Comparison of high pressure liquid chromatography and fluorescence polarization immunoassay to assess quinidine pharmacokinetics

High pressure liquid chromatography (HPLC) and fluorescence polarization immunoassay (FPIA) were compared in a quinidine pharmacokinetic study. Six healthy male subjects received single oral doses of regular release (RR) quinidine sulfate, sustained release (SR) quinidine bisulfate and the same dose of the SR product with food (SR-F). Serum was collected for 49 h after each dose and analysed for quinidine by HPLC and FPIA. Using HPLC, there were no statistically significant differences between dosing regimens with respect to area under the curve (AUC) or terminal rate constant (K). The RR dose resulted in a higher peak plasma concentration (Cp) and shorter time to peak (Tp) than either of the SR doses (p less than 0.01). Food had no apparent effect on the bioavailability of the SR product. When using FPIA, food administration was found to increase the AUC for the SR product (p less than 0.05) and all three dosage regimens resulted in a different Cp and Tp (p less than 0.001). When comparing all pharmacokinetic parameters determined by each assay, FPIA resulted in a significantly lower K (p less than 0.01). Orthogonal regression of all serum quinidine concentrations showed that FPIA = 0.926 (HPLC) + 0.06 (r = 0.971, p less than 0.001). Analysis of quinidine concentrations in different concentration ranges revealed that FPIA overestimated HPLC for concentrations less than 1 micrograms ml-1 (p less than 0.001) and underestimated HPLC for concentrations greater than 2 micrograms ml-1 (p less than 0.01). Although the use of FPIA is appropriate for quinidine therapeutic drug monitoring, HPLC is the preferred assay method for assessing pharmacokinetic parameters in single dose bioavailability studies.     

Performance of a fluorescence polarization immunoassay system evaluated by therapeutic monitoring of four drugs.

Fluorescence polarization immunoassays (FPIA) for amikacin, gentamicin, quinidine, and theophylline (supplied by Roche Diagnostic Systems, made using a Cobas Fara centrifugal analyzer) were evaluated and compared with widely used monitoring analysis methods. For each drug, the between-assay imprecision was ascertained by calibration on the day of assay and by a stored calibration curve made at the beginning of the study. The precision of the amikacin and theophylline assays was acceptable [total coefficient of variation (CV) less than 7.5%] at all concentrations tested for each calibration mode. Imprecision of quinidine and gentamicin assays was significant at low concentrations (1.9 mg/L): total CV = 9.0% for quinidine assessed with stored calibration curve and total CV greater than 8.5% for gentamicin measured with the two calibration modes. The calibration curves for all four assays had a good stability (greater than 30 days). Linear regression analysis demonstrated close agreement between the FPIA (y) and the following comparative techniques (x): Abbott TDx assay for amikacin and gentamicin (r = 0.988, r = 0.974, respectively); Stratus fluorometric enzyme immunoassay for quinidine (r = 0.979); and EMIT Syva assay for theophylline (r = 0.993). It is concluded that fluorescence polarization immunoassay is a rapid and reliable method for the therapeutic monitoring of the four drugs tested. Moreover, the use of reagents on an instrument that can be implemented for a wide range of chemistries has significant advantages and cost benefits over dedicated instruments.     

NMR spectroscopic determination of preferred conformations of quinidine and hydroquinidine

NMR spectra of quinidine (I), hydroquinidine (II), and their respective acetyl derivatives (III and IV) were compared. The chemical shifts of some protons in I differed from those of their counterparts in II. These values were concentration dependent in I and II; they were similar in III and IV but not concentration dependent. The implications of these findings and the correlation of the NMR data with the preferred conformations are discussed.     

Quinidine-induced inhibition of the fast transient outward K+ current in rat melanotrophs.

1. The effect of quinidine on the fast-activating, fast-inactivating potassium current (IK(f] in acutely dissociated melanotrophs of the adult rat pituitary was examined. Macroscopic currents were measured by use of the whole-cell configuration of the patch clamp technique. 2. Bath application of quinidine caused a dose-dependent reduction of the peak amplitude of IK(f). The Kd for blockade of IK(f) at 0 mV was estimated to be 41 +/- 5.6 microM. 3. Quinidine elicited a dose-dependent increase of the rate of the decay of IK(f) and this effect was enhanced by membrane depolarization. The possibility that this phenomenon reflects an open channel blocking reaction is discussed. 4. Quinidine also caused a 5 mV hyperpolarizing shift of the steady-state inactivation curve and increased the half-time for recovery from inactivation. Quinidine did not affect the onset of inactivation measured at -30 mV. 5. Internal quinidine did not appear substantially to affect either the peak amplitude or kinetics of IK(f). 6. A study of some structural analogues showed that hydroquinidine and quinacrine had effects similar to those of quinidine. The effect of quinacrine on the amplitude and kinetics of IK(f) was also pH-dependent. Cinchonine, which bears a close structural resemblance to quinidine, was much less effective as a blocker of IK(f).     

New enzyme-linked chemiluminescent immunosorbent digoxin assay is free from interference of Chinese medicine DanShen

DanShen is a traditional Chinese medicine indicated for cardiovascular diseases. The potential interference of DanShen with serum digoxin measurement was investigated using a new enzyme-linked chemiluminescent immunosorbent (ECLIA) digoxin assay. Aliquots of drug-free serum were supplemented with ethyl acetate extract of DanShen (4 different brands studied), and apparent digoxin concentrations were measured by the ECLIA as well as fluorescence polarization immunoassay (FPIA) and a turbidimetric assay for comparison. Mice were also fed 4 DanShen preparations and apparent digoxin concentrations were subsequently measured. In another experiment, serum pools containing digoxin were further supplemented with DanShen extracts and digoxin concentrations were measured again by all 3 assays. No apparent digoxin concentration was observed when aliquots of drug-free serum pools were supplemented with DanShen and digoxin concentrations were measured by the ECLIA or the turbidimetric assay. In contrast, significant apparent digoxin concentrations were observed using FPIA, and the highest apparent digoxin concentration was observed with brand 4 of DanShen extract. Similarly, when mice were fed with this herb, significant apparent digoxin concentrations were also observed using FPIA, but neither ECLIA nor turbidimetric assay showed any apparent digoxin concentration. When aliquots of digoxin pool were further supplemented with various DanShen extract, the apparent digoxin concentrations were significantly increased when FPIA was used. In contrast, digoxin concentrations in the presence of DanShen extract compared well with the digoxin concentration of the original pool when ECLIA or turbidimetric assay was used. We conclude that DanShen does not interfere with serum digoxin measurement using a more recently released ECLIA digoxin assay.     

Comparison of bioassay, high-performance liquid chromatography, and fluorescence polarization immunoassay for quantitative determination of vancomycin in serum

This investigation was designed to compare three assay techniques, the traditional bioassay (agar diffusion), and two more recent techniques, high-performance liquid chromatography (HPLC) and fluorescence polarization immunoassay (FPIA), for the determination of vancomycin concentrations in serum. One hundred clinical samples obtained from patients receiving vancomycin were assayed by each method. The results from each assay were compared using linear regression analysis. The resultant correlation coefficients were as follows: 0.9996 for the HPLC versus FPIA, 0.7773 for the FPIA versus bioassay, and 0.7779 for HPLC versus bioassay. The FPIA technique was the easiest and fastest of the three methods; FPIA and HPLC were the most accurate.     

Rapid detection of oleander poisoning using digoxin immunoassays: comparison of five assays

Oleander is an ornamental shrub that grows in the United States, Australia, India, Sri Lanka, China, and other parts of the world. All parts of the plant are poisonous because the presence of cardiac glycoside oleandrin. Despite its toxicity, oleander extract is used in folk medicines. Because of its structural similarity, oleandrin cross-reacts with the fluorescence polarization immunoassay (FPIA) for digoxin. We studied the potential of detecting oleandrin in serum using 5 common digoxin immunoassays (FPIA, MEIA, both from Abbott; Beckman digoxin assay on Synchron LX, Chemiluminescent assay, CLIA from Bayer Diagnostics) and a recently FDA-approved turbidimetric assay on the ADVIA 1650 analyzer (Bayer). Aliquots of drug-free and digoxin-like immunoreactive substances (DLIS)-free serum pools were supplemented with ethanol extract of oleander leaves or oleandrin (Sigma Chemicals) in amounts expected in vivo after severe overdose. We observed significant apparent digoxin concentration with FPIA, Beckman, and the new turbidimetric assay (1 mL drug-free serum supplemented with 5.0 microL of oleander extract: apparent digoxin 2.36 ng/mL by the FPIA, 0.32 ng/mL by the MEIA, 0.93 ng/mL by the Beckman, 0.82 ng/mL by the new turbidimetric assay). The CLIA showed no cross-reactivity. Similar observations were made when serum pools were supplemented with oleandrin. Because cross reactivity should be tested in the presence of the primary analyte, we supplemented serum pools prepared from patients receiving digoxin with oleander extract or oleandrin. The measured digoxin concentrations were falsely elevated with the FPIA, Beckman, and turbidimetric assays, the highest false elevation being observed with the FPIA. Surprisingly, apparent digoxin concentrations were falsely lowered when MEIA was used. Digibind neutralizes free apparent digoxin concentration in vitro in serum pools supplemented with oleander extract, and this effect can be measured by the FPIA. We conclude that FPIA is most sensitive to detect the presence of oleander in serum. In contrast, the CLIA (no cross-reactivity) should be used for monitoring digoxin in a patient receiving digoxin and self-medicated with a herbal remedy containing oleander.     

Quantification of sirolimus and everolimus by immunoassay techniques: test specificity and cross-reactivity evaluation

The possible cross-reactivity of immunoassays with structurally-related drugs was investigated. Innofluor Certican (FPIA) calibrators were measured by using IMx Sirolimus assay (MEIA) and MEIA Sirolimus calibrators were analysed by using FPIA Certican assay. Drug concentrations were measured in 95 and 100 samples from renal transplanted patients (RTP) on sirolimus or everolimus treatment by using immunoassays and LC/ESI-MSMS. A high cross-reactivity was found both for MEIA and FPIA. High correlation degrees, confirmed by the Bland-Altman and the Eksborg tests, were found between drug concentrations measured in real samples by both immunoassays (r = 0.909 and r = 0.970, respectively). LC/ESI-MSMS analysis of samples containing sirolimus showed no positivity for everolimus. Similarly, samples from patients on treatment with everolimus resulted negative as far as regards sirolimus. MEIA and FPIA could be considered mutually reliable and accurate alternatives for the specific-drug immunoassay. It should be noticed that in patients switching from one drug to the other unreal overestimation of the blood levels of the current administered immunosuppressant can occur.     

New enzyme-linked immunosorbent digoxin assay on the ADVIA IMS 800i system is virtually free from interference of endogenous digoxin-like immunoreactive factors

Endogenous digoxin-like immunoreactive factors (DLIF) may crossreact with antidigoxin antibody and falsely elevate immunoassay results. Recently, a new enzyme-linked immunosorbent chemiluminescent assay for digoxin has been available for use on the ADVIA IMS (Integrated Modular System) 800i analyzer (Bayer Diagnostics). We studied potential interference of DLIF with this new digoxin assay. We analyzed 30 serum specimens from patients who have pathologic conditions that may increase serum DLIF concentrations. These patients were never exposed to digoxin or other agents that may lead to a measurable digoxin concentration. We also analyzed 10 specimens from neonates, 10 cord blood specimens, and 10 amniotic fluid specimens. Apparent digoxin concentrations were measured using the new enzyme-linked immunosorbent digoxin assay (IMS-Digoxin), a fluorescence polarization immunoassay (FPIA), and also a chemiluminescent immunoassay (CLIA, run on ACS:180(R) system from Bayer Diagnostics). We observed measurable apparent digoxin levels with the FPIA in 4 uremic patients (range 0.21-0.36 ng/mL, digoxin equivalent), 7 patients with liver disease (range 0.21-0.72 ng/mL), and 3 patients in the third trimester of pregnancy (0.22-0.66 ng/mL). We also observed measurable DLIF concentrations with the FPIA in 2 neonates (0.22 and 0.36 ng/mL), 5 cord blood specimens (range 0.21-1.18 ng/mL), and 5 amniotic fluid specimens (0.21-0.50 ng/mL). None of these DLIF-positive specimens showed any measurable digoxin concentration using the IMS-Digoxin or the CLIA assay. When serum specimens containing elevated concentrations of DLIF but no digoxin (as measured by FPIA) were supplemented with known concentrations of digoxin, we observed falsely elevated digoxin concentrations, as expected, only by the FPIA. In contrast, we observed a good agreement between the target and observed concentrations when the new IMS-Digoxin or the CLIA assay was used. We conclude that the IMS-Digoxin assay is free from interference of DLIF.     

GLC determination of quinidines in human plasma.

Human plasma was made alkaline and extracted with methylene chloride. To the extract was added the internal standard, cinchonine, followed by evaporation to dryness. The resultant residue was dissolved in a methanolic solution containing trimethylanilinium hydroxide. This solution was assayed by GLC for quinidines (quinidine and hydroquinidine). Evaluation of the method over a 0.5-10-mug/ml range in human plasma gave an overall precision and accuracy of +/- 4.5% (RSD and RE). Plasma of several patients was analyzed by the present method as well as by a fluorometric method for the level of quinidines. Results from the two methods were comparable.     

Comparison of the fluorescence polarization immunoassay and the microbiological assay methods for the determination of gentamicin concentration in human serum

The performance of the fluorescence polarization immunoassay (FPIA) was compared with that of a microbiological assay for the measurement of serum gentamicin concentrations. Within-run precision from duplicate assays of two concentrations (4 and 8 micrograms/ml) using FPIA and the microbiological assay yielded coefficients of variation (r) of 2.62%, 1.76% (n = 12) and 8.06%, 6.87% (n = 12), respectively. Day-to-day precision was estimated by repetitive analysis of 4 and 8 micrograms/ml control samples over a 3-week period. Coefficients of variation (r) were 2.57%, 3.09% (n = 8) and 10.71%, 14.20% (n = 8) for FPIA and the microbiological assay, respectively. Linear regression analysis performed on data from parallel determinations on 143 patient samples by the two methods showed correlations in the order of 0.74. The FPIA offers a rapid, efficient, and accurate system for therapeutic monitoring of gentamicin serum levels.     

Analytic performance evaluation of a new turbidimetric immunoassay for carbamazepine on the ADVIA 1650 analyzer: effect of carbamazepine 10,11-epoxide

Carbamazepine, an anticonvulsant, requires therapeutic drug monitoring. Recently Bayer HealthCare, Diagnostics Division released a turbidimetric immunoassay of carbamazepine on the ADVIA 1650 analyzer. We evaluated the analytic performance of this assay by comparing values obtained with this new assay in sera of 54 patients receiving carbamazepine with the values obtained by using a widely used fluorescence polarization immunoassay (FPIA) and a chemiluminescent immunoassay (CLIA). The new turbidimetric immunoassay for carbamazepine showed excellent precision. The low control showed a total CV of 4.9% (mean 2.86, SD 0.14 microg/mL), the medium control demonstrated a total CV of 3.5% (mean 7.79, SD 0.27 microg/mL), and the high control showed a total CV of 4.8% (mean 16.15, SD 0.78 microg/mL). The assay was linear up to a carbamazepine concentration of 20 microg/mL. The assay showed excellent dilution recovery and recovery of samples supplemented with carbamazepine (mean recovery 102.2%). We observed an excellent correlation between the values obtained by the FPIA (x-axis) assay and the new turbidimetric (y-axis) assay (y = 0.96 x - 0.46, r = 0.99, n = 54). We also observed excellent correlation between the values obtained by the CLIA (x-axis) and the turbidimetric (y-axis) assay (y = 1.10 x -0.32, r = 0.99, n = 54). However, the slope of 1.10 was higher than the slope of 0.96 observed with the regression equation obtained by using values obtained by the FPIA and the turbidimetric assay. The positive bias obtained with the new turbidimetric assay compared with the CLIA assay resulted from lower cross reactivity of carbamazepine 10,11-epoxide, the active metabolite of carbamazepine, with CLIA. On the other hand, the cross reactivity of the metabolite is similar between the new turbidimetric assay and the FPIA assay. We conclude that the new turbidimetric assay can be used for routine monitoring of carbamazepine in clinical laboratories.     

Evaluation of the AxSYM monoclonal cyclosporin assay and comparison with radioimmunoassay

Recently, a semiautomated fluorescence polarization immunoassay (FPIA) for determination of parent cyclosporin (CsA) has been developed for the Abbott AxSYM system. The new CsA assay measures the drug from an extracted whole blood specimen. The authors report here the evaluation of this new assay and the comparison with a previously validated radioimmunoassay (RIA) method (CYCLO-Trac SP). To assess the imprecision, the authors used tri-level controls supplied by both Abbott and Bio-Rad manufacturers. The within-run CV ranged from 4.4% to 7.3% and the between-day CV ranged from 4.4% to 7.6%. Mean recovery of the drug from clinical specimens spiked with kit calibrators was 108.4%. Fluorescence polarization immunoassay AxSYM (y) was correlated to RIA (x) by using 132 trough blood specimens (44 renal, 44 liver, and 44 heart) from transplant recipients and resulted in the following Passing-Bablok linear regression equation: y = 6.7 + 0.97x, r = 0.989, S(x/y) = 12.9. The percentage of overestimation (mean, range) by FPIA AxSYM versus RIA was (3.8%, range -17.7% to 39.1%). The results observed with this new method from follow-up studies in patients during the early course after transplant were not consistently higher than those obtained by RIA. These findings contrast with previously reported results that compared FPIA TDx assay with RIA. The authors conclude that FPIA AxSYM is a precise method for measuring CsA and offers results similar to those obtained by RIA with a marked reduction in assay time.     

Comparison of a monoclonal antibody fluorescent polarization immunoassay with monoclonal antibody radioimmunoassay for cyclosporin determination in whole blood.

We have compared the whole blood concentrations of parent cyclosporin A (CsA) using monoclonal fluorescence polarization immunoassay (FPIA) and radioimmunoassay (RIA) as well as polyclonal FPIA in kidney, heart, and bone marrow transplant patients (n = 89). A good correlation was found between monoclonal FPIA and monoclonal RIA (r = 0.96) and a slightly worse one between polyclonal and monoclonal FPIA (r = 0.90). The interassay coefficient of variation was satisfactory for all the methods-less than 5% for monoclonal FPIA. The monoclonal FPIA assay with Abbott TDx appears to provide rapid, precise, and accurate measurement of parent CsA. It is therefore useful for therapeutic monitoring of CsA in whole blood in kidney, heart, and bone marrow transplant patients.     

Comparison of high pressure liquid chromatography and fluorescence polarization immunoassay methods in a theophylline pharmacokinetic study

High pressure liquid chromatography (HPLC) and fluorescence polarization immunoassay (FPIA) were compared in a theophylline pharmacokinetic study. Eight healthy subjects received single 600-mg oral doses of two different sustained-release theophylline formulations. Fourteen blood samples were collected over 57 h after each dose, and the serum was analyzed for theophylline using both HPLC and FPIA methods. In comparing the two formulations using HPLC, there was no statistical difference in the area under the curve (AUC), terminal rate constant (k), or time of peak. However, there was a 13% difference in peak theophylline concentration (p less than 0.05). The same statistical conclusions were made for all parameters when using FPIA. When comparing the kinetic parameters determined with each assay, the AUC was 12% greater and the k was 17% smaller with FPIA (p less than 0.05). Orthogonal regression of all serum theophylline concentrations showed that FPIA = 1.04 HPLC + 0.20; r = 0.987, p less than 0.001. Stratification of serum theophylline concentrations into different ranges showed that FPIA overestimated the HPLC results in each range, but the percentage of overestimation was greater at lower concentrations (p less than 0.05). The use of FPIA seems appropriate in comparative studies of theophylline pharmacokinetics; however, the calculated kinetic parameters may differ slightly from those obtained with HPLC.     

Quinidine assays: enzyme immunoassay versus high performance liquid chromatography

Plasma quinidine results determined by enzyme multiplied immunoassay (EMIT) (Syva) were compared with results from a quinidine-specific high performance liquid chromatographic (HPLC) assay. During a clinical study, 16 patients with stable ventricular arrhythmias were treated with three oral quinidine preparations given during three consecutive 3-day periods. Seventeen plasma samples were drawn from each patient at steady-state during each period. Each specimen was divided into two portions, one for assay by EMIT and the other for assay by HPLC. EMIT assays were done on a Syva Autolab 6000 System using Syva quinidine kits and bilevel Ortho Diagnostics controls. The overall mean (+/- SD) quinidine concentrations by EMIT and HPLC were 2.16 +/- 0.58 and 1.81 +/- 0.60, respectively, n = 816, with a mean overall EMIT/HPLC ratio of 1.23 +/- 0.18. Mean ratios in individual patients ranged from 1.01 to 1.56; the average of mean individual ratios was 1.23 +/- 0.13. The EMIT assay, which also reports dihydroquinidine and small amounts of quinidine metabolites as quinidine, reported quinidine values that averaged 1.2-fold greater than results from a quinidine-specific HPLC method.     

The new enzyme-linked immunosorbent digoxin assay on the ADVIA Integrated Modular System is virtually free from oleander interference

Despite known toxicity of oleander, this product is used in herbal preparations. Oleander interferes with various digoxin immunoassays. It is possible that a person taking digoxin also may take oleander-containing herbal products, and digoxin immunoassays interfering with oleander cannot be used for therapeutic monitoring of digoxin. Recently, Bayer Diagnostics introduced a new enzyme-linked chemiluminescent immunosorbent digoxin assay for application on the ADVIA IMS System (ECLIA-digoxin). We studied potential interference of oleander with this new digoxin assay and found that this assay is virtually free from oleander interference. When aliquots of drug-free serum pools were supplemented with ethyl alcohol extract of oleander leaf or pure oleandrin standard, we observed significant apparent digoxin concentration when measured by the fluorescence polarization immunoassay (FPIA) but minimal digoxin-like immunoreactivity using the ECLIA digoxin assay. Because cross-reactivity should be studied in the presence of primary analyte, we prepared 2 serum pools using sera from patients receiving digoxin. Then aliquots of first digoxin pool were supplemented with oleandrin standard and aliquots of second digoxin pool with oleander extract. We observed significant increases in apparent digoxin concentration in the presence of both oleandrin and oleander extract using the FPIA. However, we observed no statistically significant change in digoxin concentration when ECLIA digoxin assay was used, indicating that this assay is virtually free from oleander interference.     

A modified fluorescence polarization immunoassay method incorporating fat emulsion (FE-FPIA) to determine cyclosporin A concentrations in rat skin.

We have developed a simple, sensitive and reliable assay procedure for cyclosporin A (CyA), a modified fluorescence polarization immunoassay method incorporating fat emulsion (FE-FPIA), to determine the CyA content in rat skin. The conventional fluorescence polarization immunoassay (FPIA) method for CyA using a commercially available FPIA kit, TDX cyclosporine monoclonal whole blood, was modified. A fat emulsion for intravenous infusion, Intralipos, was incorporated for dissolving the CyA extracted from the skin tissue, and a mixture of MeOH/purified water was used as the sample pretreatment medium instead of the precipitation reagent in the conventional FPIA kit intended for whole blood samples. These modifications enabled us to produce a reliable and the sensitive assay of CyA in skin tissue. The reproducibility (coefficient of variation), detection limit, and assay time for FE-FPIA were below 2%, 25 ng/ml, and about 24 min/24 samples, respectively, and were comparable with those for the whole blood samples determined by the conventional FPIA. Pre-purification of samples required by the HPLC assay is not needed in the FE-FPIA method. The usefulness of the FE-FPIA method in evaluating the topical pharmacokinetics of CyA in skin is discussed.     

The TDx assay for cyclosporine and its metabolites in blood samples compared with HPLC and RIA methods

Two methodologies have been developed to monitor cyclosporine (CsA) therapy: high-performance liquid chromatography (HPLC) and radioimmunoassay (RIA). Recently, a fluorescence polarization immunoassay (FPIA) has also become commercially available for the assay of CsA and its metabolites. The authors compared the results obtained with a modified FPIA with those found with two RIAs which use a polyclonal antibody, in order to verify if the FPIA assay is suitable for routine measurements in blood samples. Moreover, the accuracy of the RIAs and of the modified FPIA was checked against the results obtained by an HPLC technique assumed as a reference assay. The FPIA assay for CsA in blood samples seems preferable to the RIAs; in fact, as far as specificity is concerned, the TDx assay is comparable to polyclonal RIAs, while the precision (both within- and between-laboratories) is significantly better. Moreover, the TDx method is easier and faster to perform (20 samples can be assayed in about 30 min, while 2-4 h are necessary with RIA), with fewer handling steps; the instrumentation is automated and the reagents are more stable and less hazardous than those used in RIA.     

Evaluation of a mouse monoclonal antibody for the determination of serum theophylline by fluorescence polarization immunoassay

Theophylline concentrations in spiked human serum and serum specimens obtained from patients with normal and impaired renal function were measured by fluorescence polarization immunoassay (FPIA) using a mouse monoclonal antitheophylline antibody. The interday coefficients of variation of the assay after 14 days were 4.2, 3.3, and 2.4% at serum theophylline concentrations in pooled human serum of 7.1, 12.2 and 26.9 mg/L. Theophylline concentrations determined by FPIA and high performance liquid chromatography (HPLC) were used to generate the following linear regression equations relating the corresponding theophylline concentrations measured by each method in serum specimens from 50 patients with normal renal function (FPIA = 1.00 HPLC + 1.01, r = 0.98) and 50 patients with end-stage renal disease (FPIA = 1.04 HPLC + 0.08, r = 0.99). In contrast to previous studies performed with a polyclonal antitheophylline antibody, the precision, accuracy, and specificity of the FPIA were adequate for analysis of theophylline in serum specimens obtained from patients with end-stage renal disease.