Cross-reactivity of fosphenytoin in two human plasma phenytoin immunoassays

¹ Department of CNS Clinical Research and Development, Parke-Davis Pharmaceutical Research, Division of Warner-Lambert Company, Ann Arbor, MI 48105.

² Department of Pathology, University of Michigan, Ann Arbor, MI 48109.

³ Department of Pharmacokinetics, Dynamics, and Metabolism, Parke-Davis Pharmaceutical Research, Division of Warner-Lambert Company, Ann Arbor, MI 48105.
^a Address correspondence to this author at: 2800 Plymouth Road, Ann Arbor, MI 48105. Fax 734-622-7428; e-mail Alan.Kugler{at}wl.com.

	Abstract

Top Abstract Introduction Materials and Methods Results and Discussion References

 
The cross-reactivity of fosphenytoin, a phosphate ester prodrug of phenytoin, was investigated in the Abbott phenytoin TDx^®/TDxFLx^TM fluorescence polarization immunoassay (TDx) and the Behring Diagnostics phenytoin Emit^® 2000 enzyme-multiplied immunoassay (Emit). The first part of our study investigating cross-reactivity utilized in vitro correlation of the two immunoassays with a validated and specific phenytoin HPLC method used to assay plasma samples prepared in several phenytoin and fosphenytoin concentration combinations. Fosphenytoin cross-reacted with both immunoassays, but to a greater extent with TDx. In the second part of the study, empirically-derived models that best explained the in vitro data were used to predict "immunoassay-derived" phenytoin concentrations in plasma samples collected from actual patients after intravenous (IV) or intramuscular (IM) fosphenytoin dosing. The greatest degree of phenytoin concentration overestimation occurred at times when fosphenytoin concentrations were highest: within 1 to 2 h after IV infusion or during the first 2 to 4 h after IM injection. It is recommended that phenytoin concentrations not be monitored using these or other potentially nonspecific immunoanalytical methods for at least 2 h after IV fosphenytoin infusion or 4 h after IM fosphenytoin injection.

	Introduction

Top Abstract Introduction Materials and Methods Results and Discussion References

 
Parenteral phenytoin is a therapy of choice for treating patients with status epilepticus, for general loading dose administration in patients with seizures, and for loading and maintenance dose administration in patients unable to take oral anticonvulsants. However, phenytoin is poorly soluble in aqueous solution, requiring an alkaline pH and organic solvent vehicle; it precipitates in commonly used intravenous (IV) fluids (1) and may crystallize at the injection site after intramuscular (IM) administration (2). Fosphenytoin (Cerebyx^®, Parke-Davis) is a recently approved, freely water-soluble phosphate ester prodrug of phenytoin that delivers phenytoin into the systemic circulation with ~100% bioavailability, after IM or IV administration. After administration to humans, fosphenytoin is rapidly converted to phenytoin, with a plasma fosphenytoin half-life typically <15 min. Because of improved solubility, administration of fosphenytoin results in better tolerance at the infusion or injection site with fewer of the complications frequently associated with parenteral phenytoin (2)(3)(4)(5)(6)(7)(8)(9)(10), while still providing the same therapeutic benefits as phenytoin.

Determination of plasma fosphenytoin concentration may be of research interest but has little clinical value because fosphenytoin is short-lived and is not pharmacologically active. In contrast, plasma phenytoin concentration is strongly correlated with seizure control and with toxicity. The therapeutic range of phenytoin is narrow and its elimination kinetics are saturable, necessitating plasma phenytoin concentration monitoring as a guide to dose individualization. Phenytoin dosage is typically titrated to attain trough therapeutic total phenytoin concentrations of 10–20 mg/L. (11) Numerous methods are commercially available for rapid determination of phenytoin concentrations in plasma. Two of the most widely used methods (12) are enzyme-multiplied immunoassay (Behring Diagnostics Emit^® 2000 and others) (13) and fluorescence polarization immunoassay (Abbott TDx^®/TDxFLx^TM and others) (14). Phenytoin and fosphenytoin have a high degree of chemical similarity. The potential exists that high fosphenytoin concentrations in plasma after IV and IM administration (rarely exceeding 200 and 40 mg/L, respectively) might lead to aberrantly high phenytoin concentration determinations when immunoassays are used. Such an overestimation of actual plasma phenytoin concentrations might lead to an inappropriately conservative dosing regimen and failure to achieve therapeutic phenytoin concentrations. Here, we describe the extent of fosphenytoin cross-reactivity with TDx and Emit assay systems. In addition, empirical models derived from this in vitro data are applied to fosphenytoin and phenytoin concentrations determined via HPLC only from actual patients. The modeling approach illustrates the potential impact of this cross-reactivity in a clinical setting. This methodology for assessing the degree of fosphenytoin cross-reactivity is expected to be generally applicable to other nonspecific analytical methods.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results and Discussion References

 
apparatus
The TDx/TDxFLx analyzer was manufactured by Abbott Laboratories. Emit 2000 phenytoin determinations (Behring Diagnostics) were performed using a MIRA analyzer (Hoffmann-La Roche Laboratories). HPLC determination of fosphenytoin and phenytoin concentrations were performed by Pharmaco-LSR on a Waters WISP 712 apparatus (Waters Corporation) utilizing a Waters µBondapak C₁₈, 10 µm, 3.9 x 150 mm column for fosphenytoin and a Whatman RAC-2, Partisil 5ODS-3, 4.6 x 100 mm column (Whatman) plus an Applied Biosystems Spheri-5, 5 µm, RP-18, 4.6 x 100 mm column in series for phenytoin.

materials
All reagents used were reagent grade and of the highest quality and purity available. Omnisolv^® HPLC grade water (EM Science) was used for HPLC and for preparation and dilution of all solutions. Fosphenytoin (75 g/L, 5-mL ampoule) was manufactured by Parke-Davis Pharmaceutical Research, Division of Warner-Lambert Company. Phenytoin was from Aldrich. Hydrochloric acid (1 mol/L) was from Mallinckrodt and sodium hydroxide (500 g/L solution) was from EM Science. Fresh (unfrozen; shipped refrigerated), EDTA-treated human plasma was purchased from the Interstate Blood Bank. Phenytoin TDx reagents and calibrators were from Abbott, and phenytoin Emit reagents and calibrators were from Behring Diagnostics.

stock solutions and plasma samples
Seven aqueous fosphenytoin stock calibrator solutions, ranging from 0 to 8 g/L, were prepared by dilution of a 75 g/L fosphenytoin ampoule with distilled water. Six phenytoin stock calibrator solutions, ranging from 0 to 4 g/L, were prepared by dilution of a stock solution of phenytoin (8 g/L in 0.1 mol/L NaOH) with 0.1 mol/L NaOH.

Fresh EDTA-treated human plasma lots were pooled and had a measured pH of 7.45. The pH was adjusted to 7.30 with 0.1 mol/L HCl to ensure that the addition of basic phenytoin stock calibrator solutions would bring the final pH to 7.40. Plasma specimens containing phenytoin and fosphenytoin were prepared in the concentration combinations shown in Table 1 by adding 0.1 mL of phenytoin and 0.2 mL of fosphenytoin stock calibrator solution to a 10-mL volumetric flask partially filled with plasma and bringing the total volume to 10 mL with plasma. After preparation, plasma samples were divided into two plastic tubes and frozen at -20 °C pending analysis. The sample collection and storage conditions used have been previously shown to maintain stability of fosphenytoin and phenytoin (15)(16)(17).

assay procedures
TDx phenytoin reagents and calibrators were used to measure phenytoin in each plasma sample with a fluorescence polarization immunoassay analyzer, according to the manufacturer's instructions. The immunoassay performed on the Abbott TDx utilizes a polyclonal antisera. Similarly, Emit phenytoin reagents and calibrators were used to measure phenytoin in each plasma sample with a MIRA analyzer, according to the manufacturer's instructions. The Emit assay uses a monoclonal antibody reagent. In both immunoassays, samples containing combinations of drug that assayed at concentrations above the reporting range of the assay (40 mg/L) were diluted with EDTA-anticoagulated plasma and reanalyzed to obtain a quantitative result. For the specimens with added fosphenytoin and phenytoin listed in Table 1 , the immunoassay analyses were performed in triplicate. The duplicate set of plasma samples described above was also analyzed by a validated HPLC assay to determine baseline phenytoin and fosphenytoin concentrations for subsequent calculation of fosphenytoin cross-reactivity in the TDx and Emit assays. The HPLC assay was based on the method of Gerber et al. (18). Briefly, 1 mL of plasma was deproteinized with 2 mL of acetonitrile. The supernatant was then partitioned between methylene chloride (4 mL) and an ammonium phosphate buffer (0.5 mL, 0.067 mol/L (NH₄)₂HPO₄, pH 6.5). The resulting aqueous phase was analyzed for fosphenytoin, and the resulting organic phase was analyzed for phenytoin after evaporation. Fosphenytoin was quantitated by UV spectroscopy ( = 214 nm) after separation on a Waters µBondapak C₁₈ column at 35 °C in a mobile phase containing 700 mL/L 0.05 mol/L (NH₄)₂HPO₄, pH 6.5, and 300 mL/L methanol. Phenytoin was quantitated by UV spectroscopy ( = 220 nm) after separation on a Whatman RAC-2, Partisil 5ODS-3 column and a PE Applied Biosystems Spheri-5 RP-18 column in series in a mobile phase containing 530 mL/L methanol and 470 mL/L (NH₄)₂HPO₄, pH 7.0. Mean measured concentrations of quality-control samples analyzed with study samples deviated -3.32% to 14.0% and -4.10% to 2.89% from nominal concentrations for fosphenytoin and phenytoin, respectively. Precision of quality-control samples, determined as coefficients of variation, ranged from 4.11% to 4.63% and 2.68% to 9.65% for fosphenytoin and phenytoin, respectively.

data analysis
Plasma fosphenytoin and phenytoin concentration data were converted to molar concentrations before cross-reactivity calculations. In the calculations, molecular weights of 252.3 and 362.3 were used for phenytoin and fosphenytoin, respectively. Cross-reactivity of fosphenytoin with the phenytoin immunoassays was defined as the "phenytoin" concentration determined from the immunoassay procedure ([PHT]_Immunoassay) minus the phenytoin concentration determined by the validated phenytoin HPLC assay procedure ([PHT]_HPLC) divided by the fosphenytoin concentration determined from the fosphenytoin HPLC assay procedure ([FOS]_HPLC), expressed as a percentage (multiplied by 100%; Eq. 1 ).

(1)

To assess the impact of any cross-reactivity from Eq. 1 in an actual clinical setting, models were fit to the TDx or Emit in vitro data obtained for the plasma specimens listed in Table 1 . All models were derived empirically using SAS (SAS Release 6.12, SAS Institute Inc. Each model used the mean TDx or Emit immunoassay-determined phenytoin concentrations as the dependent variable and the HPLC-measured phenytoin and fosphenytoin concentrations as the independent variables. Using data obtained from both prepared plasma samples and actual patient specimens, the adequacy of each derived model was explored further by plotting model-predicted concentrations vs measured immunoassay phenytoin concentrations.

	Results and Discussion

Top Abstract Introduction Materials and Methods Results and Discussion References

 
All nominal and experimentally determined phenytoin and fosphenytoin concentrations are listed in Table 1 . Fosphenytoin cross-reacted with both assays, but the extent of cross-reaction was much greater with TDx. The degree of fosphenytoin cross-reactivity in the TDx immunoassay decreased with increasing fosphenytoin concentration. For example, in the absence of phenytoin, 4 mg/L fosphenytoin produced 120% cross-reactivity, whereas 75 mg/L fosphenytoin produced only 22% cross-reactivity. In the presence of phenytoin, the more clinically relevant situation, the same relationship held true. However, the magnitude of fosphenytoin cross-reactivity increased with increasing phenytoin concentration. In the absence of phenytoin, 4 mg/L fosphenytoin produced 120% cross-reactivity. The same concentration of fosphenytoin produced 193% cross-reactivity in the presence of 5 mg/L phenytoin and 315% cross-reactivity in the presence of 20 mg/L phenytoin. Thus, phenytoin appears to facilitate the cross-reactivity of fosphenytoin in the TDx assay.

Although fosphenytoin also exhibited cross-reactivity in the Emit assay, the degree of cross-reactivity was far less than that observed in the TDx assay. Furthermore, the fosphenytoin cross-reactivity profiles presented in Table 1 do not exhibit the nonlinearities (saturability and facilitation) observed for the TDx assay. In general, the cross-reactivity of fosphenytoin in the Emit assay was <10% except in some samples with low fosphenytoin concentrations, in which cross-reactivity estimates are more variable, probably because of analytical error.

The phenomena of difference in effect of added phenytoin to cross-reactivity of the two immunoassays has been described elsewhere for both digoxin and phenytoin. Matheke and Valdes (19) and Miller and Valdes (20) showed that when cross-reactivity of digoxin-like immunoreactive factor was tested in immunoassays for digoxin, the effect of added digoxin was variable among assays. In one assay, added digoxin increased the cross-reactivity of digoxin-like immunoreactive factor, whereas in another there was a decreased effect. The authors speculated that the effect was due to differential affinity of the analyte and cross-reactant for each antibody and would "be observed even with monoclonal antibodies". Rainey et al. (21) observed, using the TDx phenytoin assay, that there was an increased cross-reactivity with the major phenytoin metabolite in the presence of phenytoin.

The substantial cross-reactivity of fosphenytoin in both phenytoin immunoassays and the potential for cross-reactivity in all immunoassays suggest that this interaction should be assessed before use of a phenytoin immunoassay in a clinical setting where fosphenytoin is being administered. To evaluate the potential impact of the observed cross-reactivity in the clinical setting, empirical models were derived that best explained the TDx and Emit in vitro data, using the immunoassay-determined phenytoin concentrations ([PHT]_TDx and [PHT]_Emit, respectively) as the dependent variables and HPLC-measured phenytoin and fosphenytoin concentrations ([PHT]_HPLC and [FOS]_HPLC, respectively) as independent variables. The TDx and Emit in vitro data were best explained by Eqs. 2 , and 3 , respectively.

(2)

(3)

The adequacy of each model was tested by construction of 95% confidence intervals for the four numerical coefficients in Eq. 2 (identified as TDx coefficients A-D in Table 2 ) and the three numerical coefficients in Eq. 3 (identified as Emit 2000 coefficients A-C in Table 2 ). The predicted vs observed phenytoin concentrations from the in vitro experiments were then plotted (Fig. 1 ). The models fit the in vitro data extremely well. The empirical models were then evaluated further by plotting predicted vs assayed phenytoin concentrations for a set of plasma specimens from individuals receiving fosphenytoin. The data in Fig. 2 illustrate that the observed immunoassay results for actual patient specimens agree quite well with values predicted by the formulas, especially so for the TDx assay.

View larger version (16K): [in this window] [in a new window]   Figure 1. Model-predicted phenytoin concentration vs observed phenytoin concentration for plasma samples (Table 1 ). (Top) TDx/TDxFLx; (Bottom), Emit 2000. Solid lines are the lines of regression (TDx: y = 1.00x - 0.350, Syx = 1.78; Emit: y = 0.987x + 0.329, Syx = 0.884).

View larger version (15K): [in this window] [in a new window]   Figure 2. Model-predicted phenytoin concentration vs observed immunoassay phenytoin concentrations for actual patient specimens. (Top) TDx/TDxFLx; (Bottom) Emit 2000. Data points are means of duplicate measurements. Solid lines are the lines of regression (TDx: y = 0.970x + 0.885, Syx = 3.64; Emit: y = 0.881x + 0.960, Syx = 1.70).

Lastly, the empirical models were used to predict "immunoassay-derived" phenytoin concentrations in 1185 plasma specimens from 144 subjects and patients participating in fosphenytoin clinical trials. The plasma specimens had been assayed previously for phenytoin and fosphenytoin by HPLC. The difference between the phenytoin concentration predicted using Eqs. 2 , and 3 and the observed (HPLC) concentration was calculated and plotted against time to examine the over-prediction of phenytoin concentrations in the presence of fosphenytoin, using the immunoassay procedures (Fig. 3 ). Fig. 3 illustrates that the greatest degree of phenytoin concentration overestimation occurs at times immediately after IV infusion or shortly after IM injection. Although conversion to phenytoin begins immediately with IV fosphenytoin administration, initial plasma fosphenytoin concentrations may approach 200 mg/L before declining with a half-life of <15 min. In general, plasma fosphenytoin concentration-time profiles after IM administration are lower and more sustained than those after IV administration.

View larger version (18K): [in this window] [in a new window]   Figure 3. Phenytoin concentration difference (model-predicted concentration - HPLC-derived concentration) vs time, using TDx/TDxFLx (Top) and Emit 2000 (Bottom) immunoassay-derived data. Time referenced from the end of IV infusion (+) or time of IM injection ().

For treatment of status epilepticus, or for loading dose administration in patients with seizures, the goal of IV or IM fosphenytoin is to obtain a rapid therapeutic concentration of phenytoin in blood. After IV administration, maximum fosphenytoin concentrations have been observed to occur near the end of the infusion (18). In the same study, the concentration of phenytoin reached 90% of its maximum 12 min after the end of IV fosphenytoin infusion, whereas dose-dependent peak phenytoin concentrations occurred at a median times of 30–45 min (18). Because of the longer half-life of phenytoin, concentrations slowly decrease from peak values. Other reports also support the fact that substantial concentrations of free or total phenytoin are attained rapidly after IV infusion or IM injection (16)(22)(23). Because of this, plasma phenytoin concentrations could conceivably be monitored soon after administration of fosphenytoin to verify that therapeutic concentrations were achieved. However, after IV or IM administration, our data derived from 144 patients indicated that fosphenytoin cross-reactivity is clinically significant for up to 4 h in TDx determinations and for 1 h in Emit determinations. These findings reemphasize the importance of evaluating fosphenytoin cross-reactivity before use of an immunoassay to titrate phenytoin concentrations during fosphenytoin administration. It should be noted that, in a recent HPLC method development study (16), discrepant phenytoin results between TDx and HPLC were reported for an IV fosphenytoin-treated patient, which are explained by the data in Table 1 and Fig. 3 .

It could be argued that the cross-reactivity for diluted samples in Table 1 may not be the same at the higher concentrations and that multiplication of results to correct for dilution could lead to errors in the coefficients derived for the empirical model defined in Eqs. 2 , and 3 . Although experiments like those in Table 1 could be designed such that the combined cross-reactivity of the analytes would not exceed the analytic range, there were several reasons why we chose to drive empirical equations for all data points in the table. First, for TDx data a substantial number of the test samples required dilutions to bring results within assay range. Second, specimens from Table 1 that required dilution were known to contain concentrations of phenytoin or fosphenytoin that have been reported in clinical studies (16)(18). Third, in actual practice, clinical laboratories will generally dilute any results exceeding the assay limit and report a value that is corrected for the dilution factor, regardless of immunoreactivity that may be concentration dependent. Thus, the apparent plasma phenytoin concentrations in Table 1 represent the concentrations that would realistically be observed by a laboratory. Because of these considerations, we attempted to derive mathematical equations that reflected real life circumstances.

However, to validate further that these two mathematical equations could be used to accurately predict immunoassay phenytoin values, we also derived a second set of formulas based solely on apparent phenytoin concentrations from undiluted specimens in Table 1 . Although some subtle changes in the coefficient values were observed when these alternative equations were applied to actual undiluted patient specimens from the data set shown in Fig. 2 , the slope and intercept for the regression data were nearly the same (TDx slope, 0.942; intercept, -0.179; Emit slope, 0.873; intercept 0.342) as the values obtained for the original models illustrated in Fig. 2 .

Chromatographic assay methods (e.g., HPLC) accurately quantitate phenytoin concentrations in biological fluids in the presence of fosphenytoin. However, even with specific assay methods, phenytoin concentrations measured before conversion of fosphenytoin is complete will not reflect the phenytoin concentrations ultimately achieved. It is, therefore, advised that blood samples to assess phenytoin concentration should not be obtained for at least 2 h after IV fosphenytoin infusion or 4 h after IM fosphenytoin injection. (22).

In conclusion, considerable fosphenytoin cross-reactivity occurs in the TDx and Emit phenytoin immunoassays. The potential for this interaction exists with any phenytoin immunoassay and has its most profound impact immediately after IV fosphenytoin infusion or within a few hours of IM fosphenytoin injection, while fosphenytoin concentrations are substantial. The methodology presented should prove to be universally applicable in assessing the extent of cross-reactivity of other assays.

	References

Top Abstract Introduction Materials and Methods Results and Discussion References

 

1. Cloyd JC, Bosch DE, Sawchuk RJ. Concentration-time profile of phenytoin after admixture with small volumes of intravenous fluids. Am J Hosp Pharm 1978;35:45-48. [Abstract]
2. Serrano EE, Wilder BJ. Intramuscular administration of diphenylhydantoin. Arch Neurol 1974;31:276-278. [Abstract/Free Full Text]
3. Earnest MP, Marx JA, Drury LR. Complications of intravenous phenytoin for acute treatment of seizures. JAMA 1983;249:762-765. [Abstract/Free Full Text]
4. Cloyd JC, Gumnit RJ, McLain W, Jr. Status epilepticus: the role of intravenous phenytoin. JAMA 1980;244:1479-1481. [Abstract/Free Full Text]
5. Hauser WA. Status epilepticus: epidemiologic considerations. Neurology 1990;40(Suppl 2):9-13. [Web of Science][Medline] [Order article via Infotrieve]
6. Leppik IE, Boucher BA, Wilder BJ, Murthy VS, Watridge C, Graves NM, et al. Pharmacokinetics and safety of a phenytoin prodrug given IV or IM in patients. Neurology 1990;40:456-460. [Abstract/Free Full Text]
7. Ramsay RE. Pharmacokinetics and clinical use of parenteral phenytoin, phenobarbital, and paraldehyde. Epilepsia 1989;30(Suppl 2):S1-S3.
8. Perrier D, Rapp R, Young B, Kostenbauder H, Cady W, Pancorbo S, Hackman J. Maintenance of therapeutic phenytoin plasma levels via intramuscular administration. Ann Intern Med 1976;85:318-321.
9. Wilder BJ, Serrano EE, Ramsay RE, Buchanan RA. A method for shifting from oral to intramuscular diphenylhydantoin administration. Clin Pharmacol Ther 1974;16:507-513. [Web of Science][Medline] [Order article via Infotrieve]
10. Wilder BJ, Ramsay RE. Oral and intramuscular phenytoin. Clin Pharmacol Ther 1976;19(3):360-364. [Web of Science][Medline] [Order article via Infotrieve]
11. Levine M, Chang T. Therapeutic drug monitoring of phenytoin: rationale and current status [Review]. Clin Pharmacokinet 1990;19:341-358. [Web of Science][Medline] [Order article via Infotrieve]
12. . College of American Pathologists. Therapeutic drug monitoring–general, comprehensive survey, set Z-B 1997:1-31 CAP Northfield, IL. .
13. Oellerich M. Enzyme immunoassays in clinical chemistry: present status and trends [Review]. J Clin Chem Clin Biochem 1980;18:197-208. [Web of Science][Medline] [Order article via Infotrieve]
14. Lu-Steffes M, Jolley M, Pittluck G, Pannas H, Olive D, Wang CHJ, et al. Fluorescence polarization immunoassays of phenytoin and phenobarbital [Abstract]. Clin Chem 1981;27:1093.
15. Cao Z, Stapp D, Linder MW, Miller JJ, Valdes R. Conversion of fosphenytoin to phenytoin in vitro [Abstract]. Clin Chem 1997;43:S219-S220.
16. Cwik MJ, Liang M, Deyo K, Andrews C, Fischer J. Simultaneous rapid high-performance liquid chromatographic determination of phenytoin and its prodrug, fosphenytoin in human plasma and ultrafiltrate. J Chromatogr B 1997;693:407-414.
17. Varia SA, Schuller S, Sloan KB, Stella VJ. Phenytoin prodrugs III: water soluble prodrugs for oral and/or parenteral use. J Pharm Sci 1984;73:1068-1073. [Web of Science][Medline] [Order article via Infotrieve]
18. Gerber N, Mays DC, Donn KH, Laddu A, Guthrie RM, Turlapaty P, et al. Safety, tolerance and pharmacokinetics of intravenous doses of the phosphate ester of 3-hydroxymethyl-5, 5-diphenylhydantoin: a new prodrug of phenytoin. J Clin Pharmacol 1988;28:1023-1032. [Abstract]
19. Matheke ML, Valdes R, Jr. Antibody specificity reduces interferences by endogenous digoxin-like immunoreactive factors. J Clin Immunoass 1989;12:115-121.
20. Miller JJ, Valdes R, Jr. Methods for calculating crossreactivity in immunoassays. J. Clin Immunoass 1992;15:97-107.
21. Rainey PM, Rogers KE, Roberts WL. Metabolite and matrix interference in phenytoin immunoassays. Clin Chem 1996;42:1645-1653. [Abstract/Free Full Text]
22. Browne TR, Kugler AR, Eldon MA. Pharmacology and pharmacokinetics of fosphenytoin. Neurology 1996;46:S3-S7.
23. Pellock JM. Fosphenytoin use in children. Neurology 1996;46:S14-S16. [Medline] [Order article via Infotrieve]