| Article Access Statistics|
| Viewed||5128 |
| Printed||134 |
| Emailed||4 |
| PDF Downloaded||150 |
| Comments ||[Add] |
| Cited by others ||1 |
Click on image for details.
|Year : 2014 | Volume
| Issue : 3 | Page : 285-289
Phenytoin toxicity in patients with traumatic brain injury
Ananth P Abraham1, Ajay Vidyasagar2, Jayseelan Lakshmanan3, Shalini Nair1, Mathew Joseph1
1 Department of Neurological Sciences, Christian Medical College, Vellore, India
2 Department of Intensive Care Medicine, KMCH Speciality Hospital, Erode, Tamil Nadu, India
3 Department of Biostatistics, Christian Medical College, Vellore, India
|Date of Submission||23-Dec-2013|
|Date of Decision||20-Jan-2014|
|Date of Acceptance||27-May-2014|
|Date of Web Publication||18-Jul-2014|
Department of Neurological Sciences, Christian Medical College Hospital, Vellore - 632 004, Tamil Nadu
Source of Support: None, Conflict of Interest: None
Background: We observed that in patients with traumatic brain injury (TBI) who did not improve as expected, serum levels of phenytoin were in the toxic range and that their sensorium improved with modification of the dose. This led us to study the usage of phenytoin in patients with TBI. Aims: To determine the prevalence of phenytoin toxicity in TBI patients and to study the suitability of using ideal body weight (IBW) to guide phenytoin dosing. Setting and Design: Neurotrauma unit of a tertiary care centre in India. Prospective data collection from an already established protocol of drug level monitoring. Materials and Methods: The study cohort included 100 consecutive adult patients with mild or moderate TBI who were administered phenytoin based on IBW. Trough serum phenytoin and albumin levels were measured on day 4 after administration of the loading dose and actual body weight obtained when it was possible. Statistical Analysis: Chi-square was used for comparing categorical variables, student's t-test for continuous variables and multivariate regression analysis to obtain independent risk factors. Results: Clinical toxicity was observed in 15% of patients and biochemical toxicity in 36%, with a significant association between the two (P < 0.01). Using multivariate analysis, abdominal girth ≤75 cm (P = 0.07), neck circumference ≤34 cm (P = 0.025) and IV dose proportion ≥80% (P = 0.003) were independent risk factors for biochemical toxicity. The plot between actual weight and IBW showed that toxicity occurred when IBW was higher than actual weight. Conclusion: The prevalence of biochemical phenytoin toxicity was high, with independent risk factors being a higher proportion of IV administration and overestimation of weight by IBW. Clinical suspicion of phenytoin toxicity was a good predictor of biochemical toxicity.
Keywords: Ideal body weight, phenytoin, toxicity, traumatic brain injury
|How to cite this article:|
Abraham AP, Vidyasagar A, Lakshmanan J, Nair S, Joseph M. Phenytoin toxicity in patients with traumatic brain injury. Neurol India 2014;62:285-9
| » Introduction|| |
Phenytoin remains the most widely used anti-epileptic drug in post-traumatic seizure prophylaxis due to its cost-efficacy. However it has an unpredictable metabolism due to its zero-order or nonlinear pharmacokinetics and high affinity for plasma proteins. Wide variation has been reported in serum levels, especially in intensive care unit (ICU) patients in whom co-administration of multiple drugs and possible organ dysfunction can alter bioavailability.  Since actual body weight (ABW) is often difficult to obtain in patients admitted with traumatic brain injury (TBI), phenytoin dosing is usually based on ideal body weight (IBW) calculations. Therapeutic drug monitoring, the gold standard for guiding phenytoin dosage is not routinely available at many trauma centres in developing countries. The aims of this study were: To study the suitability of using IBW to guide phenytoin dosing, and to assess if biochemical toxicity could be predicted clinically in patients with TBI.
| » Materials and Methods|| |
One hundred consecutive TBI patients admitted to the neurotrauma unit who met the criteria for this study were included for analysis. The study was approved by the Institutional Review Board (Research and Ethics Committee, IRB Min No: 8535, dated 30.10.2013). Inclusion criteria were: (1) age ≥18 years; (2) clinical or radiological features that by our protocol required administration of anticonvulsants; and (3) moderate or mild head injury with Glasgow Comma Scale (GCS) score ≥9/15 admitted within 48 hours of injury. Exclusion criteria were: (1) exposure to phenytoin at another facility or previous phenytoin use; (2) surgical intervention; and (3) significant other injury causing hemodynamic or respiratory instability.
Loading dose of phenytoin was administered intravenously (IV) at 15-20 mg/kg over 20-30 minutes (50 mg/min). Maintenance dose was given at 4-6 mg/kg/day in three divided doses IV until the sensorium was adequate enough for oral medications. Dilantin© (Phenytoin sodium injection) and Eptoin© (Phenytoin sodium tablet) were drugs used in this study. Phenytoin was administered as per the IBW calculated using the Devine formula:  IBW (male) = 50.0 kg + 2.3 × each inch >5 feet IBW (female) = 45.5 kg + 2.3 × each inch >5 feet.
Neck circumference (just below the level of the laryngeal prominence) and abdominal girth (at the umbilicus) were measured in each patient at admission using a non-stretchable plastic tape, and ABW obtained whenever possible. Clinical suspicion of phenytoin toxicity was documented if the senior author felt that the patient was drowsier than expected considering the clinico-radiological presentation. Trough levels of serum phenytoin with concurrent serum albumin were obtained at least 72 hours after the administration of the loading dose. Adjusted phenytoin levels were calculated using the Sheiner-Tozer equation:  C corrected = C measured /(0.2 × serum albumin) +1 (C - concentration of phenytoin in μg/ml). In patients with end stage renal disease, Glomerular Filtration Rate (GFR) <15 ml/min/1.73 m 2 , the equation is: C corrected = C measured /(0.1 × serum albumin) +1. Statistical analysis was done using SPSS 16.0. The association between clinical toxicity and biochemical toxicity was done using Chi-square test with Yates correction. Univariate analysis was done for age, ABW, IBW, neck circumference, abdominal girth, GCS score at admission, loading dose of phenytoin, total maintenance dose, proportion of drug given IV, daily maintenance dose per unit IBW, and maintenance dose per unit ABW. Student's t-test was used to compare the mean (SD) of the above variables between the patients with and without biochemical phenytoin toxicity. In the multivariate analysis using logistic regression, abdominal girth, neck circumference, height, GCS score at admission and IV dose proportion were included. A P - value less than 0.05 was considered as significant for all analyses. The 95% confidence interval (CI) for median was calculated using bootstrap method with resampling of 1000 times. This was done using R software.
| » Results|| |
The mean age was 38.8 years (SD = 14.4) and 81 were male. All patients had a GCS score above eight with a mean of 12.6 (SD = 2.2). ABW was obtained in 57 patients. The phenytoin dosing was based on calculated IBW as shown in [Table 1]. The average daily maintenance dose (mg/kg/day) was seen to be higher when calculated with actual body weight (5.5 vs. 4.6). The respective median with 95% CI was 4.6 (4.4-4.8) and 5.3 (4.8-5.4). The mean serum phenytoin level value also changed when adjusted for serum albumin-the unadjusted mean was 15.1 μg/ml (SD = 6.6) and the mean adjusted phenytoin level was 19.2 (SD = 9.4).
A total of 36 patients had an adjusted serum phenytoin level that was in the toxic range as shown in [Table 2]. Clinical toxicity was suspected in 15 patients, of whom 13 had an adjusted serum phenytoin level that was greater than 20 μg/ml and none had serum levels in the subtherapeutic range (<10 μg/ml). There was a statistically significant association between clinical suspicion of toxicity and biochemical toxicity (P < 0.001). The diagnostic test characteristics of clinical suspicion were a sensitivity of 36%, specificity of 97%, positive predictive value of 87% and a negative predictive value of 73%.
On univariate analysis the risk factors for toxicity with significance were: ABW, abdominal girth, proportion of the maintenance dose given IV and GCS at admission [Table 3]. Among those who had their actual weight measured, those that developed toxicity were on average 10 kg lighter than those who did not develop biochemical toxicity (55.1 kg versus 65.1 kg; P = 0.001). The scatter plot [Figure 1] shows that most of the patients who developed toxicity had an actual weight lower than the calculated ideal body weight. On multivariate analysis, lower abdominal girth, neck circumference, and larger proportion of maintenance dose given IV (>80%) were shown to be independent risk factors for toxicity [Table 4]. Patients who had an abdominal girth ≤75 cm were 2.7 times more likely to have a serum phenytoin level over 20 μg/ml as compared to those whose abdominal girth was more than 75 cm. Patients who had a neck circumference ≤34 cm were 4.1 times more likely to have biochemical toxicity as compared to those with a neck circumference more than 34 cm. Those who had ≥80% of their total maintenance dose given IV had 5.2 times more chance of biochemical toxicity as compared to those who had received less than 80% IV.
|Table 3: Factors associated with corrected phenytoin levels >20 ìg/ml: Univariate analysis|
Click here to view
|Table 4: Factors associated with phenytoin toxicity (>20 ìg/ml): Multivariate analysis (logistic regression)|
Click here to view
In this study, there were 11 patients with impaired renal function (GFR 40-60 ml/min/1.73 m 2 ) amongst them only two had phenytoin toxicity. Five patients had mildly deranged liver function tests, in which none had hyperbilrubinemia, but all five patients had toxic serum phenytoin levels.
| » Discussion|| |
Phenytoin was first introduced in the market in 1938, and though there has been harsh criticism with regard to certain adverse drug reactions as well as its unpredictable pharmacokinetics, it remains arguably, the most cost-effective anti-epileptic drug (AED) in use today in the developing world.  The role of AED in preventing early post-traumatic seizures has been well established.  The study by Haltiner and colleagues had shown that the use of prophylactic phenytoin for 1 or 2 weeks reduces the incidence of early post-traumatic seizure without a significant increase in drug-related side effects.  Despite the availability of many newer AEDs, phenytoin has stood the test of time. ,
Phenytoin blocks voltage-sensitive sodium channels leading to delay in neuronal electrical recovery from inactivation. Phenytoin is largely protein (mostly albumin) bound (90%) and free phenytoin levels are affected by serum albumin concentration. The Sheiner-Tozer equation corrects the total phenytoin level for the corresponding serum albumin level, and has been found to be a useful tool to guide phenytoin dosing. ,, Phenytoin is primarily metabolised in the liver (>95%). Stowe et al., showed that children with severe, neurotrauma were found to have markedly altered protein binding and phenytoin metabolism,  and Markowsky et al., also demonstrated that phenytoin binding was significantly more variable with critical and convalescent patients with head injuries than healthy volunteers.  Following rapid IV infusion, phenytoin usually reaches therapeutic range within 18 hours. 
Neurological manifestations of phenytoin toxicity vary from nystagmus, drowsiness, dysarthria, ataxia coma, and paradoxical seizures at very high levels. Fine nystagmus occurs even at therapeutic levels. Since it is usually not feasible to ascertain the ABW of a patient with TBI, we administer phenytoin according to height based IBW calculation using the Devine formula.  Prior to this study, it was our observation that a number of patients with unexplained drowsiness had high adjusted phenytoin levels and that they improved with holding of phenytoin. It was this finding that prompted us to undertake this study. The main drawback of the Devine formula is that it suggests IBW values that are too low in women, but is still one of the most commonly used in India.
Correction of measured phenytoin levels to the serum albumin increased the mean concentration of phenytoin by >4 μg/ml. Thereby, the prevalence of biochemical toxicity increased from 20% to 36%. Since large segments of the Indian population are likely to have low serum albumin, this correction is essential. Clinical suspicion of toxicity was found to be a highly specific (97%) predictor of biochemical toxicity with a positive predictive value of 87%. However, the test was not sensitive (36%), since biochemical toxicity was not suspected in 23 patients. In these patients it was felt by the senior author that the computed tomography (CT) scan findings could explain the patient's altered sensorium.
On univariate analysis, lower ABW, smaller abdominal girth, a larger proportion of drug administered IV, and lower GCS score at admission were found to besignificantly associated with toxic phenytoin levels. On multivariate analysis low admission GCS score lost its significance and the independent risk factors for toxicity were neck circumference less than 34 cm, IV dosing proportion over 80% and possible abdominal girth less than 75 cm. A possible explanation for this finding is that patients with lower GCS score would have received higher a higher proportion of IV phenytoin.
These results suggest that thinly built patients (with narrow waist and neck) are more likely to develop toxicity. Majority of patients with phenytoin toxicity had ABW less than 70 kg and their ABW was less than IBW [Figure 1]. Thus, height based IBW formulas may lead to toxicity in lean patients. This has relevance to our country where malnutrition is a problem and majority of the population may be lean. Since it is difficult to obtain actual weights in critically ill patients, IBW could be calculated using abdominal girth, neck circumference, and height. Further and much larger studies are needed however, to substantiate this approach.
Mean plasma half life of phenytoin is similar whether given orally or IV.  In this study IV administration of phenytoin led to more toxicity although drug levels were measured 8 hours after the last dose. The oral bioavailability of phenytoin varies from 80-95%.  Studies have shown a decrease of serum phenytoin concentrations when it is co-administered with enteral feeding formulations. However, four randomised controlled trials in healthy volunteers failed to show any possible interaction.  Serum phenytoin levels may drop by as much as 70% when administered via nasogastric tube along with enteral feeds. , In our study enteral phenytoin was only administered orally. It is also possible that oral diet might interfere with phenytoin absorption.
The main limitation of this study was the measurement of a single phenytoin level as opposed to serial measurements which would have provided a better insight into the variability of serum levels. Another limitation was that the ABW could be measured only in 57 patients, with several patients not being able to stand at the time of discharge due to associated injuries or other causes. Also it would have been ideal if free phenytoin levels were also measured to corroborate the use of adjusted phenytoin levels.
The most important conclusion to be drawn from this analysis is that strictly following a height based IBW formula will result in phenytoin toxicity in undernourished patients, who form a significant part of our patient population. However it is still a more accurate method than "guess-timating" the body weight, if it is applied with a degree of common sense. It may be able to reduce this toxicity by deriving a better formula for IBW using abdominal girth and neck circumference so that the calculated weight is not a gross overestimation. Clinical suspicion of phenytoin toxicity has been found to be a very specific test, and we would suggest that measurement of the serum phenytoin level form part of the evaluation of unexplained drowsiness in patients with TBI.
| » References|| |
|1.||Pillai LV, Vaidya N, Khade AD, Hussainy S. Variability of serum phenytoin levels in critically ill head injured patients in intensive care unit. Indian J Crit Care Med 2008;12:24-7. |
|2.||Pai MP, Paloucek FP. The origin of the "ideal" body weight equations. Ann Pharmacother 2000;34:1066-9. |
|3.||Martin E, Tozer TN, Sheiner LB, Riegelman S. The clinical pharmacokinetics of phenytoin. J Pharmacokinet Biopharm 1977;5:579-96. |
|4.||Chisholm D. Cost-effectiveness of first-line antiepileptic drug treatments in the developing world: A population-level analysis. Epilepsia 2005;46:751-9. |
|5.||The Brain Trauma Foundation. The American Association of Neurological Surgeons. The Joint Section on Neurotrauma and Critical Care. Role of antiseizure prophylaxis following head injury. J Neurotrauma 2000;17:549-53. |
|6.||Haltiner AM, Newell DW, Temkin NR, Dikmen SS, Winn HR. Side effects and mortality associated with use of phenytoin for early posttraumatic seizure prophylaxis. J Neurosurg 1999;91:588-92. |
|7.||Pieracci FM, Moore EE, Beauchamp K, Tebockhorst S, Barnett CC, Bensard DD, et al. A cost-minimization analysis of phenytoin versus levetiracetam for early seizure pharmacoprophylaxis after traumatic brain injury. J Trauma Acute Care Surg 2012;72:276-81. |
|8.||Cotton BA, Kao LS, Kozar R, Holcomb JB. Cost-utility analysis of levetiracetam and phenytoin for posttraumatic seizure prophylaxis. J Trauma 2011;71:375-9. |
|9.||Tandon M, Pandhi P, Garg SK, Prabhakar SK. Serum albumin-adjusted phenytoin levels: An approach for predicting drug efficacy in patients with epilepsy, suitable for developing countries. Int J Clin Pharmacol Ther 2004;42:550-5. |
|10.||Krasowski MD, Penrod LE. Clinical decision support of therapeutic drug monitoring of phenytoin: Measured versus adjusted phenytoin plasma concentrations. BMC Med Inform Decis Mak 2012;12:7. |
|11.||Mlynarek ME, Peterson EL, Zarowitz BJ. Predicting unbound phenytoin concentrations in the critically ill neurosurgical patient. Ann Pharmacother 1996;30:219-23. |
|12.||Stowe CD, Lee KR, Storgion SA, Phelps SJ. Altered phenytoin pharmacokinetics in children with severe, acute traumatic brain injury. J Clin Pharmacol 2000;40:1452-61. |
|13.||Markowsky SJ, Skaar DJ, Christie JM, Eyer SD, Ehresman DJ. Phenytoin protein binding and dosage requirements during acute and convalescent phases following brain injury. Ann Pharmacother 1996;30:443-8. |
|14.||Salem RB, Wilder BJ, Yost RL, Doering PL, Lee C. Rapid infusion of phenytoin sodium loading doses. Am J Hosp Pharm 1981;38:354-7. |
|15.||Gugler R, Manion CV, Azarnoff DL. Phenytoin: Pharmacokinetics and bioavailability. Clin Pharmacol Ther 1976;19:135-42. |
|16.||Richens A. Clinical pharmacokinetics of phenytoin. Clin Pharmacokinet 1979;4:153-69. |
|17.||Au Yeung SC, Ensom MH. Phenytoin and enteral feedings: Does evidence support an interaction? Ann Pharmacother 2000;34:896-905. |
|18.||Bauer LA. Interference of oral phenytoin absorption by continuous nasogastric feedings. Neurology 1982;32:570-2. |
|19.||O′Hagan M, Wallace SJ. Enteral formula feeds interfere with phenytoin absorption. Brain Dev 1994;16:165-7. |
[Table 1], [Table 2], [Table 3], [Table 4]
|This article has been cited by|
||Combination therapy of levetiracetam and gabapentin against nonconvulsive seizures induced by penetrating traumatic brain injury
| ||Xi-Chun M. Lu,Ying Cao,Andrea Mountney,Zhilin Liao,Deborah A. Shear,Frank C. Tortella |
| ||Journal of Trauma and Acute Care Surgery. 2017; 83: S25 |
|[Pubmed] | [DOI]|