Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
CLINICAL TOXICOLOGY 8(6), pp. 623-631 (1975)
Central Nervous System Salicylate
JAMES R. REED, Ph.D., and PAUL A. PALMISANO, M.D.
Departments of Pharmacology and Pediatrics University of Alabama School of Medicine Birmingham, Alabama
INTRODUCTION Therapy for acute salicylate overdose has long been complicated by the poor correlation between serum salicylate levels and clinical severity. Efforts to resolve this problem resulted in the Done monogram, which is based on the observation that extrapolated time zero serum salicylate levels do, in fact, correlate well with the severity of symptoms [l] As yet, there is limited information to explain why these initial serum levels a r e more useful prognostically. However, some recent evidence does indicate that the initial salicylate dose influences the volume of distribution (Vd), which is directly proportional to the dose [2]. Two known factors that strongly influence the salicylate Vd are plasma protein binding and plasma pH. McCann and Palmisano [3] demonstrated that salicylate Vd in dogs is relatively constant over a dose range of 27 to 147 mg/kg if Vd is calculated from serum unbound levels rather than total salicylate levels. Thus, protein binding is implicated as a major determinate of Vd. Protein binding is based on mass action kinetics, hence a potential reason for the initial concentration dependence of Vd. Hill [4] has demonstrated in r a t s that acidosis will increase the
.
623 Copyright 0 1976 by Marcel Dekker, Inc. All Rights Reserved. Neither this work nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the publisher.
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
624
REED AND PALMISAN0
entrance of salicylate into the CNS from the intravascular space. The lower pH shifts the ionized to un-ionized salicylate ratio in favor of the un-ionized form. The blood-brain bar rier is more permeable to un-ionized salicylate giving basis to the effect of serum pH on Vd. A movement of salicylate into the CNS is especially pertinent to clinical severity of salicylate overdose. Many of the major symptoms result from direct CNS effects of the drug. Kussmaul respiration is largely due to salicylate interaction with receptors in the medullary respiratory center [5]. Tinnitus, nausea, and vomiting a r e also primarily CNS effects of salicylate [6]. Throughout the clinical course of any salicylate intoxication, there is marked variability in serum pH and in serum protein binding. Clearly, this could account for the frequent disagreement between total serum salicylate levels and clinical toxicity. CSF salicylate determinations would provide more accurate assessment of toxicity by circumventing the acid-base and protein binding variables that a r e involved as one relies on total serum salicylate data. However, serial lumbar puncture cannot be routinely performed in salicylatepoisoned patients. A compromise would be the frequent measurement of unbound salicylate concentration in serum. This study purports to describe in dogs the relationship between serum-unbound salicylate levels and CSF salicylate levels during changes in serum protein binding and serum pH. In this manner, the applicability of using serum unbound levels to approximate CSF levels will be assessed. M A T E R I A L S AND M E T H O D S The protocol for a typical in vivo experiment w a s as follows: A 23-kg female mongrel dog w a s anesthetized with pentobarbital (30 mg/kg IV). A priming intravenous salicylate dose of 125 mg/kg (expressed as the free acid) was given at time zero. After 30 min a salicylate intravenous infusion at the rate of 1.28 mg/min was initiated and was continued throughout the duration of the experiment. Blood samples were taken at timed intervals from a femoral venous catheter. Simultaneous cerebrospinal fluid samples were obtained by a cisterna magna puncture. Serum was prepared from each blood sample and subjected to the ultrafiltrate procedure described by McCann [7] for determination of bound and unbound salicylate concentration using the Trinder [8] colorimetric method. The salicylate concentration in cerebrospinal fluid was also measured by the same method. In three of the dogs so studied, 100 mg/kg of sulfinpyrazone was administered by rapid IV push a t 420 min after the initial salicylate
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
CNS SALICYLATE
625
injection. Serum samples taken subsequent to sulfinpyrazone injection were subjected to the determination of Burns 191 for measurement of total sulfinpyrazone concentration. Drug determinations for both salicylate and sulfinpyrazone were adapted to microvolumetric quantities to avoid volume depletion of the dogs. In another dog, instead of sulfinpyrazone, 20 mg/kg of acetazolamide w a s administered intravenously at 307 min after initial salicylate injection. In this dog, serum arterial pH and p C 0 ~were measured at regular intervals throughout the experiment both before and after acetazolamide administration. As before, salicylate concentration in total serum, serum water, and CSF were also measured a t regular intervals over the course of the experiment. RESULTS Continuous single animal experiments were performed in which salicylate concentrations were followed closely in total serum, serum water (unbound salicylate), and cerebrospinal fluid before and after sulfinpyrazone administration. Figure 1 depicts the data from a representative experiment utilizing a 23-kg female mongrel dog anesthetized with pentobarbital sodium. Time zero i s the point at which a single dose of 125 mg/kg of salicylate was administered followed by a constant salicylate infusion. The concentration of salicylate as total serum salicylate, as unbound salicylate, and as cerebrospinal fluid salicylate a r e plotted against time from the initial salicylate injection. At 420 min after initial salicylate injection, sulfinpyrazone 100 mg/kg was administered by rapid intravenous injection. The sulfinpyrazone total serum concentration measured after this injection is included for reference. Since ample equilibration time was allowed, the relationsip w a s fairly constant among total serum salicylate, unbound serum salicylate, and cerebrospinal fluid salicylate prior to sulfinpyrazone administration. However, subsequent to intravenous sulfinpyrazone administration, there was a profound decrease in the total serum salicylate level. The total serum salicylate decrease was accompanied by an increase in the concentration of unbound serum salicylate. Moreover, there was a corresponding increase in cerebrospinal fluid salicylate concentration. The CSF salicylate concentration w a s reflected reasonably well by the unbound salicylate concentration both before and after sulfinpyrazone administration; whereas total serum salicylate concentration was decreasing over the time period in which cerebrospinal fluid salicylate concentration was increasing.
‘., REED AND PALMISANO
626
“1
[MGK] SULFINPYAAZONE
I I
30t
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
TOTAL SERUM
A SULFINPYRAZWE
I
20
I
TOTAL SERUM
ASALICYLATE UNWUNO SERUM SALK YLATE
I
I
OCSF SALICYLATE
I
1.28nq/rnin Salicy(olel I
0
0
SALICYLATE
125 m g h
I
200
I
1
I
1
500
6oo
700
300
*O%
SULFINPYRAZONE
100 m g F g
TIME (MINI
FIG. 1. Relation of unbound serum salicylate to CSF and total serum salicylate in the dog, before and after displacement of salicylate from albumin binding sites by sulfinpyrazone. Figure 2 depicts an experiment in which a similar protocol was followed except acetazolamide 20 mg/kg was administered at 307 min instead of sulfinpyrazone. Salicylate concentrations were closely followed in total serum, serum water (unbound salicylate), and cerebrospinal fluid before and after acetazolamide injection. Serum arterial pH and p C 0 ~determinations before and after acetazolamide administration a r e included on the same figure. Note that changes in cerebrospinal fluid salicylate concentrations agree with those measured in
627
CNS SALICYLATE
Acetatolamide 20
40 Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
Arterial
Pcoz
30
mg/kg
[
+
?I
20
Ar teriol PH
7.20
30 Salicylate mg O/O
Unbound
20
P
\
10
I
1.28 mg/min Salicylate Infusion
I
I I
I/
200
rlat e 125 w / k g
300 400 Time ( m i d
500
FIG. 2. Relation of unbound serum salicylate to CSF and total serum salicylate in the dog, before and after induction of acidosis by acetazolamide.
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
628
REED
AND PALMISANO
serum water until the administration of acetazolamide. Concurrent with the pH decrease from 7.32 to 7.24, the serum unbound salicylate concentration decreased over the period in which cerebrospinal fluid salicylate concentrations were actually rising, Total serum salicylate concentration declines remarkably during the time period in which salicylate concentration in the cerebrospinal fluid was increasing. DISCUSSION
The concentration of salicylate within the central nervous system may well be the critical determinant of aspirin toxicity [4]. Major symptoms of salicylate toxicity, i.e., Kussmaul respiration, nausea, vomiting, and convulsions all a r i s e as salicylate concentration inc r e a s e s in the cerebrospinal fluid [6]. Heretofore, most efforts to a s s e s s clinical salicylate toxicity have focused upon total salicylate concentration measured in whole serum. These values often correlate poorly with cerebrospinal fluid concentrations because of at least two factors that affect the exchange of salicylate molecules between the vascular space and the cerebrospinal fluid. These are: (a) serum protein binding [lo], and (b) pH-dependent ionization of the salicylate molecule [ 113. Clearly, true cerebrospinal fluid salicylate levels would be the most prognostically valuable laboratory data. However, serial lumbar punctures are unfeasible in routine management of patients. It would appear that serum unbound salicylate determinations would be a reasonable approximation of the more useful but invasive CSF determinations and much more clinically pertinent than the presently employed total serum salicylate levels. There is growing evidence that unbound serum salicylate concentration should be considered a good indicator of clinical toxicity. Recent data indicate that the initial Vd of salicylate correlates well with the clinical severity of intoxication [2]. Implicit in this understanding is that a greater Vd represents a greater central nervous system involvement. There is evidence that Vd calculated from serum unbound salicylate values is valid and constant whereas that calculated from total serum values is misleading [3]. This also indicates that unbound serum salicylate concentrations should be considered as a clinical measure of salicylate toxicity. Our study in dogs has evaluated the relative concordance between cerebrospinal fluid and serum unbound salicylate concentration throughout the course of experimental salicylism. Changes in protein binding and changes in serum pH were experimentally induced to determine if
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
CNS SALICYLATE
629
cerebrospinal fluid and serum unbound salicylate concentration might deviate, causing serum unbound salicylate concentration to become misrepresentative of cerebrospinal fluid salicylate concentration. Actually, in the absence of binding changes o r pH changes, serum unbound salicylate concentration parallelled cerebrospinal fluid salicylate concentration rather closely; the absolute difference in these concentrations being explained by the impermeability of the bloodbrain b a r r i e r to ionized salicylate molecules. As salicylate was displaced from the serum albumin binding sites by sulfinpyrazone, the secondary increase in serum unbound salicylate in serum water was rapidly reflected by an immediate increase in CSF salicylate concentration. Accordingly, the constant relationship of these concentrations w a s remarkably retained, even as protein binding of salicylate was drastically altered. Most impressive, however, w a s the effect of acidosis on the otherwise parallel relationship between serum unbound salicylate concentration and CSF salicylate concentration. A s a change in serum pH (7.32 to 7.24) w a s elicited by acetazolamide administration, serum unbound salicylate concentration decreased. Presumably, the more permeable un-ionized molecules increased in concentration at the lower pH and diffused rapidly from the vascular compartment into the CSF. Figure 2 indicates that this is, indeed, the case, since CSF salicylate concentration increased with the acute change from the ionized to the un-ionized state. Thus, the increasing Vd of salicylate w a s reflected by increased salicylate levels in the CSF. It is therefore most important to note that serum unbound salicylate does not always accurately reflect the situation in the central nervous system. In the presence of acidosis, CSF salicylate may actually increase, as serum unbound concentration is falling. Even with this obvious limitation, unbound salicylate determinations would much more accurately reflect the situation within the CNS than the standard total levels as presently employed. Clearly, under present practices, the therapist i s frequently deceived by inappropriately low total serum concentration in the face of rising levels in the CNS. SUMMARY The poor correlation between clinical salicylate toxicity and serum blood levels is reapproached in light of recent evidence linking clinical severity with initial volume of distribution (Vd). It is recognized that two variables alter salicylate Vd in such manner that serum salicylate levels are misleading (thus, the change in Vd i s not detected by
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
630
REED AND PALMISANO
present methods). These variables are serum protein binding and the pH-dependent ionized/un- ionized ratio in the unbound salicylate fraction. Measurement of salicylate concentration in the cerebrospinal fluid (CSF) would circumvent these variables, but would be clinically impractical. Thus, an alternative is sought to the inexact total serum salicylate levels and the impractical CSF salicylate levels for assessment of the severity of salicylate poisoning. This study indicates that, in dogs, serum unbound salicylate levels closely reflect CSF salicylate levels, even as a decrease in serum protein binding is in progress. However, serum unbound salicylate concentration does not reflect CSF salicylate concentration as a decrease in serum pH is elicited (CSF salicylate actually increased as serum unbound salicylate decreased). On the other hand, serum unbound salicylate measurement would seem preferable to total serum salicylate measurements now used in that the total value decreased markedly as either protein binding change o r acidosis produced a change in distribution and the resultant increase in CSF salicylate. REFERENCES
[ 11 A. K. Done, Salicylate intoxication: Significance of measure121
[3] [4] [5] [6]
[7] [8]
ments of salicylate in blood in cases of acute ingestion, Pediatrics, 26, 800 (1960). G. Levy and%. J. Yaffe, Relation between dose and apparent volume of distribution of salicylate in children, Pediatrics, 54, 713 (1974). i?l; P. McCann and P. A. Palmisano, Salicylate pharmacokinetics in the dog, Res. Commun. Chem. -Pathol. Pharmacol., 5, 17 (1973). 3. B. Hill, Salicylate intoxication, New Engl. J. Med., 288, 1110 (1973). S. M. 'Te&ey and R. M. Miller, The respiratory and circulatory actions of salicylate, Am. J. Med., 19, 498 (1955). D. M. Woodbury, "Analgesic- Antipyretics, Anti-Inflammatory Agents, and Inhibitors of Uric Acid Synthesis," in The Pharmacological Basis of Therapeutics, 4th ed. (L.Goodman and A. Gilman, ed.), MacMillan, New York, 1970, p. 317. W. P. Mccann, Problems in data collection and analysis in human pharmacokinetics, Drug Information Bull., 3,52 (1969). P. Trinder, Rapid determination of salicylate in biological fluids, Biochem. J., 57, 301 (1954).
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
CNS SALICYLATE
631
[9] J. J. Burns, T. F. Yu, A. Rittenband, J. M. Perel, A. B. Gutman, and B. B. Brodie, A potent new uricosuric agent, the sulfoxide metabolite of the phenylbutazone analogue, J. Pharmacol. Exper. 418 (1957). Therap., 19, [ 101 C. Davison and R. L. Strauss, The effect of the distribution and excretion of salicylate by drugs displacing that compound from plasma binding sites, The Pharmacologist, 3,81 (1961). [ 111 J. B. Hill, Experimental salicylate poisoning: observations on the effects of altering blood pH on tissue and plasma salicylate concentrations, Pediatrics, 47, 658 (1971).
CLINICAL TOXICOLOGY 8(6), pp. 623-631 (1975)
Central Nervous System Salicylate
JAMES R. REED, Ph.D., and PAUL A. PALMISANO, M.D.
Departments of Pharmacology and Pediatrics University of Alabama School of Medicine Birmingham, Alabama
INTRODUCTION Therapy for acute salicylate overdose has long been complicated by the poor correlation between serum salicylate levels and clinical severity. Efforts to resolve this problem resulted in the Done monogram, which is based on the observation that extrapolated time zero serum salicylate levels do, in fact, correlate well with the severity of symptoms [l] As yet, there is limited information to explain why these initial serum levels a r e more useful prognostically. However, some recent evidence does indicate that the initial salicylate dose influences the volume of distribution (Vd), which is directly proportional to the dose [2]. Two known factors that strongly influence the salicylate Vd are plasma protein binding and plasma pH. McCann and Palmisano [3] demonstrated that salicylate Vd in dogs is relatively constant over a dose range of 27 to 147 mg/kg if Vd is calculated from serum unbound levels rather than total salicylate levels. Thus, protein binding is implicated as a major determinate of Vd. Protein binding is based on mass action kinetics, hence a potential reason for the initial concentration dependence of Vd. Hill [4] has demonstrated in r a t s that acidosis will increase the
.
623 Copyright 0 1976 by Marcel Dekker, Inc. All Rights Reserved. Neither this work nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the publisher.
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
624
REED AND PALMISAN0
entrance of salicylate into the CNS from the intravascular space. The lower pH shifts the ionized to un-ionized salicylate ratio in favor of the un-ionized form. The blood-brain bar rier is more permeable to un-ionized salicylate giving basis to the effect of serum pH on Vd. A movement of salicylate into the CNS is especially pertinent to clinical severity of salicylate overdose. Many of the major symptoms result from direct CNS effects of the drug. Kussmaul respiration is largely due to salicylate interaction with receptors in the medullary respiratory center [5]. Tinnitus, nausea, and vomiting a r e also primarily CNS effects of salicylate [6]. Throughout the clinical course of any salicylate intoxication, there is marked variability in serum pH and in serum protein binding. Clearly, this could account for the frequent disagreement between total serum salicylate levels and clinical toxicity. CSF salicylate determinations would provide more accurate assessment of toxicity by circumventing the acid-base and protein binding variables that a r e involved as one relies on total serum salicylate data. However, serial lumbar puncture cannot be routinely performed in salicylatepoisoned patients. A compromise would be the frequent measurement of unbound salicylate concentration in serum. This study purports to describe in dogs the relationship between serum-unbound salicylate levels and CSF salicylate levels during changes in serum protein binding and serum pH. In this manner, the applicability of using serum unbound levels to approximate CSF levels will be assessed. M A T E R I A L S AND M E T H O D S The protocol for a typical in vivo experiment w a s as follows: A 23-kg female mongrel dog w a s anesthetized with pentobarbital (30 mg/kg IV). A priming intravenous salicylate dose of 125 mg/kg (expressed as the free acid) was given at time zero. After 30 min a salicylate intravenous infusion at the rate of 1.28 mg/min was initiated and was continued throughout the duration of the experiment. Blood samples were taken at timed intervals from a femoral venous catheter. Simultaneous cerebrospinal fluid samples were obtained by a cisterna magna puncture. Serum was prepared from each blood sample and subjected to the ultrafiltrate procedure described by McCann [7] for determination of bound and unbound salicylate concentration using the Trinder [8] colorimetric method. The salicylate concentration in cerebrospinal fluid was also measured by the same method. In three of the dogs so studied, 100 mg/kg of sulfinpyrazone was administered by rapid IV push a t 420 min after the initial salicylate
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
CNS SALICYLATE
625
injection. Serum samples taken subsequent to sulfinpyrazone injection were subjected to the determination of Burns 191 for measurement of total sulfinpyrazone concentration. Drug determinations for both salicylate and sulfinpyrazone were adapted to microvolumetric quantities to avoid volume depletion of the dogs. In another dog, instead of sulfinpyrazone, 20 mg/kg of acetazolamide w a s administered intravenously at 307 min after initial salicylate injection. In this dog, serum arterial pH and p C 0 ~were measured at regular intervals throughout the experiment both before and after acetazolamide administration. As before, salicylate concentration in total serum, serum water, and CSF were also measured a t regular intervals over the course of the experiment. RESULTS Continuous single animal experiments were performed in which salicylate concentrations were followed closely in total serum, serum water (unbound salicylate), and cerebrospinal fluid before and after sulfinpyrazone administration. Figure 1 depicts the data from a representative experiment utilizing a 23-kg female mongrel dog anesthetized with pentobarbital sodium. Time zero i s the point at which a single dose of 125 mg/kg of salicylate was administered followed by a constant salicylate infusion. The concentration of salicylate as total serum salicylate, as unbound salicylate, and as cerebrospinal fluid salicylate a r e plotted against time from the initial salicylate injection. At 420 min after initial salicylate injection, sulfinpyrazone 100 mg/kg was administered by rapid intravenous injection. The sulfinpyrazone total serum concentration measured after this injection is included for reference. Since ample equilibration time was allowed, the relationsip w a s fairly constant among total serum salicylate, unbound serum salicylate, and cerebrospinal fluid salicylate prior to sulfinpyrazone administration. However, subsequent to intravenous sulfinpyrazone administration, there was a profound decrease in the total serum salicylate level. The total serum salicylate decrease was accompanied by an increase in the concentration of unbound serum salicylate. Moreover, there was a corresponding increase in cerebrospinal fluid salicylate concentration. The CSF salicylate concentration w a s reflected reasonably well by the unbound salicylate concentration both before and after sulfinpyrazone administration; whereas total serum salicylate concentration was decreasing over the time period in which cerebrospinal fluid salicylate concentration was increasing.
‘., REED AND PALMISANO
626
“1
[MGK] SULFINPYAAZONE
I I
30t
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
TOTAL SERUM
A SULFINPYRAZWE
I
20
I
TOTAL SERUM
ASALICYLATE UNWUNO SERUM SALK YLATE
I
I
OCSF SALICYLATE
I
1.28nq/rnin Salicy(olel I
0
0
SALICYLATE
125 m g h
I
200
I
1
I
1
500
6oo
700
300
*O%
SULFINPYRAZONE
100 m g F g
TIME (MINI
FIG. 1. Relation of unbound serum salicylate to CSF and total serum salicylate in the dog, before and after displacement of salicylate from albumin binding sites by sulfinpyrazone. Figure 2 depicts an experiment in which a similar protocol was followed except acetazolamide 20 mg/kg was administered at 307 min instead of sulfinpyrazone. Salicylate concentrations were closely followed in total serum, serum water (unbound salicylate), and cerebrospinal fluid before and after acetazolamide injection. Serum arterial pH and p C 0 ~determinations before and after acetazolamide administration a r e included on the same figure. Note that changes in cerebrospinal fluid salicylate concentrations agree with those measured in
627
CNS SALICYLATE
Acetatolamide 20
40 Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
Arterial
Pcoz
30
mg/kg
[
+
?I
20
Ar teriol PH
7.20
30 Salicylate mg O/O
Unbound
20
P
\
10
I
1.28 mg/min Salicylate Infusion
I
I I
I/
200
rlat e 125 w / k g
300 400 Time ( m i d
500
FIG. 2. Relation of unbound serum salicylate to CSF and total serum salicylate in the dog, before and after induction of acidosis by acetazolamide.
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
628
REED
AND PALMISANO
serum water until the administration of acetazolamide. Concurrent with the pH decrease from 7.32 to 7.24, the serum unbound salicylate concentration decreased over the period in which cerebrospinal fluid salicylate concentrations were actually rising, Total serum salicylate concentration declines remarkably during the time period in which salicylate concentration in the cerebrospinal fluid was increasing. DISCUSSION
The concentration of salicylate within the central nervous system may well be the critical determinant of aspirin toxicity [4]. Major symptoms of salicylate toxicity, i.e., Kussmaul respiration, nausea, vomiting, and convulsions all a r i s e as salicylate concentration inc r e a s e s in the cerebrospinal fluid [6]. Heretofore, most efforts to a s s e s s clinical salicylate toxicity have focused upon total salicylate concentration measured in whole serum. These values often correlate poorly with cerebrospinal fluid concentrations because of at least two factors that affect the exchange of salicylate molecules between the vascular space and the cerebrospinal fluid. These are: (a) serum protein binding [lo], and (b) pH-dependent ionization of the salicylate molecule [ 113. Clearly, true cerebrospinal fluid salicylate levels would be the most prognostically valuable laboratory data. However, serial lumbar punctures are unfeasible in routine management of patients. It would appear that serum unbound salicylate determinations would be a reasonable approximation of the more useful but invasive CSF determinations and much more clinically pertinent than the presently employed total serum salicylate levels. There is growing evidence that unbound serum salicylate concentration should be considered a good indicator of clinical toxicity. Recent data indicate that the initial Vd of salicylate correlates well with the clinical severity of intoxication [2]. Implicit in this understanding is that a greater Vd represents a greater central nervous system involvement. There is evidence that Vd calculated from serum unbound salicylate values is valid and constant whereas that calculated from total serum values is misleading [3]. This also indicates that unbound serum salicylate concentrations should be considered as a clinical measure of salicylate toxicity. Our study in dogs has evaluated the relative concordance between cerebrospinal fluid and serum unbound salicylate concentration throughout the course of experimental salicylism. Changes in protein binding and changes in serum pH were experimentally induced to determine if
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
CNS SALICYLATE
629
cerebrospinal fluid and serum unbound salicylate concentration might deviate, causing serum unbound salicylate concentration to become misrepresentative of cerebrospinal fluid salicylate concentration. Actually, in the absence of binding changes o r pH changes, serum unbound salicylate concentration parallelled cerebrospinal fluid salicylate concentration rather closely; the absolute difference in these concentrations being explained by the impermeability of the bloodbrain b a r r i e r to ionized salicylate molecules. As salicylate was displaced from the serum albumin binding sites by sulfinpyrazone, the secondary increase in serum unbound salicylate in serum water was rapidly reflected by an immediate increase in CSF salicylate concentration. Accordingly, the constant relationship of these concentrations w a s remarkably retained, even as protein binding of salicylate was drastically altered. Most impressive, however, w a s the effect of acidosis on the otherwise parallel relationship between serum unbound salicylate concentration and CSF salicylate concentration. A s a change in serum pH (7.32 to 7.24) w a s elicited by acetazolamide administration, serum unbound salicylate concentration decreased. Presumably, the more permeable un-ionized molecules increased in concentration at the lower pH and diffused rapidly from the vascular compartment into the CSF. Figure 2 indicates that this is, indeed, the case, since CSF salicylate concentration increased with the acute change from the ionized to the un-ionized state. Thus, the increasing Vd of salicylate w a s reflected by increased salicylate levels in the CSF. It is therefore most important to note that serum unbound salicylate does not always accurately reflect the situation in the central nervous system. In the presence of acidosis, CSF salicylate may actually increase, as serum unbound concentration is falling. Even with this obvious limitation, unbound salicylate determinations would much more accurately reflect the situation within the CNS than the standard total levels as presently employed. Clearly, under present practices, the therapist i s frequently deceived by inappropriately low total serum concentration in the face of rising levels in the CNS. SUMMARY The poor correlation between clinical salicylate toxicity and serum blood levels is reapproached in light of recent evidence linking clinical severity with initial volume of distribution (Vd). It is recognized that two variables alter salicylate Vd in such manner that serum salicylate levels are misleading (thus, the change in Vd i s not detected by
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
630
REED AND PALMISANO
present methods). These variables are serum protein binding and the pH-dependent ionized/un- ionized ratio in the unbound salicylate fraction. Measurement of salicylate concentration in the cerebrospinal fluid (CSF) would circumvent these variables, but would be clinically impractical. Thus, an alternative is sought to the inexact total serum salicylate levels and the impractical CSF salicylate levels for assessment of the severity of salicylate poisoning. This study indicates that, in dogs, serum unbound salicylate levels closely reflect CSF salicylate levels, even as a decrease in serum protein binding is in progress. However, serum unbound salicylate concentration does not reflect CSF salicylate concentration as a decrease in serum pH is elicited (CSF salicylate actually increased as serum unbound salicylate decreased). On the other hand, serum unbound salicylate measurement would seem preferable to total serum salicylate measurements now used in that the total value decreased markedly as either protein binding change o r acidosis produced a change in distribution and the resultant increase in CSF salicylate. REFERENCES
[ 11 A. K. Done, Salicylate intoxication: Significance of measure121
[3] [4] [5] [6]
[7] [8]
ments of salicylate in blood in cases of acute ingestion, Pediatrics, 26, 800 (1960). G. Levy and%. J. Yaffe, Relation between dose and apparent volume of distribution of salicylate in children, Pediatrics, 54, 713 (1974). i?l; P. McCann and P. A. Palmisano, Salicylate pharmacokinetics in the dog, Res. Commun. Chem. -Pathol. Pharmacol., 5, 17 (1973). 3. B. Hill, Salicylate intoxication, New Engl. J. Med., 288, 1110 (1973). S. M. 'Te&ey and R. M. Miller, The respiratory and circulatory actions of salicylate, Am. J. Med., 19, 498 (1955). D. M. Woodbury, "Analgesic- Antipyretics, Anti-Inflammatory Agents, and Inhibitors of Uric Acid Synthesis," in The Pharmacological Basis of Therapeutics, 4th ed. (L.Goodman and A. Gilman, ed.), MacMillan, New York, 1970, p. 317. W. P. Mccann, Problems in data collection and analysis in human pharmacokinetics, Drug Information Bull., 3,52 (1969). P. Trinder, Rapid determination of salicylate in biological fluids, Biochem. J., 57, 301 (1954).
Clinical Toxicology Downloaded from informahealthcare.com by UB Frankfurt/Main on 09/05/14 For personal use only.
CNS SALICYLATE
631
[9] J. J. Burns, T. F. Yu, A. Rittenband, J. M. Perel, A. B. Gutman, and B. B. Brodie, A potent new uricosuric agent, the sulfoxide metabolite of the phenylbutazone analogue, J. Pharmacol. Exper. 418 (1957). Therap., 19, [ 101 C. Davison and R. L. Strauss, The effect of the distribution and excretion of salicylate by drugs displacing that compound from plasma binding sites, The Pharmacologist, 3,81 (1961). [ 111 J. B. Hill, Experimental salicylate poisoning: observations on the effects of altering blood pH on tissue and plasma salicylate concentrations, Pediatrics, 47, 658 (1971).
