journal u1 Internal Medicine 1990 : 22 7 : 5-9
Creatine kinase isoenzyme BB in the cerebrospinal fluid of patients with acute neurological diseases A. BODVARSSON, L. FRANZSON & H. BRIEM From the Departments of Medicine and Clinical Chemistry. Reykjavi’k City Hospital, lcelund
Abstract. Bodvarsson A, Franzson L, Briem H (Departments of Medicine and Clinical Chemistry, Reykjavik City Hospital, Iceland). Creatine kinase isoenzyme BB in the cerebrospinal fluid of patients with acute neurological diseases. journal ufhternal Medicine 1990; 227: 5-9.
In order to predict the outcome of patients with acute neurological symptoms at discharge, the concentration of creatine kinase isoenzyme BB (CK-BB) was determined by radioimmunoassay in the cerebrospinal fluid (CSF) of 115 consecutive patients. On admission and over the next 3 days the concentration of CK-BB was significantly increased in patients with brain death and other neurological sequelae compared with those with favourable outcome. There was a variation in time in concentrations of CKBB between diseases causing neurological sequelae. Thus, cerebrovascular haemorrhages caused highest concentrations on admission, but cerebral ischaemia due to cardiac arrest caused highest concentrations 3 days after admission. On admission the CK-BB measurements were highly specific with high predictive value of positive result when distinguishing patients with brain death and other neurological sequelae from those without complications at discharge. However, when distinguishing patients with brain death from those with other neurological sequelae, the test was most specific and had highest predictive value of a positive result 3 days after admission. Keywords: cerebrospinal fluid, creatine kinase isoenzyme BB. neurological sequelae.
Introduction Brain tissue contains creatine kinase isoenzyme BB (CK-BB) but lacks CK-MM and CK-MB [l-31. CK-BB has been shown to be present in both neurones and astrocytes [4,5]. Indeed, the concentration of CK-BB in the cerebrospinal fluid (CSF) is correlated to the quantity of neuronal necrosis found at autopsy in patients who subsequently die after cardiac arrest [61. The outcome of patients with acute symptoms of neurological disease is often difficult to predict [731. Recently, several reports have been published, indicating that the determination of CK-BB in CSF may be of value for predicting the outcome of patients with head injuries [9,l o ] , cerebral ischaemia after cardiac arrest [11+12],acute stroke [13], viral and bacterial infections of the central nervous system [14] and cerebral insults in newborn infants (4). The purpose of this study was to determine the value of CK-BB measurements in the CSF for
predicting the outcome of patients with acute neurological diseases in a general hospital.
Study population One hundred and fifteen consecutive patients (mean age 50 years; range 14-96 years) were investigated during a 14-month period at the Reykjavik City Hospital. The patients were enrolled in the study if a lumbar puncture was performed for diagnostic purposes of acute neurological symptoms. The lumbar puncture was repeated every 2 4 h when possible. For control purposes CSF was obtained at the induction of spinal anaesthesia in 1 5 patients without neurological symptoms but admitted for orthopaedic surgery. Every patient with a neurological symptom on admission was further examined by a neurologist. Cerebral infarcts and cerebrovascular haemorrhages were diagnosed by computerized tomography of the brain.
5
6
A. B O D V A R S S O N et al.
Methods A total of 2 1 8 CSF samples were obtained by lumbar puncture with the patient in the lateral recumbent position. The samples were stored at -20 "C and were analysed within a week from sampling with radioimmunoassay (CARDIO-Check,CK-B, RIA KIT, Nuclear Medical Systems Inc, California). In CSF the coefficient of variation for CK-BB measurements was 8% at 5 pg 1-', 3.5% at 10 pg 1-' and 3 % at 100 pg 1-'. The storage, freezing and thawing of the CSF samples did not significantly alter the CK-BB values. Neurological complications were assessed at discharge from the hospital and were divided into three stages: I = no complications, I1 = clinical evidence of neurological sequelae and I11 = brain death. The Mann-Whitney U-test was used for statistical analysis. A P-value < 0.05 was considered significant. Discriminating limits of the CK-BB measurements dividing the results into positive or negative were chosen empirically to maximize the clinical usefulness of the test [15]. Based on the discriminating limits, sensitivity and specificity of the test and predictive values of negative and positive results were calculated [16].
Results As shown in Fig. 1 the concentration of CK-BB in CSF was significantly increased on admission, 1 and 2 days (P < 0.001) and 3 days after admission (P < 0.002) in patients with neurological complications
(stages I1 and 111) as compared with those without complications (stage I). On admission the mean ( fSEM) CSF concentration of CK-BB in stage I was 0.64 pg 1-'f0.33 pg I-', in stage I1 45.4 pg 1-'f28.2 pgl-' and in stage I11 132.6pg1-'+99.4 pg 1-'. The mean CSF concentration increased until the third day after admission in stage 111 when it reached a maximum of 260.3 pg 1-'& 54.8 pg 1-'. Thereafter it declined and the mean concentration after 7 days was 6.7 pg 1-' f4.2 pg 1-'. On admission Table 1. Classification of 115 consecutive patients with neurological symptoms according to diagnosis and outcome No. of cases
Stage I*
44
44
Cerebral ischaemia after cardiac arrest
9
3
1
5
Cerebral infarct
18
2
15
1
7
3
2
2
Infections of CNS Bacterial Aseptic
11 8
7 8
1
3
Cerebral tumours
4
3
11
6
Diagnosis No organic disease of CNS found
Cerebrovascular haemorrhages Cerebral Subarachnoidal
Various
3
Stage I1
Stage 111
3
1
3
2
*Stage I = no neurological sequelae at discharge; stage I1 = neurological sequelae without brain death at discharge; stage 111 = brain death.
T
-7 2002
s
1
m
41 Y
u
0
I
2
3
4 5 DOY
6
7
too-
>.'7
Fig. 1. Cerebrospinal fluid creatine kinase isoenzyme BB (CK-BB) concentration (mean kSEM) at different times after admission of patients with acute neurological symptoms. Patients without Patients with neurological complication at discharge (0). sequelae at discharge (+). Patients with brain death at discharge (m).
0-
Do Y
Fig. 2. Cerebrospinal fluid creatine kinase isoenzyme BB (CK-BB) concentration (mean & SEM) of patients with intracranial cardiac arrest (n = 6) (+), and haemorrhage (n = 7) cerebral infarct (n = 16) (m) at different times after admission.
(o),
CREATINE KINASE ISOENZYMI! BB I N CSF
Day 0
Day 2
Day I
Day
1000
the difference in the CSF concentrations of CK-BB between stages I1 and I11 was not significant. However, the CK-BB concentrations were significantly higher in stage I11 than in stage 11, 1 day ( P < 0.02), 2 days ( P < 0.05) and 3 days ( P < 0.05) after admission. CK-BB could not be detected in any of the 1 5 control patients. In Table 1 the patients with acute neurological symptoms are classified according to diagnoses and outcome at discharge from the hospital. The group of patients where no organic disease of the central nervous system (CNS) could be found consisted of patients with psychological diseases, migraine, epilepsy, pneumonia etc. The variation in time in concentration of CK-BB in the CSF of the six patients with brain ischaemia after cardiac arrest, the 1 6 patients with cerebral infarcts, and the seven patients with cerebrovascular haemorrhages, who all developed complications, is shown in Fig. 2 . One of the patients with bacterial meningitis who was comatose on admission was investigated every 24 h from admission until she died 5 days later with symptoms of brain death. The concentration of CKBB was 32 pg I-' on admission, but above 400 pg I-' at every measurement until death.
3
500
200 h
72
100
-
50
0
LL
v,
"
._
40Y "
20
10 5
.I
2 I
m LA, LA,
-
LA,
I
rrm
nm Stage
Fig. 3. Distribution of creatine kinase isoenzyme BB (CK-BB) concentration in cerebrospinal fluid (CSF) of patients with acute neurological symptoms on admission (day 0) and the next 3 days after admission (days 1-3). The outcome of patients at discharge: stage I = no complication, stage I1 = neurological sequelae. and stage 111 = brain death. The discrimination limit between stage I on the one hand and stages 11 and Ill on the other is shown by (--) and between stages I1 and 111 by (. . . ..).
Table 2. The sensitivity. specificity and predictive values of CK-BB determinations in the CSF of patients with acute neurological symptoms for the discrimination between those with neurological sequelae and those without Result* Time of investigation Day Day Day Day
0 1 2 3
(DLt = 5pgI-') ( D L = IOpgI-') ( D L = IOpgl-') ( D L = lOpgI-')
'TP = true positive result, FP
= false
Predictive valuet Sensitivity
Specificity
PVneg
PVpos
TP
FP
TN
FN
(%)
(%I
(%)
(%I
9 11 12 8
0 1 1 1
45 19 14 5
10
47 65 75 100
100 95 93 83
82 76 78 100
100
positive result, TN
6
4 0
= true
92 92 89
negative result, FN = false negative result.
t PVneg = predictive value of a negative result, PVpos = predictive value of a positive result. $ DL = discrimination limit.
Table 3. The sensitivity. specificity and predictive values of CK-BB determinations in the CSF of patients with acute neurological symptoms for the discrimination between those with brain death and those with other neurological sequelae Result*
Predictive value1
Time of investigation
PVneg
PVpos
FP
TN
FN
Sensitivity (%)
Specificity
TP
(%I
(%I
(%I
Day Day Day Day
3 6 8 6
2 3 1 0
53 24 18 8
3 1 2 0
50 86 80 100
96 89 95 100
95 96 90 100
60 67 89 100
0 (DLt = ZOpgI-') 1 ( D L = 60pgI-') 2 ( D L = 60pgI-') 3 (DL = 6 0 p g I-')
'TP = true positive result, FP = false positive result, TN
=
true negative result, FN = false negative result.
t PVneg = predictive value of a negative result, PVpos = predictive value of a positive result. $ DL = discrimination limit.
7
8
A. BODVARSSON et al.
None of the patients with cerebral tumour had any significant increase of the concentration of CK-BB. In Fig. 3 the limits of the CK-BB concentrations distinguishing stage I from stages I1 and I11 on the one hand and stage I11 from stages I and I1 on the other are shown. The sensitivity, specificity and predictive values are given in Tables 2 and 3 for the test at different discrimination limits and different times from admission.
Discussion This study confirms previous reports of a significant increase of CK-BB concentration in the CSF of patients with neurological disease leading to neurological sequelae [4, 9-14, 171. The overall results show that the concentration of CK-BB in the CSF correlates with the severity of the neurological damage. Varying concentrations of CK-BB were found in the CSF of patients with different underlying diseases at a given time. Thus, in patients with neurological sequelae, cerebrovascular haemorrhages caused the highest level of CK-BB on admission, but cerebral ischaemia due to cardiac arrest caused highest levels 2 days after admission. Patients with cerebral infarcts had, in general, lower CK-BB levels and reached the highest levels 2 days after admission, but most of these patients could not be followed for more than 48 h since they were treated with anticoagulants. These temporal differences in the concentration of CK-BB in the CSF are in accordance with previous findings [11,13]. Recently the value of determination of CK-BB levels in the CSF for predicting the clinical outcome in patients following cardiac resuscitation has been questioned [18]. However, in that study precise information on the CSF sampling time after the cardiac resuscitation, which may influence the predictive value of the test, was not stated. Indeed, initial brain damage may be assessed by extrapolating from an observed CK-BB value within 24 h after severe head trauma when taking into account the exponential decay of the CK-BB activity in the CSF after the injury [lo]. The variable increase in CK-BB concentration in time between different aetiological groups of patients is not completely understood. A delayed vascular factor has been suggested as an explanation for the protracted enzyme release after global ischaemia associated with cardiac arrest and the fact that small, focal lesions in functionally important regions of the
brain may cause focal motor defects with minimal enzyme release [6]. This may explain the low concentrations of CK-BB in the CSF found in some patients in stage 11. However, we found high concentrations of CK-BB in the CSF of two patients in stage 11. The high concentrations of CK-BB in the CSF of patients with subarachnoidal and cerebral haemorrhages on admission is unlikely to be caused by blood contamination since CK-BB is not found in white blood cells [19] and only in small amounts in red blood cells [20] and serum of healthy individuals [17, 211. Therefore, it may indicate that structural damage to the CNS occurs early in the disease process of cerebrovascular accidents. The sustained high concentrations of CK-BB in the CSF of a patient with bacterial meningitis might indicate an ongoing disease process leading to continuing structural damage of the CNS. The four patients with cerebral tumour showed no significant increase of the concentration of CK-BB in the CSF. Three of them did not develop any neurological sequelae. It is, however, unclear why the faurth patient who suffered from oligodendroglioma and died because of brain oedema did not have any significant increase of the concentration of CK-BB. The purpose of this study was to investigate the value of CK-BB measurements in the CSF for predicting the outcome of patients with acute neurological symptoms in a general hospital. Therefore, discriminating limits were chosen to obtain high specificity with reasonable sensitivity of the test. Generally, on admission the CK-BB measurements were highly specific with a high predictive value of a positive result when distinguishing patients with brain death and other neurological sequelae (stages I11 and 11) from those without complications at discharge (stage I). However, when distinguishing patients with brain death (stage 111) from those with other neurological sequelae (stage 11) the test was most specific with the highest predictive value of a positive result 3 days after admission. Therefore as a general rule we recommend at least two samples of CSF, that is, on admission and 3 days after admission. In conclusion, the determination of CK-BB in the CSF is highly specific for brain damage and seems to be a useful test for predicting the outcome of patients with acute neurological symptoms in a general hospital. However, the time factor after admission and the nature of the underlying disease must be taken into consideration when the test is applied.
C R E A T I N E K I N A S E I S O E N Z Y M E BB I N C S F
References 1 Tsung SH. Creatine kinase isoenzyme patterns in human tissue obtained at surgery. Clin Chem 1976: 2 2 : 173-5. 2 Chandler WL. Clayson KJ. Longstreth WT. Fine JS. Creatine kinase isoenzymes in human cerebrospinal fluid and brain. Clin Chem 1 9 8 4 : 30: 1804-6. 3 Petronia RRL, Maas AHJ, van Veelen CWM. Staal GEJ, Isoenzymes of creatine kinase in extracts of various parts and regions of the human central nervous system. Clin Chem 1980: 6 : 760-2. 4 Worley G. Lipman B. Gewolb IH et al. Creatine kinase brain isoenzyme : relationship of cerebrospinal fluid concentration to the neurologic condition of newborns and cellular localization in the human brain. Pediatrics 1985 : 76 : 1 5-2 1, 5 Pfeiffer FE. Homburger HA. Yanagihara T. Creatine kinase BB isoenzyme in CSF neurological diseases : measurement by radioimmunoassay. Arch Neurol 1983: 40: 169-72. 6 Vaagenes P. Kjekshus J. Torvik A. The relationship between cerebrospinal fluid creatine kinase and morphologic changes in the brain after transient cardiac arrest. Circulation 1980: 61 : 1194-9. 7 Sadove MS. Yon MK, Hollinger PH. Johnston KS, Phillips FL. Severe prolonged hypoxic episode with complete recovery. / Am Med Assoc 1961 : 175: 1102-4. 8 Bates D,Caronna JJ, Cartlidge NEF et a/. A prospective study of nontraumatic coma: methods and results in 310 patients. Ann Neurol 1977: 2 : 211-20. 9 Hans P. Born JD. Chapelle JP, Milbouw G. Creatine kinase isoenzymes in severe head injury. / Neurosurg 1983: 58: 689-92. 10 Hans P. Born JD, Albert A. ‘Extrapolated’creatine kinases-BB isoenzyme activity in assessment of initial brain damage after severe head injury. J Neurosurg 1 9 8 7 : 6 6 : 714-7. 1 1 Kjekshus JK, Vaagenes P. Hetland 0. Assessment of cerebral injury with spinal fluid creatine kinase (CSF-CK) in patients after cardiac resuscitation. Scan / Lab Invest 1980: 4 0 : 437-44. 12 Longstreth WT. Clayson KJ. Sumi SM. Cerebrospinal fluid and serum creatine kinase BB activity after out-of-hospital cardiac arrest. Neurology 1 9 8 1 : 31: 455-8.
9
13 Vaagenes P. Urdal P. Melvoll R. Valnes K. Enzyme level changes in the cerebrospinal fluid of patients with acute stroke. Arch Neurol 1 9 8 6 : 43: 357-62. 1 4 Briem H, Lindquist L. Lundbergh P. Sego E, Skoldenberg B. Creatine kinase isoenzyme BB in cerebrospinal fluid from patients with meningitis and encephalitis. / Inject Dis 1 9 8 3 : 148:180. 15 Hofvendahl S. Gerhardt W, Ljungdal L. Olsen JS. Are ‘predictive values ’ from different materials comparable ? Acta Med Scand 1 9 7 8 : 623(suppl): 1 17-20. 16 Galen RS. Gambino SR. Beyond Normality : the Predictive Value and Eficiency of Medical Diagnoses. New York: John Wiley, 1 9 7 5 : 10-14. 17 Bell RD. Rosenberg RN. Ting R, Mukherjee A. Stone MJ. Willerson JT.Creatine kinase BB isoenzyme levels by radioimmunoassay in patients with neurological disease. Ann Neurol 1978: 3 : 52-9. 18 Clemmensen P. Strandgaard S. Rasmussen S. Grande P. Cerebrospinal fluid creatine kinase isoenzyme BB levels do not predict the clinical outcome in patients unconscious following cardiac resuscitation. Clin Cardiol 1987; 1 0 : 235-6. 19 Cho HW. Prilipko L. Meltzer HY et a/. Isoenzymes of creatine phosfokinase in white blood cells. Experientia 1977: 33: 166-7. 2 0 Thompson RJ. Graham JG, McQeen INF et a/. Radioimmunoassay of brain type creatine kinase isoenzyme in human tissues and in serum of patients with neurological disorders. Neurol Sci 1 9 8 0 : 47:241-51. 21 Phillips JP, Jones HM, Hitchcock R. A d a m N. Thompson RJ. Radioimmunoassay of serum creatine kinase BB as index of brain damage after head injury. Br Med J 1980; 281 : 777-9. Received 2 7 December 1988. accepted 9 June 1989.
Correspondence : Haraldur Briem. Medical Department, Reykjvik City Hospital, 108 Reykjvik. Iceland.
Creatine kinase isoenzyme BB in the cerebrospinal fluid of patients with acute neurological diseases A. BODVARSSON, L. FRANZSON & H. BRIEM From the Departments of Medicine and Clinical Chemistry. Reykjavi’k City Hospital, lcelund
Abstract. Bodvarsson A, Franzson L, Briem H (Departments of Medicine and Clinical Chemistry, Reykjavik City Hospital, Iceland). Creatine kinase isoenzyme BB in the cerebrospinal fluid of patients with acute neurological diseases. journal ufhternal Medicine 1990; 227: 5-9.
In order to predict the outcome of patients with acute neurological symptoms at discharge, the concentration of creatine kinase isoenzyme BB (CK-BB) was determined by radioimmunoassay in the cerebrospinal fluid (CSF) of 115 consecutive patients. On admission and over the next 3 days the concentration of CK-BB was significantly increased in patients with brain death and other neurological sequelae compared with those with favourable outcome. There was a variation in time in concentrations of CKBB between diseases causing neurological sequelae. Thus, cerebrovascular haemorrhages caused highest concentrations on admission, but cerebral ischaemia due to cardiac arrest caused highest concentrations 3 days after admission. On admission the CK-BB measurements were highly specific with high predictive value of positive result when distinguishing patients with brain death and other neurological sequelae from those without complications at discharge. However, when distinguishing patients with brain death from those with other neurological sequelae, the test was most specific and had highest predictive value of a positive result 3 days after admission. Keywords: cerebrospinal fluid, creatine kinase isoenzyme BB. neurological sequelae.
Introduction Brain tissue contains creatine kinase isoenzyme BB (CK-BB) but lacks CK-MM and CK-MB [l-31. CK-BB has been shown to be present in both neurones and astrocytes [4,5]. Indeed, the concentration of CK-BB in the cerebrospinal fluid (CSF) is correlated to the quantity of neuronal necrosis found at autopsy in patients who subsequently die after cardiac arrest [61. The outcome of patients with acute symptoms of neurological disease is often difficult to predict [731. Recently, several reports have been published, indicating that the determination of CK-BB in CSF may be of value for predicting the outcome of patients with head injuries [9,l o ] , cerebral ischaemia after cardiac arrest [11+12],acute stroke [13], viral and bacterial infections of the central nervous system [14] and cerebral insults in newborn infants (4). The purpose of this study was to determine the value of CK-BB measurements in the CSF for
predicting the outcome of patients with acute neurological diseases in a general hospital.
Study population One hundred and fifteen consecutive patients (mean age 50 years; range 14-96 years) were investigated during a 14-month period at the Reykjavik City Hospital. The patients were enrolled in the study if a lumbar puncture was performed for diagnostic purposes of acute neurological symptoms. The lumbar puncture was repeated every 2 4 h when possible. For control purposes CSF was obtained at the induction of spinal anaesthesia in 1 5 patients without neurological symptoms but admitted for orthopaedic surgery. Every patient with a neurological symptom on admission was further examined by a neurologist. Cerebral infarcts and cerebrovascular haemorrhages were diagnosed by computerized tomography of the brain.
5
6
A. B O D V A R S S O N et al.
Methods A total of 2 1 8 CSF samples were obtained by lumbar puncture with the patient in the lateral recumbent position. The samples were stored at -20 "C and were analysed within a week from sampling with radioimmunoassay (CARDIO-Check,CK-B, RIA KIT, Nuclear Medical Systems Inc, California). In CSF the coefficient of variation for CK-BB measurements was 8% at 5 pg 1-', 3.5% at 10 pg 1-' and 3 % at 100 pg 1-'. The storage, freezing and thawing of the CSF samples did not significantly alter the CK-BB values. Neurological complications were assessed at discharge from the hospital and were divided into three stages: I = no complications, I1 = clinical evidence of neurological sequelae and I11 = brain death. The Mann-Whitney U-test was used for statistical analysis. A P-value < 0.05 was considered significant. Discriminating limits of the CK-BB measurements dividing the results into positive or negative were chosen empirically to maximize the clinical usefulness of the test [15]. Based on the discriminating limits, sensitivity and specificity of the test and predictive values of negative and positive results were calculated [16].
Results As shown in Fig. 1 the concentration of CK-BB in CSF was significantly increased on admission, 1 and 2 days (P < 0.001) and 3 days after admission (P < 0.002) in patients with neurological complications
(stages I1 and 111) as compared with those without complications (stage I). On admission the mean ( fSEM) CSF concentration of CK-BB in stage I was 0.64 pg 1-'f0.33 pg I-', in stage I1 45.4 pg 1-'f28.2 pgl-' and in stage I11 132.6pg1-'+99.4 pg 1-'. The mean CSF concentration increased until the third day after admission in stage 111 when it reached a maximum of 260.3 pg 1-'& 54.8 pg 1-'. Thereafter it declined and the mean concentration after 7 days was 6.7 pg 1-' f4.2 pg 1-'. On admission Table 1. Classification of 115 consecutive patients with neurological symptoms according to diagnosis and outcome No. of cases
Stage I*
44
44
Cerebral ischaemia after cardiac arrest
9
3
1
5
Cerebral infarct
18
2
15
1
7
3
2
2
Infections of CNS Bacterial Aseptic
11 8
7 8
1
3
Cerebral tumours
4
3
11
6
Diagnosis No organic disease of CNS found
Cerebrovascular haemorrhages Cerebral Subarachnoidal
Various
3
Stage I1
Stage 111
3
1
3
2
*Stage I = no neurological sequelae at discharge; stage I1 = neurological sequelae without brain death at discharge; stage 111 = brain death.
T
-7 2002
s
1
m
41 Y
u
0
I
2
3
4 5 DOY
6
7
too-
>.'7
Fig. 1. Cerebrospinal fluid creatine kinase isoenzyme BB (CK-BB) concentration (mean kSEM) at different times after admission of patients with acute neurological symptoms. Patients without Patients with neurological complication at discharge (0). sequelae at discharge (+). Patients with brain death at discharge (m).
0-
Do Y
Fig. 2. Cerebrospinal fluid creatine kinase isoenzyme BB (CK-BB) concentration (mean & SEM) of patients with intracranial cardiac arrest (n = 6) (+), and haemorrhage (n = 7) cerebral infarct (n = 16) (m) at different times after admission.
(o),
CREATINE KINASE ISOENZYMI! BB I N CSF
Day 0
Day 2
Day I
Day
1000
the difference in the CSF concentrations of CK-BB between stages I1 and I11 was not significant. However, the CK-BB concentrations were significantly higher in stage I11 than in stage 11, 1 day ( P < 0.02), 2 days ( P < 0.05) and 3 days ( P < 0.05) after admission. CK-BB could not be detected in any of the 1 5 control patients. In Table 1 the patients with acute neurological symptoms are classified according to diagnoses and outcome at discharge from the hospital. The group of patients where no organic disease of the central nervous system (CNS) could be found consisted of patients with psychological diseases, migraine, epilepsy, pneumonia etc. The variation in time in concentration of CK-BB in the CSF of the six patients with brain ischaemia after cardiac arrest, the 1 6 patients with cerebral infarcts, and the seven patients with cerebrovascular haemorrhages, who all developed complications, is shown in Fig. 2 . One of the patients with bacterial meningitis who was comatose on admission was investigated every 24 h from admission until she died 5 days later with symptoms of brain death. The concentration of CKBB was 32 pg I-' on admission, but above 400 pg I-' at every measurement until death.
3
500
200 h
72
100
-
50
0
LL
v,
"
._
40Y "
20
10 5
.I
2 I
m LA, LA,
-
LA,
I
rrm
nm Stage
Fig. 3. Distribution of creatine kinase isoenzyme BB (CK-BB) concentration in cerebrospinal fluid (CSF) of patients with acute neurological symptoms on admission (day 0) and the next 3 days after admission (days 1-3). The outcome of patients at discharge: stage I = no complication, stage I1 = neurological sequelae. and stage 111 = brain death. The discrimination limit between stage I on the one hand and stages 11 and Ill on the other is shown by (--) and between stages I1 and 111 by (. . . ..).
Table 2. The sensitivity. specificity and predictive values of CK-BB determinations in the CSF of patients with acute neurological symptoms for the discrimination between those with neurological sequelae and those without Result* Time of investigation Day Day Day Day
0 1 2 3
(DLt = 5pgI-') ( D L = IOpgI-') ( D L = IOpgl-') ( D L = lOpgI-')
'TP = true positive result, FP
= false
Predictive valuet Sensitivity
Specificity
PVneg
PVpos
TP
FP
TN
FN
(%)
(%I
(%)
(%I
9 11 12 8
0 1 1 1
45 19 14 5
10
47 65 75 100
100 95 93 83
82 76 78 100
100
positive result, TN
6
4 0
= true
92 92 89
negative result, FN = false negative result.
t PVneg = predictive value of a negative result, PVpos = predictive value of a positive result. $ DL = discrimination limit.
Table 3. The sensitivity. specificity and predictive values of CK-BB determinations in the CSF of patients with acute neurological symptoms for the discrimination between those with brain death and those with other neurological sequelae Result*
Predictive value1
Time of investigation
PVneg
PVpos
FP
TN
FN
Sensitivity (%)
Specificity
TP
(%I
(%I
(%I
Day Day Day Day
3 6 8 6
2 3 1 0
53 24 18 8
3 1 2 0
50 86 80 100
96 89 95 100
95 96 90 100
60 67 89 100
0 (DLt = ZOpgI-') 1 ( D L = 60pgI-') 2 ( D L = 60pgI-') 3 (DL = 6 0 p g I-')
'TP = true positive result, FP = false positive result, TN
=
true negative result, FN = false negative result.
t PVneg = predictive value of a negative result, PVpos = predictive value of a positive result. $ DL = discrimination limit.
7
8
A. BODVARSSON et al.
None of the patients with cerebral tumour had any significant increase of the concentration of CK-BB. In Fig. 3 the limits of the CK-BB concentrations distinguishing stage I from stages I1 and I11 on the one hand and stage I11 from stages I and I1 on the other are shown. The sensitivity, specificity and predictive values are given in Tables 2 and 3 for the test at different discrimination limits and different times from admission.
Discussion This study confirms previous reports of a significant increase of CK-BB concentration in the CSF of patients with neurological disease leading to neurological sequelae [4, 9-14, 171. The overall results show that the concentration of CK-BB in the CSF correlates with the severity of the neurological damage. Varying concentrations of CK-BB were found in the CSF of patients with different underlying diseases at a given time. Thus, in patients with neurological sequelae, cerebrovascular haemorrhages caused the highest level of CK-BB on admission, but cerebral ischaemia due to cardiac arrest caused highest levels 2 days after admission. Patients with cerebral infarcts had, in general, lower CK-BB levels and reached the highest levels 2 days after admission, but most of these patients could not be followed for more than 48 h since they were treated with anticoagulants. These temporal differences in the concentration of CK-BB in the CSF are in accordance with previous findings [11,13]. Recently the value of determination of CK-BB levels in the CSF for predicting the clinical outcome in patients following cardiac resuscitation has been questioned [18]. However, in that study precise information on the CSF sampling time after the cardiac resuscitation, which may influence the predictive value of the test, was not stated. Indeed, initial brain damage may be assessed by extrapolating from an observed CK-BB value within 24 h after severe head trauma when taking into account the exponential decay of the CK-BB activity in the CSF after the injury [lo]. The variable increase in CK-BB concentration in time between different aetiological groups of patients is not completely understood. A delayed vascular factor has been suggested as an explanation for the protracted enzyme release after global ischaemia associated with cardiac arrest and the fact that small, focal lesions in functionally important regions of the
brain may cause focal motor defects with minimal enzyme release [6]. This may explain the low concentrations of CK-BB in the CSF found in some patients in stage 11. However, we found high concentrations of CK-BB in the CSF of two patients in stage 11. The high concentrations of CK-BB in the CSF of patients with subarachnoidal and cerebral haemorrhages on admission is unlikely to be caused by blood contamination since CK-BB is not found in white blood cells [19] and only in small amounts in red blood cells [20] and serum of healthy individuals [17, 211. Therefore, it may indicate that structural damage to the CNS occurs early in the disease process of cerebrovascular accidents. The sustained high concentrations of CK-BB in the CSF of a patient with bacterial meningitis might indicate an ongoing disease process leading to continuing structural damage of the CNS. The four patients with cerebral tumour showed no significant increase of the concentration of CK-BB in the CSF. Three of them did not develop any neurological sequelae. It is, however, unclear why the faurth patient who suffered from oligodendroglioma and died because of brain oedema did not have any significant increase of the concentration of CK-BB. The purpose of this study was to investigate the value of CK-BB measurements in the CSF for predicting the outcome of patients with acute neurological symptoms in a general hospital. Therefore, discriminating limits were chosen to obtain high specificity with reasonable sensitivity of the test. Generally, on admission the CK-BB measurements were highly specific with a high predictive value of a positive result when distinguishing patients with brain death and other neurological sequelae (stages I11 and 11) from those without complications at discharge (stage I). However, when distinguishing patients with brain death (stage 111) from those with other neurological sequelae (stage 11) the test was most specific with the highest predictive value of a positive result 3 days after admission. Therefore as a general rule we recommend at least two samples of CSF, that is, on admission and 3 days after admission. In conclusion, the determination of CK-BB in the CSF is highly specific for brain damage and seems to be a useful test for predicting the outcome of patients with acute neurological symptoms in a general hospital. However, the time factor after admission and the nature of the underlying disease must be taken into consideration when the test is applied.
C R E A T I N E K I N A S E I S O E N Z Y M E BB I N C S F
References 1 Tsung SH. Creatine kinase isoenzyme patterns in human tissue obtained at surgery. Clin Chem 1976: 2 2 : 173-5. 2 Chandler WL. Clayson KJ. Longstreth WT. Fine JS. Creatine kinase isoenzymes in human cerebrospinal fluid and brain. Clin Chem 1 9 8 4 : 30: 1804-6. 3 Petronia RRL, Maas AHJ, van Veelen CWM. Staal GEJ, Isoenzymes of creatine kinase in extracts of various parts and regions of the human central nervous system. Clin Chem 1980: 6 : 760-2. 4 Worley G. Lipman B. Gewolb IH et al. Creatine kinase brain isoenzyme : relationship of cerebrospinal fluid concentration to the neurologic condition of newborns and cellular localization in the human brain. Pediatrics 1985 : 76 : 1 5-2 1, 5 Pfeiffer FE. Homburger HA. Yanagihara T. Creatine kinase BB isoenzyme in CSF neurological diseases : measurement by radioimmunoassay. Arch Neurol 1983: 40: 169-72. 6 Vaagenes P. Kjekshus J. Torvik A. The relationship between cerebrospinal fluid creatine kinase and morphologic changes in the brain after transient cardiac arrest. Circulation 1980: 61 : 1194-9. 7 Sadove MS. Yon MK, Hollinger PH. Johnston KS, Phillips FL. Severe prolonged hypoxic episode with complete recovery. / Am Med Assoc 1961 : 175: 1102-4. 8 Bates D,Caronna JJ, Cartlidge NEF et a/. A prospective study of nontraumatic coma: methods and results in 310 patients. Ann Neurol 1977: 2 : 211-20. 9 Hans P. Born JD. Chapelle JP, Milbouw G. Creatine kinase isoenzymes in severe head injury. / Neurosurg 1983: 58: 689-92. 10 Hans P. Born JD, Albert A. ‘Extrapolated’creatine kinases-BB isoenzyme activity in assessment of initial brain damage after severe head injury. J Neurosurg 1 9 8 7 : 6 6 : 714-7. 1 1 Kjekshus JK, Vaagenes P. Hetland 0. Assessment of cerebral injury with spinal fluid creatine kinase (CSF-CK) in patients after cardiac resuscitation. Scan / Lab Invest 1980: 4 0 : 437-44. 12 Longstreth WT. Clayson KJ. Sumi SM. Cerebrospinal fluid and serum creatine kinase BB activity after out-of-hospital cardiac arrest. Neurology 1 9 8 1 : 31: 455-8.
9
13 Vaagenes P. Urdal P. Melvoll R. Valnes K. Enzyme level changes in the cerebrospinal fluid of patients with acute stroke. Arch Neurol 1 9 8 6 : 43: 357-62. 1 4 Briem H, Lindquist L. Lundbergh P. Sego E, Skoldenberg B. Creatine kinase isoenzyme BB in cerebrospinal fluid from patients with meningitis and encephalitis. / Inject Dis 1 9 8 3 : 148:180. 15 Hofvendahl S. Gerhardt W, Ljungdal L. Olsen JS. Are ‘predictive values ’ from different materials comparable ? Acta Med Scand 1 9 7 8 : 623(suppl): 1 17-20. 16 Galen RS. Gambino SR. Beyond Normality : the Predictive Value and Eficiency of Medical Diagnoses. New York: John Wiley, 1 9 7 5 : 10-14. 17 Bell RD. Rosenberg RN. Ting R, Mukherjee A. Stone MJ. Willerson JT.Creatine kinase BB isoenzyme levels by radioimmunoassay in patients with neurological disease. Ann Neurol 1978: 3 : 52-9. 18 Clemmensen P. Strandgaard S. Rasmussen S. Grande P. Cerebrospinal fluid creatine kinase isoenzyme BB levels do not predict the clinical outcome in patients unconscious following cardiac resuscitation. Clin Cardiol 1987; 1 0 : 235-6. 19 Cho HW. Prilipko L. Meltzer HY et a/. Isoenzymes of creatine phosfokinase in white blood cells. Experientia 1977: 33: 166-7. 2 0 Thompson RJ. Graham JG, McQeen INF et a/. Radioimmunoassay of brain type creatine kinase isoenzyme in human tissues and in serum of patients with neurological disorders. Neurol Sci 1 9 8 0 : 47:241-51. 21 Phillips JP, Jones HM, Hitchcock R. A d a m N. Thompson RJ. Radioimmunoassay of serum creatine kinase BB as index of brain damage after head injury. Br Med J 1980; 281 : 777-9. Received 2 7 December 1988. accepted 9 June 1989.
Correspondence : Haraldur Briem. Medical Department, Reykjvik City Hospital, 108 Reykjvik. Iceland.
