Absence of particle-induced coronary vasoconstriction during cardioplegic infusion: Is it desirable to use a microfilter in the infusion line?

J THORAC

CARDIOVASC SURG

1991;101:473-80

Absence of particle-induced coronary vasoconstriction during cardioplegic infusion: Is it desirable to use a microfilter in the infusion line? Many cardiac surgical units now use a microfilter in the infusion line for de6very of crystalloid cardioplegic solution to protect against the potential hazards of particulatecontamination of cardioplegic solution. The aim of this group of studieswas to determine the effects of particulate contamination of cardioplegic solutiol6, in order to establish whether a microfilter is needed in the infusion line. Total particle counts performed on two commercial cardioplegic solutions were low, but there weresufficient particles greater than 10 #Lm in diameter to cause coronary vasoconstriction. In isolated rat hearts a 2o-minute infusion of St, Thomas' Hospital cardioplegic solution produced a progressive reduction in coronary flow, which was not prevented by the inclusion of a 0.8 #Lm filter in the infusion line. Two studies were performed on canine hearts to determine the effects of unfiltered cardioplegic solution on coronary vascular resistance. In the first, cardioplegic solution at 20° C was ~ for 20 minutes at a constant pressure of 50 mm Hg and flow rate was measured. In the second, the same solution at 4° C was ~ at a constant flow rate for 21/ 2 minutes and the inf'ttiion pressure was measured. In neither study did coronary vascular resistance rise. A final cHnical study involving patients undergoing coronary bypass surgery compared the effects on coronary resistance of inf'ttiion at a constant flow of filtered verses unfiltered cardioplegic solution (n = 10 in each group). There was a similar rise in coronary perl'usion pressure in both groups during the inf'ttiion. We conclude that there is insufficient evidence of particle-induced coronary vasoconstriction to justify the expense of a microfilter in the cardioplegic inf'ttiion line. Christopher Munsch, FRCS,a Franklin Rosenfeldt, MD,a FRACS,a Victor Chang, MB, BS,a Mark Newman, FRACS,a and Bruce Davis, FRACS,b Melbourne. Australia

h e possibility that microscopic particles in cardioplegic solutions coulddamage the heart was first raised by Robinson and colleagues' in 1984. They observed that large infusions of unfiltered commercial cardioplegic solutions into isolated rat hearts led to coronaryvasoconstriction and impaired recovery of contractility. This effect was avoided by using a 0.8 #Lm microfilter i~ the infusion line. On the basisof this workthey concluded "a From The Baker Medical Research Institute,' and The Alfred and St. Francis Xavier Cabrini Hospitals," Melbourne, Australia. Supported by grants from the National Health and Medical Research Council, the National Heart Foundation of Australia, and the National Heart Research Fund of Great Britain. Received for publication May I, 1989. Accepted for publication March I, 1990. Address for reprints: F. L. Rosenfeldt, MD, Baker Medical Research Institute, PO Box 348, Prahran, Victoria 3181, Australia.

12/1/21156

strongcasecan be made for the routine inclusion of a 0.8

#Lm filterin the finalcardioplegic infusion linefor cardiac

surgicalprocedures." Subsequentexperiments bya group including the original investigators indicated that this impairmentof coronaryflow was attributable to a small number of contaminant particles of 10 #Lm or more in diameter.? The authors alsorecognized the limitations of the rat heart modelused in both studiesand emphasized the needto duplicatetheir observations withother species and in clinical studies. Regardless of these reservations and despitethe absenceof further experimental data, it becamea standard clinical practice in many cardiac surgical units to use a 0.2 #Lm filter in the cardioplegic infusion line. In this series of studieswe assessed the current extent of particulatecontamination of commercial cardioplegic solutions. We also hoped to establish whether such contaminationhad any effectoncoronaryflow notonlyin the 473

The Journal of Thoracic and Cardiovascular

474 Munsch et al.

Surgery

- - Filtered

140

"2

120

0

100

C

o ?fi 3 0

u::: >-

80 60

~

40

(;

20

c 0

0

----- Unfiltered

CARDIOPLEGIC INFUSION

REPERFUSION

I

0 2

4

6

8

10

12

14

16

18

20

22

24

26

Time (minutes)

Fig. 1. Coronary flow rates during infusion of filtered and unfiltered cardioplegic solutions into rat hearts, followed at 20 minutes by reperfusion with buffer solution. Flow rates are plotted as a percentage of the initial (control) flow rate. Each point represents the mean of the values for six hearts and the bars indicate the SEM.

25

20

en .>< <,

l:

"e

15

<,

]

s~

10

~

g ,.. 16l: 0

15

o

1

N

5

o

5

10

15

20

25

Time (minutes)

Fig. 2. Coronary flow rate during infusion of unfiltered cardioplegic solution at a constant pressure into dog hearts. On completion of the infusion, coronary vasoconstriction was produced by a bolus injection of I mg norepinephrine (N). Each point represents the mean of the values for five hearts and the bars indicate the SEM.

rat heart but alsoin canine and human hearts. It wasour ultimate intentionto determinewhetherthe clinical useof a microfilter in the cardioplegia infusion line is justified. Methods Cardioplegic solutions. In these studies we used a St Thomas' Hospital No. I cardioplegic solution prepared by injecting the contents of a 20 ml ampule of commercially prepared car-

dioplegic additive and 10 mmol sodium bicarbonate (both David Bull Laboratories, Mulgrave, Australia) into a I L bag of commercial Ringer's solution (Travenol, Toongabbie, Australia). The final composition of this solution was KCI, 20 mmol/ L; MgCh, 16 mmol/L; procaine hydrochloride, I mmol/L; and NaHC03, 10 mmol/L, The solution had a pH of7.7. In the rat study and the constant pressure infusion dog study, a procainefree additive was prepared in the Alfred Hospital Pharmacy, which when mixed with commercial Ringer's solution resulted in a procaine-free St. Thomas' solution with 16 mmol/L ofKCI. In addition to St. Thomas' Hospital solution, particle counts were also performed on a commercially prepared potassium cardioplegic solution used at the Royal Children's Hospital in Melbourne. This second solution consisted of750 ml cardioplegic solution base (Delta West, Bentley, Australia), to which was added 40 mI12.5% mannitol, 10 mI8.4% sodium bicarbonate, and 200 ml 0.9% sodium chloride. The final solution comprised 30 mmol/L KCI and 10 mmol/L NaHC0 3 . The solution had a pH of 7.65. Particle counting. We performed the first particle-counting study in a commercial quality-assurance laboratory (David Bull Laboratories) with an HIAC PC320 particle counter (HIAC/ Royco Instruments Division, Menlo Park, Calif.). The counter was calibrated by half-count calibration method with latex monosized spheres of 5.25 ~m and 25.7 ~m diameter.' Counts were given as the number of particles more than 5 /-Lm in size and the number of particles more than 20 /-Lm in size. For each sample six aliquots of 25 ml were counted, with the first count discarded according to standard practice. To determine the source of any particles in the final cardioplegic infusate, the following procedure was used. The first 250 ml sample was drawn directly from the bag of Ringer's solution. The cardioplegic additive and sodium bicarbonate were then added to the bag, and a second 250 ml sample was withdrawn. A third sample was then withdrawn through a cardioplegia delivery set with a coarse (I50 ~m) filter (Kal Life, Bentleigh, Australia). Finally, a fourth sample was withdrawn through the same tubing set with the addition of a 0.2 ~m commercial cardioplegia filter (Cobe Laboratories, Coburg, Australia). A second particle counting study was performed on the St.

Volume 101 Number 3

Particles in cardioplegic solution

March 1990

475

Table I. Source ofparticles in cardioplegic solution Sample Particle size (pm) ~5

~20

I (R) (particlesjml)

ll(R+A) (particlesjml)

lll(R+A+T) (particlesjml)

IV(R+A+T+F) (particlesjml)

35 ± 8.4* 0.11 ± 0.03

35 ± 8.2 0.11 ± 0.02

43 ± 9.6 0.35 ± 0.06

0.11 ± 0.04

R. Ringer's solution; R + A; Ringer's solution additive + tubing set + filter. 'All values mean ± SEM.

3 ± 0.9

+ additive; R + A + T. Ringer's solution + additive + tubing set; R + A + T + F, Ringer's solution +

Thomas' Hospital solution by the particle research laboratory at the South Australia Institute ofTechnology, also with an HIAC PC320 counter. This independent set of particle counts allowed us to validate our earlier counts and to adjust our counting windows to 2, 5, 10, and 15 .urn to more closely approximate the standards established for the United States" and United Kingdom.! Rat study. We repeated, as closely as possible, the experiments of Robinson and colleagues! on the isolated rat heart. Male Sprague-Dawley rats (body weight, 250 to 350 gm) were anesthetized with 4% halothane/oxygen and heparinized. Their hearts were excised. A modified Langendorff perfusion method" was used in which the hearts were initially perfused with oxygenated Krebs-Henseleit bicarbonate buffer at 37° C under a constant perfusion pressure of 60 cm H 20 . During a 5-minute equilibration period coronary flow was measured. The perfusion line was then clamped and procaine-free St. Thomas' cardioplegic solution at 20° C was infused into the aorta, again at a constant pressure of 60 cm H 20 . In one group of hearts an in-line 0.8 .um filter was used; in the other group the cardioplegic solution was unfiltered (n = 6 in each group). Infusion was continued for 20 minutes. The coronary effluent was collected and measured once per minute. At the end of this period, the hearts were reperfused with Krebs-Henseleit buffer at 37° C for 5 more minutes, and coronary flow was measured. Dog studies. We performed two studies to determine the effects of infusing unfiltered cardioplegic solution on the coronary vascular resistance of the canine heart. In the first the solution was infused at a constant pressure and flow was measured; in the second the infusion was given at constant flow and aortic root pressure was measured. Constant-pressure study. This study was modeled on Robinson and coworkers' protocol' but was adapted for the dog. Five mongrel dogs (weight, 17 to 25 kg) were anesthetized with pentobarbital sodium (30 mg/kg), intubated, and placed on mechanical ventilation. A median sternotomy was performed and the azygous vein was ligated. Two cannulas were inserted into the ascending aorta, one for to infuse cardioplegic solution and the other to monitor the aortic pressure. A cannula was inserted in the right atrium to collect the coronary effluent. The venae cavae, the pulmonary artery, and the aorta were occluded, and cardioplegic infusion was then commenced. Pressure in the aortic root was held constant at 50 mm Hg by a feedback control of the roller infusion pump. As in Robinson and coworkers' study, the cardioplegic solution did not contain procaine and was delivered at 20° C for 20 minutes. The flow rate of coronary effluent from the right atrium was measured. On completion of infusion, norepinephrine (l mg in 10 m\) was injected into the infusion line to test the reactivity of the coronary circulation. Constant-flow study. This study was intended to reproduce

more closely the usual clinical technique of cardioplegic administration. Eight greyhound dogs (weight, 24 to 35 kg) were anesthetized with methohexital (l mg/kg) followed by chloralose (80 rng/kg). The dogs were intubated and placed on mechanical ventilation. A median sternotomy was performed. After heparinization, the animal was cannulated for cardiopulmonary bypass. Two additional cannulas were inserted into the ascending aorta, one to infuse the cardioplegic solution and the other to monitor the aortic root pressure. Cardiopulmonary bypass was commenced, the aorta was crossclamped, and unfiltered procaine-containing St. Thomas' cardioplegic solution at 4° C was infused at a rate of 5.5 nil/kg/min for 1 V2 minutes. Thirty minutes later, a second infusion of cardioplegic solution was delivered over a period of I minute. During these infusions at constant flow, the aortic root pressure was monitored. Clinical human study. This study was carried out on patients undergoing elective, isolated coronary artery surgery. All procedures were performed with standard techniques by the same surgeon (B.D.), anesthesiologist, and perfusionist. Bypass was instituted at a flow rate of 2.4 L/min/m 2 by means of a CML membrane oxygenator (Cobe Laboratories), with systemic cooling to 28 ° C. After aortic crossclamping, 700 ml of the commercial procaine-containing St. Thomas' cardioplegic solution at 4°C was infused by roller pump into the aortic root through a 14 gauge cannula. Flow rate was held constant at 500 ml/rnin. One group of patients (n= 10) received unfiltered cardioplegic solution. A second group (n = 10) received solution filtered through a commercial 0.2 .um cardioplegia filter (Pall Biomedical, East Hills, N.Y.). Infusion pressure was monitored through a separate needle in the aortic root. Animal care. All animals used in these experiments received humane care according to the guidelines set down by the National Health and Medical Research Council of Australia. Data analysis. All data are presented as the mean ± the standard error of the mean (SEM). To detect differences between filtered and unfiltered groups and any differences occurring over the time course of the experiments, statistical analysis was performed by repeated measures analysis of variance with a general linear models procedure. Significance was set at a p value of less than 0.05.

Results Particle counting. The first set of particle counts performed on St. Thomas' Hospital cardioplegic solution (Table I) showed that the number of particles larger than 5 ",m and the number of particles larger than 20 ",m were relatively small and well within the Australian, British, and U.S. standards (Table II). Most of the particles were already present in the Ringer's solution, with minimal

476

The Joumal of Thoracic and Cardiovascular Surgery

Munsch et al.

Table II. U.S. British, and Australian standards for particulate contamination of large-volume parenteral solutions Maximum allowable particlesfml Particle size (jUrI)

u.s.

pharmacopeia

British pharmacopeia

500 80

~2 ~5

~IO

50

~W

5

Australian Health Dept. 100

2

- , Indicates no set limit.

Table III. Particle counts performed on St. Thomas's Hospital cardioplegic solution in the present study compared with the solution tested by Hearse and

colleagues'

Particle coumfml

Particle size

Present study

Hearse study

2-5 5-10 10-15* >15*

36 7 2

773

Total

47

(um}

2

188 7 2 970

'Particles larger than 10 I'm have been reported to constrict coronary arteries in the isolated rat heart.

additional amountsoriginating in the cardioplegia delivery set. As expected, the 0.2 #Lm filter was effective in removing theseparticles. Countsin the RoyalChildren's Hospitalsolution weresimilarto thosein theSt. Thomas' Hospital solution, with 39 ± 5.4 particles larger than 5 #Lm and 2.6 ± 0.1 particles larger than 20 #Lm. The second set of particle counts performed on St. Thomas' Hospital solution (Table III) generally agreed with the first set, but the different counting windows allowed direct comparison with counts performed by Hearse and colleagues.' Although the cardioplegic solution we tested had fewer of the smallerparticles (2 to 5 #Lm), it had comparable numbersof thoseparticles in the criticalrangeof 10 #Lm and larger (ofthe sizereported by Hearse and colleagues as causing coronary vasoconstriction in rats). Rat studies. The meancoronary flow before cardioplegicinfusion was 10.0 ± 0.46ml/minin the filtered group and 9.4 ± 0.46 ml/min in the unfiltered group (differencenot significant). Fig. 1 shows the percentage change in coronaryflow overthe 2Q-minute infusion period and subsequent 5-minute reperfusion period. In both groups there was an initial rise in coronary flow; flow increased after 2 minutes to 11.8 ± 0.53 ml/rnin in the filtered groupand to 12.1 ± 0.92 ml/min in the unfiltered group. Thisinitialrise wasfollowed bya gradualdecline overthe remainder of the infusion period. By20 minutes after the inception of infusion, flow had fallen to 7.4 ± 0.33 ml/min in the filtered groupand to 7.0 ± 0.36 ml/min in the unfiltered group. On reperfusion with warm buffer coronaryflow once again rose, to 9.3 ± 0.34 ml/min in the filtered groupand to 9.0 ± 0.40 ml/min in the unfiltered group. The decline in coronary flow during the infusion of cardioplegic solution was statistically significant for both groups of hearts (p < 0.01), but there was no significant difference between the two groups. Dog studies Constant-pressure study. Fig. 2 shows the coronary flow rate during constant-pressure infusion of unfiltered

cardioplegic solution. As the roller pumpgradually developedpressure, there wasan initial rise in flow rate. After 3 minutes a steady infusion pressure of 50 mm Hg was obtained. Coronary flow remained constant for the remainder of the infusion, implying unaltered coronary vascular resistance. Vasoconstriction induced by the injection of 1mgof norepinephrine produced a reduction in coronary flow from20.6 ± 3.2to 14.1 ± 1.6nil/min/ kg, indicating that the coronary vasculature remained reactive at the end of the cardioplegic infusion. Constant-flow study. Fig. 3 shows the effect on coronary perfusion pressure of infusing cardioplegic solution at a constantflow rate. Therewasno increase in pressure duringthe 1Y2-minute infusion, nor during the l-minute reinfusion. Byimplication there wasno increase in coronary vascular resistance. Clinical human study. The effects on patients of infused filtered and unfiltered cardioplegic solutions is shown in Fig. 4. During the first 20 seconds of the infusion, there was a significant rise in coronary perfusion pressure in both filtered and unfiltered groups (p < 0.001). After 20 seconds, pressure was similar for bothgroups, witha pressure of78.9 ± 4.3mmHg forthe group receiving filtered solution and a pressure of 75.5 ± 7.2 mm Hg for the group receiving unfiltered solution. After the steepinitialincrease in pressure, there was a second phaseof more gradual increase. Coronary perfusion pressure after 80 seconds of infusion was 95.5 ± 14.3 mm Hg for the group receiving unfiltered solution and 101 ± 15.3 mm Hg for the groupreceiving filtered solution. There was no significant difference between groups. Discussion Garvan and Gunner;' in Australia in 1964, were the firstto express concern about the presence of particles in intravenous fluids. They identified a variety of contami-

Volume 101

Particles in cardioplegic solution 4 7 7

Number 3 March 1990

"

120

60

r

E

.§

:c

50

E

.s

"'5

'" 40

a'""c:

~

80

c

60

0

30

a:

20

0 .(jj :::l

D.-

Ql

sa.

»,

~0

10

o

0

100

:::l UJ UJ

.2 iii

·iii

I I-I-LLr r. J I r.-/ /1 1.)....1.... --(l'T'l"r,r-r--r J~1"I'-!"I 1

CD

40

»

ac g

(;

=

<, /

1 .-

Vl

-

Filtered

---- Unfiltered

(1

20

U

Initial Infusion

Reinfusion Time (seconds)

Fig. 3. Coronary pressure during infusion of unfiltered cardioplegic solution at a constant flow rate into greyhound hearts during the induction of arrest and during subsequent reinfusion. Each point represents the mean of the values for eight hearts and the bars indicate the SEM.

nantsincommercially preparedintravenous solutions and speculated about their potential hazards, as have subsequent authors.t '? The infusion of fluids contaminated with particles of starch, silica, or glasshas been implicated in damage to the brain, kidneys, spleen, and liver. I I This phenomenon appears to have been uncommon, presumablybecauseof the filtering effects of the lungs.Subsequently, efforts by pharmaceutical companies and the introduction of quality-control regulations have combined to produce refinements in formulation, filtration, and packaging. These refinements have dramatically reduced particulate contaminationin modern parenteral fluids. 4, 5 However, the widespread use of hypothermic cardioplegic arrest in cardiacsurgeryin whichthe cardioplegic solution is infused directly into the coronary circulation has led to a renewed interest in possible particulate contamination ofparenteralfluids. In 1981 Maclzonald!2 reported particulatedebrisin crystalloid cardioplegia but madeno commenton its biologic effects. In 1984Robinson,Braimbridge, and Hearse! raised the possibility that particulatecontamination of cardioplegic solutions could damage the heart. With isolated rat hearts and a modified Langendorffperfusion method, they found that continuous infusion of unfiltered cardioplegic solutions at constant pressure resulted in a progressive reduction in coronary flow; filtration of these solutions through a 0.8 JLm filterconsiderably reduced, but did not entirelyeliminate, this effect. They also found, in a separate study of hypothermic ischemia with multidose cardioplegia, that hearts receiving filtered solutions recovered 90%of function after 180 minutes of ischemia, whereas hearts receiving unfiltered solutions failed to recover at all.

o 10

20

30

40

50

60

70

80

Crossclamp time (seconds)

Fig. 4. Clinical study during coronary bypass surgery. Mean coronary perfusion (aortic root) pressure during infusion at a constant flow rate of filtered cardioplegic solution in 10 patients and unfiltered solution in 10 patients. Each point represents the mean and the bars indicate the SEM.

Hearse and coworkers? suggested in a later study that the reductionin coronaryflow wasdue to activevasoconstriction. Subsequentinfusion offiltered solutionreversed this impairmentofftow.Thus it is unlikely that this effect resultedfrom physical occlusion of the coronaryvasculature. They also found that vasoactive agents, such as nifedipine, procaine, and adenosine triphosphate,reduced this effect. In a series of studies with filters of varying porosity, they also showed that this transient coronary vasoconstriction wascaused by a relatively small number of particlesgreater than 10 JLIn in diameter. The authors concluded that, in order to optimize myocardial preservation, consideration should be given to the use of a microfilter in the cardioplegic infusion line. Nonetheless, they recognized the limitations of the isolated rat heart modeland emphasized the need both for further studies ofother animal species and for clinicalstudiesof humans. The relevance of thesetwostudiesofisolatedrat hearts to clinical situations is questionable. Both studies used cardioplegic solutions at a higher temperature (20 0 C), a lower infusion pressure, and a much higher dose than usual in clinical practice. The usual clinical dosage of cardioplegic solution is 1000rnIj300 gm heart tissueover a 2-minute infusion period (3.33 rnIjgm). During the courseof theseexperiments in rats, each heart received an average volume of 178 rnl cardioplegic solutionover 20 minutes. Assuminga heart weight of 1 gm, this volume represents more than 50 times the clinicaldose. The onlyother studyin this field of whichweare aware was reported by Sellevold and Jynge.':' With a rat heart model almost identicalto that of Robinson and cowork-

478

Munsch et al.

ers, they also showed that filtration of cardioplegic solution partially eliminated impairment of coronary flow, and that the additionof steroids had a vasodilating effect. These authors also carefully emphasized the danger of applyingresultsfrom experiments performedon onespeciesto the treatment of another. Despitethesecautionary notes, it was not long beforecardioplegic microfilters for clinical use were produced. After persuasive marketing their use rapidly became routine in many cardiac surgical units. We felt that these rat heart studieswere enlightening but did not provide sufficient evidence to justify the clinical use of a microfilter. The aim of the present study, therefore, was to determine the extent and clinical relevanceof particle-induced coronaryvasoconstriction, and to determine whether a microfilter in the cardioplegia infusion lineis clinically necessary. This aim wasaccomplished through a seriesofseparateexperiments. First,we determined the level and sourceof particulate contamination in two commercially produced cardioplegic solutions.Second,we attempted to duplicatethe experiments performedby Robinson and Coworkers on the rat heart. Third, we carried out two studieson the effects of unfiltered cardioplegic solution on caninehearts. Finally, and probablymostimportant,weexamined the clinical effects of filtered and unfiltered solutions on coronaryvascular resistance in humans undergoing coronary bypass surgery. We found that the cardioplegic solutions we tested contained,as expected, a low level of particulate debris, wellwithinthe standards prescribed bythe formularies of the United States and United Kingdom.v ' Compared with the solutions tested by Hearse and colleagues, the solutions we tested containedfewer small particles(<10 JLm in diameter) but similar numbers of large particles (> 10 JLm in diameter). Only particles of this latter size were shownin Hearse and coworkers' study- as causing coronary vasoconstriction in the isolated rat heart; we therefore believed the solutions were comparable and should produce similar biologic effects. The precipitation of unsoluble particles of calcium carbonate is a theoretic possibility after the addition of sodium bicarbonate to a Ringer's solution base containing calcium chloride. Adding a wide range of sodium bicarbonate concentrations to Ringer's solution in our laboratory, we found no microscopic evidence of precipitation of calcium carbonate. As shown in Table I, the additionof 10 mmolsodiumbicarbonateto 1 L Ringer's solution did not increaseparticlecount for any of the size ranges.Presumably, therefore, the concentrations of calcium chlorideand sodiumbicarbonateused in preparing

The Journal -ot Thoracic and Cardiovascular Surgery

St. Thomas' solution do not reach the level required to induceprecipitation. In our studyof the isolated rat heart, wesaw an initial transient increase in coronary flow during cardioplegic infusion, followed by a steady gradual decline. The pattern of this change in coronary flow was essentially the same as that in Hearse and coworkers' experiments with unfiltered solutions. However, we saw no difference between hearts receiving filtered solution and hearts receiving unfiltered solution, indicating that, although infusion of cardioplegic solution produced coronaryvasoconstriction in the rat heart, this effect was not caused by particulatecontamination of the solution. In addition, in the constant-pressure infusion study of dogs modeled on Robinson and coworkers' rat experiments, unfiltered cardioplegic solution did not produce any vasoconstriction. In the second series of experiments with dogs, we attempted to approximatemoreclosely the methods routinelyemployed in heart operations. Under these conditions, we still did not detect an increase in coronaryvascular resistance either during infusion or during subsequent reinfusion with unfiltered cardioplegic solution. The final and crucial portion of this study was the investigation of the clinical effecton coronaryresistance of filtering cardioplegic solution during coronary artery bypass surgery. We saw an immediateincreasein aortic pressure during the initialphaseof cardioplegic infusion, whichwas the same for both filtered and unfiltered solutions. Presumably, the time needed for the pump to generate a constant pressure and for the aortic root to fill played some part in this initial increase. Pressure rose more slowly for the remainder of the infusion period in both groups,thus indicating an increasein coronary vascular resistance. A commercial cardioplegia filterdid not affectcoronaryvascularresistance in this clinical setting. Besides the putative effects of particles, several other well-known factorsundoubtedly affectcoronaryvascular resistance duringcardiopulmonary bypass and cardioplegic infusion. Local controlof the coronary circulation is closely relatedto myocardial metabolism. Adjustments in coronary vascular resistance resulting from changes in myocardial oxygen content are probably mediated by wayofvasoactive metabolites suchas adenosine." Hypothermia produces coronaryvasodilatation, both in isolated coronary arteries" and in the working heart," presumably because of the direct effect of temperature on vascular smooth muscle. Hypothermia attenuates or abolishes autoregulation of the coronaryvasculature. 17, 18 In general,experimental interventions that alter myocar-

Volume 101 Number 3 March 1990

dial oxygen consumption have variable effects on the degreeof autoregulation. 19 It wouldthereforebe difficult to predictthe state of the coronarycirculation during the induction and maintenance of hypothermic, ischemic arrest. Hiratzka and colleagues'? have recently shown that a reduction in coronary vascular resistance occurs after bypass and cardioplegic arrest in both humans and dogs, and that progressive coronary vasodilatation also occurs in anaesthetized dogswithopenchestssimplywith the passage of time. The effects of intracoronaryadministrationof potassium on the coronaryvasculaturewerestudied by Murray and Sparks." Bolus injections of 40 ~mol KCI produced a transient increasein bloodflow. This result appears to confirm the conventional viewof potassium in low concentrations as a vasodilator. However, work in our institutionhasshown that potassiumin higherconcentrations, similarto thosefound in cardioplegic solutions ( -1.6 log M,or 25mmoljL), isa potentcoronaryvasoconstrictor. 22 The addition of procaine to the cardioplegic solution at the usual concentration of 1 mmoljL probablyhas little, if any, protective effect against coronary spasm; Hearse and colleagues- found that the maximum vasodilator effect of procainewas attained at a concentration of 10 mmoljL. In the clinicalsituationthere are additionalfactorsthat couldaffectcoronaryvascularresistance. The severity of the coronary artery disease and the extent of collateral development have variable effects on the resistance to cardioplegic infusion and the distribution of the solution. It is clear that any or all of these factors could have played complementary or conflicting roles in our experiments. In addition, differences between animal models and humans must be taken into consideration. This consideration does not, however, obscure the importance of the lack of evidence for particle-induced coronary vasoconstriction in any of our studies, including the clinical studies. There have been reports of septicemia in patients who have received bacteriologically contaminated cardioplegic solutions.P: 24 Theoretically, a cardioplegia filter will remove microorganisms in addition to particulate debris, but filtration cannot be assumed to ensure sterility and doesnot remove preformedtoxins. Microbiologic control measures duringpreparationare far moreimportantthan is filtration during delivery.i" What is the relevance of these studies of crystalloid cardioplegia to the technique of blood cardioplegia now used so widely? Obviously it is not possible to use a microfilter in the administrationlinefor bloodcardioplegia. Nevertheless, if harmful particleswerepresentin the

Particles in cardioplegic solution 4 7 9

components of a bloodcardioplegic solution, these would find their way into the coronary microcirculation just as witha crystalloid solution. The onlywayto overcome such a problem would be to pass the components through a microfilter before mixing them with the blood. Our studies indicate that neither the base solution nor the additives for cardioplegic solutions contain harmful particles. Therefore such filtering proceduresare not necessary for either bloodor crystalloid cardioplegia. Similar considerations applyto solutions usedto prime the heartj lungmachine,sincethe bypassprime reachesthe cardiac and other microcirculations without passingthrough the pulmonary"filter." Although there is someevidence that it is beneficial to recirculatethe primingsolutionthrough a 5 ~m filter to removedebris from the oxygenator and associated components'> beforethe start of cardiopulmonary bypass, there is no evidence that priming solutions shouldbe passedthrougha 0.8 ~m filtersuchas that often advocatedfor cardioplegic solutions. In today's financial climate, cardiac surgical units are under ever-increasing pressure to contain expenditure. The useof unnecessary equipmentis an unwelcome additionalexpense. It isapparent that cardioplegic filterswere the subject of aggressive marketing techniquesafter the publication of a series of experiments of limited clinical relevance and before investigation of their value in clinicalpractice.We haveattempted to resolve this difficulty. On the basis of our studies, we concludethat the use of a microfilter in the cardioplegic infusion line during cardiac surgical procedures is not clinically justified. The authors gratefully acknowledge the assitance of James McMillan, Christine Boyes,Lesley Langley, and Dr. Ian Smith. Dr. Steven Farrish of Monash Medical School provided assistance with statistical analysis. Weare also indebted to the statT of David Bull Laboratories and to Professor J. Ralston of the South Australia Institute of Technology for assistance in particle counting. The Pall cardioplegia filters were kindly donated by Cardiac and Surgical Australia Pty, Camberwell, Yictoria, Australia. REFERENCES 1. Robinson LA, Braimbridge MY, Hearse DJ. The potential hazard of particulate contamination of cardioplegic solutions. J THORAC CARDIOVASC SURG 1984;87:48-58. 2. Hearse DJ, Erol C, Robinson LA, Maxwell MP, Braimbridge MY. Particle-induced coronary vasoconstriction during cardioplegic infusion. Characterization and possible mechanisms. J THoRAc CARDIOVASC SURG 1985;89:428-

38. 3. Sensor theory and calibration. HIAC PC320 Particle Counter instruction book. Menlo Park, California: HIAC/ Royco Instruments Division.

The Journal of Thoracic and Cardiovascular Surgery

480 Munsch et 01.

4. The United States Pharmacopeia, vol 21. 21st revision. Rockville, Maryland: United States Pharmacopeia Convention Inc, 1985:1257. 5. The British Pharmacopoeia, vol II. London: Her Majesty's Stationery Office, 1980: Appendix XIII, A120. 6. LangendorffO. Untersuchungen am iiberlebeden Saugethierherzen. Pflugers Arch 1895;61:281-332. 7. Garvan JM, Gunner BW. The harmful effects of particles in intravenous fluids. Med J Aust 1964;2:1-6. 8. Dratz RG, Graf J. Identifying particle contaminants. Bull Parenter Drug Assoc 1974;28:35-52. 9. Leong AS. Particulate contamination in intravenous therapy and extracorporeal systems. Med J Aust 1982:2:30910. 10. Falchuk KH, Peterson L, McNeil BJ. Microparticulateinduced phlebitis. Its prevention by in-line filtration. N Engl J Med 1985;312:78-82. 11. Turco S, Davis NM. Clinical significance of particulate matter: a review ofthe literature. Hosp Pharm 1973;8:13740. 12. MacDonald JL. Is crystalloid cardioplegia a source of particulate debris? Am Acad Cardiovasc Perfus Proc 1981;2:20-24. 13. Sellevold OFM, Jynge P. Steroids and cardioplegia: effects of glucocorticoids upon vascular resistance during cardioplegic perfusion. Thorac Cardiovasc Surg 1987;35:30711. 14. Rubio R, Berne RM. Release of adenosine by the normal myocardium in dogs and its relationship to the regulation of coronary resistance. Circ Res 1969;25:407-15. 15. Cruickshank EWH, Subba Rau A. Reactions of isolated systemic and coronary arteries. J Physiol (Lond) 1927;64:65-77. 16. Hardin RA, Scott JB, Haddy FJ. Effect of cardiac cooling

17.

18.

19. 20.

21.

22.

23.

24.

25.

on coronary vascular resistance in normothermic dogs. Am J Physiol 1960;199:163-6. Badeer HS. Influence of cooling the heart on reactive hyperemia of the coronary bed in the heart-lung preparation. Circ Res 1965;16:19-25. Chitwood WR, Sink JD, Hill RC, Wechsler AS, Sabiston DC. The effects of hypothermia on myocardial oxygen consumption and transmural coronary blood flow in the potassium-arrested heart. Ann Surg 1979;190:106-16. Dole WP. Autoregulation of the coronary circulation. Prog Cardiovasc Dis 1987;29:293-323. Hiratzka LF, Eastham CL, Carter JG, et al. The effects of cardiopulmonary bypass and cold cardioplegia on coronary flow velocity and the reactive hyperemic response in patients and dogs. Ann Thorac Surg 1988;45:474-81. Murray PA, Sparks HV. The mechanism of Kt-induced vasodilation of the coronary vascular bed of the dog. Circ Res 1979;42:35-42. He G-W, Angus JA, Rosenfeldt FL. Reactivity of the canine isolated internal mammary artery, saphenous vein, and coronary artery to constrictor and dilator substances: relevance to coronary bypass graft surgery. J Cardiovasc PharmacoI1988;12:12-22. Hughes CF, Grant AF, Leckie BD, Baird DK. Cardioplegic solution: a contamination crisis. J THORAC CARDIaVASC SURG 1986;91:296-302. Talbot GH, Miller DE, Doorley M, et al. Enterobacter cloacae--contaminated cardioplegic solution. Am J Infect Control 1984;12:239-44. Brooks JD, Beauchamp RA, Magrath RA, Clark RE. Efficacy of particulate removal by a prebypass filter with different oxygenation systems.. J Extracorp. Technol 1979;11:175-82.