Indian Pediatrics 2003; 40:1183-1186
CSF Glucose Concentrations in Infants with Febrile Convulsions and the Possible Effect of Acetaminophen
Mahmoud Mohammadi,* Mohammad R. Mohebbi** and Fariba Naderi***
Acetaminophen, CSF glucose, Febrile seizure.
Hyperglycemia and hyperglycorrhagia are common associations of febrile seizures(1,2). Rapid release of cortisol and adrenaline after seizures as a stress reaction induces elevated glucose concentration in blood and CSF(1). Fever intensity is also independently correlated with elevated CSF glucose level following febrile seizures(3).
The effect of acetaminophen administra-tion on CSF glucose concentration is unclear. The present study was done to evaluate the relation of fever and convulsion with CSF glucose concentration and to evaluate any effect of acetaminophen administration on CSF glucose concentration.
Subjects and Methods
The records of 128 patients aged 3-18 months admitted with febrile convulsions from March 1998 to January 2001 were studied retrospectively. All patients were subjected to a diagnostic lumbar puncture. A detailed history including preadmission acetaminophen administration was recorded. Patients with CSF pleocytosis, positive CSF culture, recent head trauma, metabolic dis-orders and epilepsy were excluded (n = 11). The children were assigned to Group I (n = 83) if they had not taken acetaminophen during the 6 hours before the convulsion and Group II (n = 34) if they had received acetaminophen. Children were considered febrile if the axillary temperature was higher than 37.2ºC(4). None of the patients had received any other medication in the previous 24 hours.
A simple febrile seizure was defined as a single generalized convulsion lasting less than 15 minutes in a febrile infant. Febrile seizures were defined as complex if they did not fulfill the above criteria(5).
CSF samples were obtained on an average of 3 hours after the first convulsion. Glucose concentration in the blood and CSF was determined by o-toluidine method. The aldehyde portion of glucose gets condensed with o-toluidine and the resulting complex is measured with a spectrophotometer (Spectronic 20D, Milton Roy Company, Belgium)(6).
Correlation between the quantitative variables was analyzed by Spearman’s correlation analysis. Comaprisons between the mean CSF glucose concentrations in the febrile and afebrile, and simple and complex febrile seizures were performed using Student’s t-test. The same method was used to compare the mean CSF glucose concentrations in different groups based on sex, duration of fever, multiplicity of convulsive attacks, and acetaminophen administration. A P-value of <0.05 was considered statistically significant.
Group I consisted of 83 children who were not administered acetaminophen during 6 hours before convulsion. In this group 70 children were febrile at the time of admission while 13 gave a history of fever at the time of convulsion but were afebrile at admission. In this group the mean CSF glucose concentration was significantly higher in febrile as compared to afebrile children (75.32 ± 2.94 mg/dL vs 66.15 ± 6.20 mg/dL; P = 0.014.
There was a linear correlation between CSF glucose concentration and body temperature in children who had not taken acetaminophen before admission (r = 0.515, P <0.001).
However a negative correlation (r = –0.389, P = 0.023) was found between the CSF glucose concentation and body temperature in those who had taken acetaminophen before admission (Group II; n = 34). The mean CSF glucose concentration was somewhat higher in Group II (78.68 mg/dL, n = 34), than Group I (73.89 mg/dL, n = 83) but the difference was not statistically significant (P = 0.076).
There was no statistically significant difference between CSF glucose concentration in the simple and complex febrile seizure groups and also between convulsive attacks longer than 15 minutes and shorter attacks. There was no statistically significant difference (73.53 mg/dL vs 75.79 mg/dL; P = 0.415) betwee the CSF glucose concentation of those with multiple convulsions (n = 32) compared to those with a single seizure (n = 84).
The duration of fever had a negative correlation (r = –0.175, P = 0.08) with CSF glucose concentration (n = 101). As expected, there was a linear correlation between the glucose concentrations in the blood and CSF (r = 0.244, P = 0.012) in 105 patients in whom the data was available. The mean CSF glucose concentration was not significantly different (77.38 mg/dL vs 73.00 mg/dL; P = 0.07) in boys (n = 61) and girls (n = 56).
Glucose enters the CSF by a facilitated transport system located in the choroid plexus(7,8). CSF glucose concentration has previously been shown to be directly proportional to that of blood(1,9). Our study reaffirms this correlation. The results show that in febrile and convulsive cases the CSF glucose concentration is increased, but we were not able to elucidate the role of each individual factor.
Hyperglycemia is regarded as a consequence of convulsion-induced release of both cortisol and adrenaline(1). Interleukin 1 beta (IL-1), an endogenous pyrogenic cytokine, inhibits insulin release(10,11) and stimulates the secretion of cortisol(11,12). This may explain the correlation between CSF glucose concentration and body temperature in the present report. Significant difference in mean CSF glucose concentrations in the convulsive infants presenting with and without fever may also be explained by IL-1 and its effects.
Our results show a negative correlation between the CSF glucose levels and body temperature in the patients administered acetaminophen prior to admission. This may be due to the non-specific o-toluidine method used for determination of CSF glucose. Additionally, acetaminophen preparation might have changed blood or CSF glucose levels. Alternatively, acetaminophen may be more effective in lowering temperature in children with higher CSF glucose levels.
Prospective studies to evaluate the effect of acetaminophen on the blood and CSF glucose concentration are warranted.
We especially thank Professor Bijan Djahanguiri for valuable discussions. We are also grateful to Dr. A.H. Kashi and Dr. M.R. Ebrahimkhani for their help with the statistical analysis.
Contributors: MM conducted and supervised the study. MRM conceived and carried out the study. FN supervised the study. MM and MRM reviewed the subject and drafted the paper. MM will act as the gurantor for the manuscript.
Competing interests: None stated.