cell repolarizes the membrane. Myocardial repolarization is less well coordinated for the aggregate of myocardial cells than depolarization and, therefore,results in a broad wave of repolarization (T wave). The duration of the QT interval is, therefore, dependent on

the duration of the action potential plateau. A reduction in extracellular ionized Ca + + results in a decrease in the slow outward K + current and prolongation of the plateau. Because the action potential plateau without the T wave is measured by the QoTc, the QoTc correlates better with the degree of hypocalcemia than the QTc. Whether citrate anions have any independent effect on myocardial permeability to Ca + + or K + and exert any effect on the action potential plateau independent of direct lowering of the extracellular concentration of ionized Ca + + is not well understood.
Effects on Ventricular Performance
In 1883, Sydney Ringer demonstrated that the isolated frog heart was capable of sustained contractions when suspended in saline. 5577 One year later, Dr Ringer published a second paper in the same jourmnal noting that the water used to prepare the saline of his previous study was not distilled water, but rather pipe water obtained from the New Water Company, London. 58 Chemical analysis showed it to be contaminated with appreciable quantities of calcium. Ringer repeated his experiments and discovered that sustained beating of the heart was dependent on the presence of extracellular calcium. In the century that has followed, an enormous body of information has developed which documents the importance of calcium in myocardial performance. An excellent review has been recently published.5599
For over thirty years it has been recognized that massive transfusion with citrate toxicity and hypocalcemia would be expected to decrease cardiac performance and early studies in both humans47.6o0.61 and animals49,62.63.64 supported the concept. Nevertheless, acceptance of citrate toxicity was not without some degree of controversy. 26 The development of cardiac surgery and radical cancer surgery in the 1950sprompted several investigations into the cardiovascular effects of citrate. In an early study ,Bunker examined the effect of transfusion on blood pressure and on serum levels of citrate and total calcium.9 Twenty-four patients receiving a median of 3,500 mL of blood developed serum citrate levels of 10 mg/dL. The median rate of citrate infusion was 0.92 mg/ kg/min (range 0.3 to 6). In 10 patients, systolic blood pressure fell to <100 mmHg. In a subsequent more detailed study in 1962, Bunker examined the cardiovascular effects of direct citrate

infusions into six lighhtt1ly anaesthetized adults undergoing stripping of leg varicosities. 27 The patients were challenged with citrate at rates ranging from 3.7 to 7.4 mg/kg/min for 9 to 19 minutes. Recipients developed citrate levels of 30 to 77 mg/dL. In 5 of 6 individuals, there was a decline in stroke volume and left ventricular work, with a corresponding rise in pulse. Mean arterial pressure in these patients decreased 26%. The cardiovascular abnormalities corrected promptly with stopping the citrate or with infusion of calcium chloride. The one patient who developed no measurable cardiovascular changes was the patient who received the lowest citrate challenge (3.7 mg/ kg/min) and who developed the lowest blood citrate level (30 mg/dL). In the same report, a similar effect was observed in dogs at citrate levels of 50 to 75 mg/dL. Serious cardiovascular depression or death occurred in the animals at citrate levels of $0 to 190 mg/dL and rates of citrate infusion of 10 to 15 mglkg/minute. Subsequent studies in dogs by other investigators showed similar overall results. 48 ,50,65
During the 1970s knowledge of the effects of citrate on left ventricular performance was refined. Rather than infusing citrate, these studies involved infusions of citrated blood versus recalcified ci.trated blood to which heparin was added. In one study of six patients undergoing open heart surgery,66 infusion of warm citrated blood at 150 mLi/min for three minutes resulted in a 28% decrease in cardiac output with no consistent change in mean arterial blood pressure, but with a rise in left atrial pressure. In contrast, the infusion of heparinized, recalcified, citrated blood at the same rate resulted in an 18% mean increase in cardiac output. Although administration of 200 to 300 mgm of (IV) calcium chloride had little effect on cardiac output prior to the infusion of the citrated blood, the same dose of calcium administered after the transfusion caused a mean increase of 24% in cardiac output. In the same study, the authors bIled dogs to a mean systolic arterial pressure of 50 mmHg which was maintained for 15 minutes. Infusion of citrated blood at rates ranging from 2 to 3 mLl/kg/min to 6 to 9 mLlkg/min was compared with the infusion of recalcified, heparinized blood at 6 to 9 mLl/kg/min. Although the infusion of recalcified blood consistently resulted in an increase in aortic pressure and cardiac output with little rise in left

atrial pressure, animals receiving an equal volume of citrated blood incompletely restored aortic blood pressure and cardiac output and developed considerable elevation in left atrial pressures. The more rapid the rate of citrated blood infusion, the greater the depression in left ventricular performance. Four of nine dogs receiving citrated blood at maximaIl rates of 6 to 9 mLl/kg/min suffered cardiac arresLt. Thus, the overall effect of transfusion with citrated blood was to blunt the normaIl left ventricular response to volume loading.
The blunting of the left ventricular response to volume loading resulting from the rapid administration of citrated blood was conclusively demonstrated in man in a well designed 1976 study.67 Nine patients undergoing coronary revascularization were studied. Each had normal ventricular function with ejection fractions >70 percent. Before cardiopulmonary bypass, the patients were transfused with two units of citratephosphate-dextrose (CPD) blood. Each unit was 37 C, pH 7.4 and <48 hours old. One unit was recalcified and heparinized and the other was heparinized but not recalcified. Each patient served as his own control. The order of transfusions was randomized and blinded to those recording cardiovascular measurements. Each unit of blood was transfused over three minutes (2 mL/kg/min for 3 min). After the first transfusion was administered, the equivalent volume of blood was removed, the patient allowed to return to a stabIle state, and the second unit infused. Cardiovascular measurements were taken every 45 seconds during the transfusions. The results from this study are shown in Figs 4 and 5. Although patients receiving one unit of citrated blood in three minutes increased left ventricular performance in response to volume, the magnitude of increase was blunted compared to those individuals receiving recalcified heparinized blood. This blunted response to volume loading was accompanied by a 27% decrease
in the level of ionized Ca + + at the end of the
three minute transfusion. Several factors appear capable of rendering the
myocardium more sensitive to the depressant effects of rapid infusions of citrated blood. Acidemia, hyperkalemia, and hypothermia have all been recognized to increase susceptibility to the cardiodepressant effects of citrated transfusions. Following earIly studies of cardiac autotransplanta-