Eur J Clin Pharmacol (1992) 42:181-185 EuropeanJournalof ~ ( ~ ( ~
© Springer-Verlag 1992
Temporary reversal of serum to cerebrospinal fluid glycerol concentration gradient after intravenous infusion of glycerol R. Nau ~, F.-J. Prins 1, H. Kolenda 2, and H. W. Prange 1 Department of Neurology 1 and Neurosurgery 2, University of G6ttingen, G6ttingen, FRG Received: February 4, 1991/Accepted in revised form: Juli 11, 1991
Summary. Glycerol 50 g infused i.v. over 2 to 6 h is widely used to treat cerebral oedema in patients with acute stroke. Its transit through the blood-cerebrospinal fluid barrier in subjects with uninflamed meninges has now been examined. In 7 patients with an external ventriculostomy for occlusive hydrocephalus, each of whom was given 500 ml of a 10% solution IV over 4 h, serum and CSF were repeatedly sampled during and after the infusion and glycerol was measured enzymatically, The highest serum glycerol level of 191-923 mg/1 was observed at the end of the infusion. The maximum CSF glycerol of 18.7-110.8 mg/1 was attained 0-1 h after the end of the infusion. Elimination both from serum and CSF approximated a single-exponential decay; the elimination half-life from serum was 0.29-0.56 h compared to 1.033.68 h from CSF. In six of the seven cases there was a temporary reversal of the serum/CSF concentration gradient during glycerol elimination. The ratios of the AUCs of CSF and serum, which describe the overall penetration of glycerol into CSI~. ranged from 0.09-0.31. In conclusion, the serum level of glycerol produced by giving 50 g IV glycerol over 4 h may not be sufficiently high reliably dehydrate to brain tissue in many patients, and the slow elimination of glycerol from the CSF may be related to the so-called rebound phenomenon.
Key words: Glycerol; brain oedema, serum, cerebrospinal fluid, pharmacokinetics
Cerebral oedema leading to an increased intracranial pressure is the most common cause of death in acute stroke. Glycerol administered either orally or intravenously has been used since the early sixties to reduce intracranial pressure [Cantore et al. 1964]. Since glycerol is metabolized by brain tissue [Sloviter et al. 1966], its introduction into clinical practice was associated with the notion that it might not produce the rebound phenomena known to occur with other osmotherapeutic substances. Whilst some authors did not report rebound, even after high doses of oral [Wald et al. 1982] or intravenous glyc-
erol [Pitlick et al. 1982], others observed severe rises in intracranial pressure both with continuous intravenous [Guisado et al. 1975] and discontinuous oral [Rottenberg et al. 1977] administration. Since its introduction by Meyer et al. (1971), the intravenous infusion of a 10 % glycerol solution of 500 ml per day has been studied extensively in clinical trials in acute stroke; 500 ml has either been infused continuously over 24 h [Gilsanz et al. 1975], or discontinually within 4 h [Bayer et al. 1987, Fawer et al. 1978, Frei et al. 1987, Krfimer et al. 1981], 1-2 h [Friedli et al. 1979] or 6 h [Frithz & Werner 1975, Larsson et al. 1976], but without measuring serum osmolality or glycerol concentrations, The majority of trials have yielded favourable results [Bayer et al. 1987, Fawer et al. 1978, Kr~imer et al. 1981, Friedli et al. 1979, Frithz & Werner 1975, Gilsanz et al. 1975], although some have not shown any significant benefit [Frei cr al. 1987, Larsson et al, 1976]. In the present study the pharmacokinetics of intravenous glycerol in serum and CSF was investigated and the findings are discussed in relation to the controversial clinical results.
Patients and methods Seven patients (5 f, 2 m; age 44-81 y, median 62 y) suffering from cerebral oedema of various causes, received 500 ml of 10 % glycerol in 2.5 % glucose/0.45% NaC1 (Glycerosteril ®,Fresenius, Bad Homburg, FRG), intravenously over 4 h, as part of their therapy. They had undergone external ventriculostomy for occlusive hydrocephalus. For further details of the patients see Table 1. Blood and cerebrospinal fluid (CSF) samples were collected before beginning the glycerol infusion for determination of the endogenous basal concentration, 2 and 4 h after starting it, and 0.5,1,2, 3, 5 and 8-12 h after the end of the infusion. The study protocol was approved by the Ethics Commission of the University Hospital; informed consent for participation was obtained from the nearest relative. Serum and CSF glycerol concentrations were determined by an enzymatic assay based on the conversion of glycerol to glycerol-3phosphate coupled to auxiliary enzyme reactions leading to NADHconsumption (kit from Boehringer Mannheim, FRG). The method
182 Table 1. Details of patients investigated in order of ascending maximal serum glycerol concentration after IV infusion of 50 g/4 h Patient Body Agey/ Disease weight sex (kg) A 95 44, m intracerebral haemorrhage B 85 60, f brain stem haemorrhage C 75 76, f subarachnoid haemorrhage D
70
62, m
E
70
48, f
F
70
81, f
G
50
70, f
subarachnoid haemorrhage subarachnoid haemorrhage subarachnoid haemorrhage intracerebral haemorrhage
CSF WBC/ RBC/ protein mm 3 mm 3 (mg/1) 916 17 9045 141
2
53
353
13
171
290
16
507
812
28
8021
450
17
4949
950
48
6912
had previously been validated with biological fluids [Gaab & Pflughaupt 1977; Wieland 1984]. At sample concentrations of 400 mg. 1-1 the accuracy of the test determined by an external standard was 99.7 %, and the interday coefficient of variation was 3.8 % (n = 29). At 4 rag- 1- ~the accuracy was 96.0 % and the coefficient of variation was 2.5 % (n = 8). The sensitivity was below 0.5 rag. 1-1. The procedure is specific for glycerol [Wieland 1984, information from the manufacturer]. The areas under the serum and CSF concentration-time curves (AUCs, AUCcsF) were estimated by the trapezoidal rule after subtraction of the basal concentrations from the values found during and after the glycerol infusion. The area to infinity beyond the last measured glycerol concentration above the basal level (CI) was estimated by G/k. Serum and CSF elimination rate constants were determined by log-linear regression analysis and half-lives (t~/2) as ln2/k. For these calculations, too, basal glycerol levels were first subtracted. The total clearance of glycerol from serum (CL) was calculated as dose/AUCs and the apparent volume of distribution (V) as dose/AUCs-k [Rowland & Tozer 1989].
Results E n d o g e n o u s basal concentrations of glycerol r a n g e d f r o m 4.2 to 11.6 rag-1 -~ in s e r u m ( m e d i a n = 6.6 mg. 1-1), and f r o m 0.8 to 2 . 3 m g . 1 - 1 in ventricular CSF (me-
dian = 1.4 m g . 1 1). T h e m a x i m u m s e r u m glycerol concentrations were o b s e r v e d at the end of the infusions and r e a c h e d 191 to 923 m g . 1-1 ( m e d i a n 436 rag. 1-1). T h e p e a k levels in CSF (18.7 to 110.8 mg .1-1, m e d i a n 53.4 mg-1-1) were m e a s u r e d 0 to i h after t e r m i n a t i o n of the glycerol infusion (Fig. 1, Table 2). T h e elimination of glycerol f r o m s e r u m and CSF app r o x i m a t e d to a single-exponential decay (Fig. 1). S e r u m elimination half-lives (tl/2S) were estimated as 0.29 to 0.56 h ( m e d i a n 0.37 h, n = 7) c o m p a r e d to CSF elimination half-lives (h/2csv) of 1.03 to 3.68 h ( m e d i a n 1.91 h; Table 2). T h e tl/2CSFwas 3- to 11-times higher than the h/2s, T h e long tmcsv resulted in a t e m p o r a r y reversal of the b l o o d - C S F gradient of glycerol concentration in six of the seven patients during the glycerol elimination phase (Fig. 1). D u e to the different shapes of the concentration-time curves, the ratio b e t w e e n corresponding glycerol CSF and s e r u m concentrations (Ccsv/Cs) was not constant. In fact, CcsF/Cs varied b e t w e e n 0.03 (Pat. E, 2 h after starting the glycerol infusion) and 6.4 (Pat. D, 3 h after the end of the glycerol infusion). T h e ratios of AUCcsv and A U C s , representing the overall p e n e t r a t i o n of glycerol into the CSF, ranged f r o m 0.09 to 0.31 (median 0.26, n =7). A U C c s d A U C s was correlated with the basal CcsdCs m e a s u r e d b e f o r e beginning of the glycerol infusion ( S p e a r m a n ' s r a n k order correlation rs=0.83, P < 5 % ) . Basal CcsF/Cs ranged f r o m 0.14 to 0.46 (median 0.17, n = 7). N e i t h e r the p e n e t r a t i o n of glycerol into CSF, expressed by A U C c s v / A U C s , nor the ratio of the basal e n d o g e n o u s glycerol concentrations, were positively correlated with p a r a m e t e r s used in clinical practice to characterize the condition of the b l o o d - C S F barrier, e,g. the C S F protein content.
Discussion In the p r e s e n t study the glycerol elimiation half-life f r o m CSF (tl/2CSF) was 1.03-3.68 h c o m p a r e d to 0.29-0.56 h in s e r u m (tvzs). T h e tl/as values agreed well with the values r e p o r t e d by others: 0.87 h [Shafrir & G o r i n 1963], 0.180.21h [Pelkonen et al. 1967], 0.5 h [Senior & L o r i d a n 1968], 0.29-0.39 h [Tibbling 1969], 0.48 h [ O l b e r m a n n et al. 1977]. D i v e r g e n c e s arose f r o m m e a s u r e m e n t s after I V boluses of different sizes, after steady state conditions, or
Table 2. Pharmacokinetics of glycerol in serum and CSF after a single infusion of 50 g/4 h Patient
A B C D E F G
Serum C~x (rag-1 -~) 191 237 295 436 478 631 923
Cerebrospinal fluid tm (h) 0.35 0.33 0.29 0.41 0.37 0.48 0.56
V (1) 45.3 32.8 21.4 22.6 18.7 18.2 13.6
CL (1.h -1) 89.5 69.1 51.5 37.7 35.0 26.2 16.9
AUCs (rag-h-1 -~) 559 723 971 1327 1428 1909 2959
time of peak CSF concentration, (h) after beginning the glycerol infusion b Penetration into CSF expressed as AUCcsv/AUCs
Cmax
tm~xa
tl/~
(mg. 1
© Springer-Verlag 1992
Temporary reversal of serum to cerebrospinal fluid glycerol concentration gradient after intravenous infusion of glycerol R. Nau ~, F.-J. Prins 1, H. Kolenda 2, and H. W. Prange 1 Department of Neurology 1 and Neurosurgery 2, University of G6ttingen, G6ttingen, FRG Received: February 4, 1991/Accepted in revised form: Juli 11, 1991
Summary. Glycerol 50 g infused i.v. over 2 to 6 h is widely used to treat cerebral oedema in patients with acute stroke. Its transit through the blood-cerebrospinal fluid barrier in subjects with uninflamed meninges has now been examined. In 7 patients with an external ventriculostomy for occlusive hydrocephalus, each of whom was given 500 ml of a 10% solution IV over 4 h, serum and CSF were repeatedly sampled during and after the infusion and glycerol was measured enzymatically, The highest serum glycerol level of 191-923 mg/1 was observed at the end of the infusion. The maximum CSF glycerol of 18.7-110.8 mg/1 was attained 0-1 h after the end of the infusion. Elimination both from serum and CSF approximated a single-exponential decay; the elimination half-life from serum was 0.29-0.56 h compared to 1.033.68 h from CSF. In six of the seven cases there was a temporary reversal of the serum/CSF concentration gradient during glycerol elimination. The ratios of the AUCs of CSF and serum, which describe the overall penetration of glycerol into CSI~. ranged from 0.09-0.31. In conclusion, the serum level of glycerol produced by giving 50 g IV glycerol over 4 h may not be sufficiently high reliably dehydrate to brain tissue in many patients, and the slow elimination of glycerol from the CSF may be related to the so-called rebound phenomenon.
Key words: Glycerol; brain oedema, serum, cerebrospinal fluid, pharmacokinetics
Cerebral oedema leading to an increased intracranial pressure is the most common cause of death in acute stroke. Glycerol administered either orally or intravenously has been used since the early sixties to reduce intracranial pressure [Cantore et al. 1964]. Since glycerol is metabolized by brain tissue [Sloviter et al. 1966], its introduction into clinical practice was associated with the notion that it might not produce the rebound phenomena known to occur with other osmotherapeutic substances. Whilst some authors did not report rebound, even after high doses of oral [Wald et al. 1982] or intravenous glyc-
erol [Pitlick et al. 1982], others observed severe rises in intracranial pressure both with continuous intravenous [Guisado et al. 1975] and discontinuous oral [Rottenberg et al. 1977] administration. Since its introduction by Meyer et al. (1971), the intravenous infusion of a 10 % glycerol solution of 500 ml per day has been studied extensively in clinical trials in acute stroke; 500 ml has either been infused continuously over 24 h [Gilsanz et al. 1975], or discontinually within 4 h [Bayer et al. 1987, Fawer et al. 1978, Frei et al. 1987, Krfimer et al. 1981], 1-2 h [Friedli et al. 1979] or 6 h [Frithz & Werner 1975, Larsson et al. 1976], but without measuring serum osmolality or glycerol concentrations, The majority of trials have yielded favourable results [Bayer et al. 1987, Fawer et al. 1978, Kr~imer et al. 1981, Friedli et al. 1979, Frithz & Werner 1975, Gilsanz et al. 1975], although some have not shown any significant benefit [Frei cr al. 1987, Larsson et al, 1976]. In the present study the pharmacokinetics of intravenous glycerol in serum and CSF was investigated and the findings are discussed in relation to the controversial clinical results.
Patients and methods Seven patients (5 f, 2 m; age 44-81 y, median 62 y) suffering from cerebral oedema of various causes, received 500 ml of 10 % glycerol in 2.5 % glucose/0.45% NaC1 (Glycerosteril ®,Fresenius, Bad Homburg, FRG), intravenously over 4 h, as part of their therapy. They had undergone external ventriculostomy for occlusive hydrocephalus. For further details of the patients see Table 1. Blood and cerebrospinal fluid (CSF) samples were collected before beginning the glycerol infusion for determination of the endogenous basal concentration, 2 and 4 h after starting it, and 0.5,1,2, 3, 5 and 8-12 h after the end of the infusion. The study protocol was approved by the Ethics Commission of the University Hospital; informed consent for participation was obtained from the nearest relative. Serum and CSF glycerol concentrations were determined by an enzymatic assay based on the conversion of glycerol to glycerol-3phosphate coupled to auxiliary enzyme reactions leading to NADHconsumption (kit from Boehringer Mannheim, FRG). The method
182 Table 1. Details of patients investigated in order of ascending maximal serum glycerol concentration after IV infusion of 50 g/4 h Patient Body Agey/ Disease weight sex (kg) A 95 44, m intracerebral haemorrhage B 85 60, f brain stem haemorrhage C 75 76, f subarachnoid haemorrhage D
70
62, m
E
70
48, f
F
70
81, f
G
50
70, f
subarachnoid haemorrhage subarachnoid haemorrhage subarachnoid haemorrhage intracerebral haemorrhage
CSF WBC/ RBC/ protein mm 3 mm 3 (mg/1) 916 17 9045 141
2
53
353
13
171
290
16
507
812
28
8021
450
17
4949
950
48
6912
had previously been validated with biological fluids [Gaab & Pflughaupt 1977; Wieland 1984]. At sample concentrations of 400 mg. 1-1 the accuracy of the test determined by an external standard was 99.7 %, and the interday coefficient of variation was 3.8 % (n = 29). At 4 rag- 1- ~the accuracy was 96.0 % and the coefficient of variation was 2.5 % (n = 8). The sensitivity was below 0.5 rag. 1-1. The procedure is specific for glycerol [Wieland 1984, information from the manufacturer]. The areas under the serum and CSF concentration-time curves (AUCs, AUCcsF) were estimated by the trapezoidal rule after subtraction of the basal concentrations from the values found during and after the glycerol infusion. The area to infinity beyond the last measured glycerol concentration above the basal level (CI) was estimated by G/k. Serum and CSF elimination rate constants were determined by log-linear regression analysis and half-lives (t~/2) as ln2/k. For these calculations, too, basal glycerol levels were first subtracted. The total clearance of glycerol from serum (CL) was calculated as dose/AUCs and the apparent volume of distribution (V) as dose/AUCs-k [Rowland & Tozer 1989].
Results E n d o g e n o u s basal concentrations of glycerol r a n g e d f r o m 4.2 to 11.6 rag-1 -~ in s e r u m ( m e d i a n = 6.6 mg. 1-1), and f r o m 0.8 to 2 . 3 m g . 1 - 1 in ventricular CSF (me-
dian = 1.4 m g . 1 1). T h e m a x i m u m s e r u m glycerol concentrations were o b s e r v e d at the end of the infusions and r e a c h e d 191 to 923 m g . 1-1 ( m e d i a n 436 rag. 1-1). T h e p e a k levels in CSF (18.7 to 110.8 mg .1-1, m e d i a n 53.4 mg-1-1) were m e a s u r e d 0 to i h after t e r m i n a t i o n of the glycerol infusion (Fig. 1, Table 2). T h e elimination of glycerol f r o m s e r u m and CSF app r o x i m a t e d to a single-exponential decay (Fig. 1). S e r u m elimination half-lives (tl/2S) were estimated as 0.29 to 0.56 h ( m e d i a n 0.37 h, n = 7) c o m p a r e d to CSF elimination half-lives (h/2csv) of 1.03 to 3.68 h ( m e d i a n 1.91 h; Table 2). T h e tl/2CSFwas 3- to 11-times higher than the h/2s, T h e long tmcsv resulted in a t e m p o r a r y reversal of the b l o o d - C S F gradient of glycerol concentration in six of the seven patients during the glycerol elimination phase (Fig. 1). D u e to the different shapes of the concentration-time curves, the ratio b e t w e e n corresponding glycerol CSF and s e r u m concentrations (Ccsv/Cs) was not constant. In fact, CcsF/Cs varied b e t w e e n 0.03 (Pat. E, 2 h after starting the glycerol infusion) and 6.4 (Pat. D, 3 h after the end of the glycerol infusion). T h e ratios of AUCcsv and A U C s , representing the overall p e n e t r a t i o n of glycerol into the CSF, ranged f r o m 0.09 to 0.31 (median 0.26, n =7). A U C c s d A U C s was correlated with the basal CcsdCs m e a s u r e d b e f o r e beginning of the glycerol infusion ( S p e a r m a n ' s r a n k order correlation rs=0.83, P < 5 % ) . Basal CcsF/Cs ranged f r o m 0.14 to 0.46 (median 0.17, n = 7). N e i t h e r the p e n e t r a t i o n of glycerol into CSF, expressed by A U C c s v / A U C s , nor the ratio of the basal e n d o g e n o u s glycerol concentrations, were positively correlated with p a r a m e t e r s used in clinical practice to characterize the condition of the b l o o d - C S F barrier, e,g. the C S F protein content.
Discussion In the p r e s e n t study the glycerol elimiation half-life f r o m CSF (tl/2CSF) was 1.03-3.68 h c o m p a r e d to 0.29-0.56 h in s e r u m (tvzs). T h e tl/as values agreed well with the values r e p o r t e d by others: 0.87 h [Shafrir & G o r i n 1963], 0.180.21h [Pelkonen et al. 1967], 0.5 h [Senior & L o r i d a n 1968], 0.29-0.39 h [Tibbling 1969], 0.48 h [ O l b e r m a n n et al. 1977]. D i v e r g e n c e s arose f r o m m e a s u r e m e n t s after I V boluses of different sizes, after steady state conditions, or
Table 2. Pharmacokinetics of glycerol in serum and CSF after a single infusion of 50 g/4 h Patient
A B C D E F G
Serum C~x (rag-1 -~) 191 237 295 436 478 631 923
Cerebrospinal fluid tm (h) 0.35 0.33 0.29 0.41 0.37 0.48 0.56
V (1) 45.3 32.8 21.4 22.6 18.7 18.2 13.6
CL (1.h -1) 89.5 69.1 51.5 37.7 35.0 26.2 16.9
AUCs (rag-h-1 -~) 559 723 971 1327 1428 1909 2959
time of peak CSF concentration, (h) after beginning the glycerol infusion b Penetration into CSF expressed as AUCcsv/AUCs
Cmax
tm~xa
tl/~
(mg. 1
