Radioimmunoassay for monitoring zidovudine in dried blood spot specimens

¹ Newborn Screening Quality Assurance Program, National Center for Environmental Health, and
² National Center for HIV, Sexually Transmitted Diseases, and Tuberculosis Prevention, Centers for Disease Control and Prevention, Atlanta, GA.
^a Address correspondence to this author at: Centers for Disease Control and Prevention, F-19, 4770 Buford Hwy., NE, Atlanta, GA 30341-3724. Fax 770-488-4255; e-mail (SMTP) jvm0{at}.cdc.gov.

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
We modified and evaluated a RIA for serum and used the modified RIA to measure zidovudine in dried blood spot specimens (DBSs) routinely collected for newborn screening and tested anonymously for maternally acquired HIV antibodies in the national HIV Seroprevalence Survey Among Childbearing Women. DBS calibration and quality-control materials were used to adapt the serum assay to the DBS matrix. The assay had a limit of detection of 24 µg/L serum and was used to measure zidovudine from both whole DBSs and the eluate remaining after HIV antibody screening. We initiated a pilot study to investigate the assay's performance and assess its potential to determine the implementation of the US Public Health Service recommendations that HIV-infected pregnant women and newborns receive zidovudine treatment to reduce the risk of perinatal HIV transmission.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
The antiretroviral drug zidovudine has been measured in human serum or plasma by enzyme immunoassay (EIA) and time-resolved fluoroimmunoassay (1), by HPLC (2)(3)(4)(5)(6), and by fluorescence polarization immunoassay (3).¹ Zidovudine and zidovudine phosphates have been measured from peripheral mononuclear cells by a combined HPLC-RIA method (7). Currently, no method exists for the measurment of zidovudine in dried blood spot specimens (DBSs).

We modified a commercially available RIA kit for zidovudine in serum (8)(9) to quantitatively measure zidovudine in DBSs and DBS eluates. Calibration and quality-control (QC) DBS materials were prepared from whole blood containing known concentrations of zidovudine, which allowed us to accurately measure zidovudine concentrations in known and unknown DBSs. The zidovudine concentration of our calibration materials was confirmed by mass spectrometry. We also used the assay to measure zidovudine in the eluate that remained after testing DBSs for HIV antibodies. The residual eluate represented a diluted specimen that contained a much lower concentration than the original DBS.

Zidovudine has been shown to reduce the perinatal transmission of HIV (10). In 1994, the US Public Health Service issued recommendations that zidovudine be administered to HIV-infected pregnant women throughout their pregnancy and during labor, and to newborns during the first 6 weeks of life (11). We developed this method of measuring zidovudine in DBSs that had been tested anonymously for maternally acquired HIV antibodies in the the national HIV Seroprevalence Survey Among Childbearing Women (12) as a means of evaluating the public health impact of the recommendations. These specimens were collected from infants 24–48 h of age as part of routine screening for metabolic and genetic disorders. Because the half-life of zidovudine in newborns is 10 times greater than that of zidovudine in mothers (13)(14), we theorized that enough zidovudine should be available for measurement in DBSs and their eluates. To test these specimens, we developed an assay based on DBS calibrators and QC materials.

We report the outcome of adapting a serum RIA for use in analyzing DBSs and DBS eluates and discuss the potential application of this method to investigate zidovudine therapy among HIV-infected pregnant mothers and their newborns to reduce perinatal HIV transmission.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
modification of serum zidovudine assay for dbss
We used the Zidovudine-Trac^{TM 125}I RIA Kit (Incstar), a direct equilibrium assay that includes an isotopically labeled zidovudine derivative for the dose–response indicator; a method synopsis follows. Diluted serum specimens were mixed with I-labeled zidovudine and rabbit anti-zidovudine antibody. After an incubation step, a precipitating reagent containing a secondary goat anti-rabbit serum was added to each specimen. The antigen–antibody complexes were centrifuged to form pellets. The supernatants were decanted, and the radioactivity from the pellets was counted in a gamma counter. A dose–response curve was developed from liquid kit calibrators of known zidovudine concentrations.

To adapt the serum assay for DBSs, we initially punched a zidovudine-negative DBS (1/4" punch) into each of six zidovudine liquid kit calibrators and allowed the DBSs to elute into the liquid calibrators overnight at room temperature. Calibration curves were prepared for both the liquid calibrators and for the liquid calibrators that contained a DBS punch; results from unknown specimens were determined against these curves.

We also prepared zidovudine-supplemented DBS calibrators to correct for the DBS matrix. We added zidovudine (Sigma) into fresh, adult whole blood containing EDTA with the hematocrit adjusted to 50%. We prepared seven blood pools with an aqueous stock solution to give a calibration range of 0.06 to 70 ng of zidovudine per 1/4" DBS punch (which was the equivalent of 11 to 13 000 µg/L serum). Each pool was spotted onto filter paper cards (Scheicher & Schuell grade 903, lot W-901) and air-dried overnight at room temperature. DBS QC materials were also prepared from zidovudine-supplemented whole-blood pools (QC pools 1, 2, and 3) and from adult donors receiving zidovudine therapy. The assay was conducted by the overnight elution (room temperature) of DBS calibrators and controls, in duplicate, in 300 µL of kit diluent. Two hundred microliters per specimen were transferred to another tube, and I-labeled zidovudine tracer and rabbit anti-zidovudine antibody were added. After a 2-h incubation, goat anti-rabbit precipitating complex was added. After a 30-min incubation, bound and unbound tracer were separated by centrifugation and decanting of the supernatant. Precipitates were counted in a gamma counter (ICN Biomedicals), and a dose–response curve was calculated for the zidovudine DBS calibrators.

stability of zidovudine in dbss
Zidovudine-supplemented DBSs were placed in zip-closure bags (Bitran^®, Com-Pac International) at 4 °C, ambient temperature, and 45 °C in the presence and absence of desiccant (Minipax^®, Multiform Desiccants). The desiccant-containing bags were maintained at a humidity <40%. A duplicate set of specimens was kept at -20 °C as a control. Specimens were removed from each temperature at daily intervals for 1 week and at monthly intervals for 3 months. After removal, they were stored at -20 °C until analysis.

determination of limit of detection and assay cutoff
The limit of detection for the DBS zidovudine assay was determined by using linear regression to plot the standard deviations for a range of zidovudine concentrations vs response (15). Our limit of detection was determined to be 3 times the y-intercept, or 24 µg/L serum (16).

To evaluate the US Public Health Service recommendations for the use of zidovudine to reduce perinatal transmission of HIV, we tested the residual eluate remaining after HIV antibody testing of newborn DBSs. We did not have access to the original DBSs collected for newborn screening for metabolic disorders. Because the specimens to be tested consisted of previously diluted DBS eluates, an assay cutoff value for the screening of these eluates for zidovudine was calculated by analyzing 175 eluates that were negative for HIV antibodies. Eluates were retrieved from storage at -20 °C and thawed at room temperature. Volumes of eluates varied considerably, but most specimens had at least 40 µL available for assay. This volume was added to 460 µL of kit diluent, and then 200 µL of the diluted specimens were assayed, in duplicate, as described above. The response data, calculated as the ratio of bound counts per minute to the counts per minute of a zero DBS calibrator, were converted to [log(response)]. The mean [log(response)] of the HIV antibody-negative specimens was 0.022, and the 99% confidence limits were -0.005 and 0.049. The response data convert to zidovudine concentrations of nondetectable for the lower 99% confidence limit, and 24 µg/L serum for the upper limit. We chose the upper [log(response)] confidence limit of 0.049 as the assay cutoff value and used it as a corroborating indicator for the zidovudine-positive status of subsequently tested specimens along with the assay limit of detection.

matched dbs and plasma specimens
To evaluate the capacity of the RIA to detect zidovudine in patient DBSs, we used whole blood containing EDTA that had been collected from 38 adult patients receiving zidovudine therapy. Sex, age, weight, height, dose, time of last dose, and time of blood collection were noted for each patient. We dispensed 110 µL of whole blood from each patient onto filter paper cards and let the blood dry overnight. Plasma from each specimen was collected by centrifugation (3000g, 20 min), followed by incubation of the plasma specimens at 56 °C for 30 min to inactivate HIV. Aliquots (1.5 mL) from each specimen were removed, and the remaining plasma was stored at -20 °C for later use. Matched DBS and plasma specimens were analyzed by making a 1:21 dilution of the plasma specimens and assaying 200 µL of the diluted specimen according to the manufacturer's protocol, and by punching 1/4" DBS disks, in duplicate, into test tubes and following the procedure for DBSs as previously outlined.

determination of zidovudine from hiv-positive and hiv-negative dbs eluates
HIV antibody-positive residual DBS eluates collected from a state participating in the HIV Seroprevalence Survey Among Childbearing Women were retrieved from storage (-20 °C). These eluates had been tested by EIA (Genetic Systems, Sanofi Diagnostics Pasteur) and confirmed HIV antibody-positive by Western blot (Pageblot HIV-1, Genetic Systems). Several HIV-negative DBS eluates from the same time period were randomly selected and blindly inserted into the specimen set for testing. In addition, DBS QC materials with high and low zidovudine concentrations were eluted according to Genetic Systems' protocol, placed in identical containers, and labeled in a manner similar to that of the actual eluates, and then these eluates were blindly inserted, at a separate laboratory site, into the specimen set to make up 10% of the total specimens. DBS eluates were assayed by the modified zidovudine RIA as described.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Assay performance characteristics.
Performance characteristics of the Zidovudine-Trac ¹²⁵I RIA Kit modified for DBSs were determined. The primary modifications of the serum assay were matrix adjustments with the development of DBS calibrators, eluate optimization, and the elution procedure for DBSs. Although elution of zidovudine from DBSs appears to occur within a few hours, overnight elution is routinely used to ensure complete elution with patient DBSs that vary in age and environmental exposures. Calibration curves calculated from zidovudine-negative DBSs added to liquid kit calibrators and those calculated from zidovudine DBS calibrators were similar. Table 1 shows both the intraassay and interassay precision, with the CV ranging from 8% to 13% for DBS QC pools 1, 2, and 3 (n = 10) for the intraassay precision and the interassay precision CV ranging from 8% to 14% for the same QC pools. Table 2 shows the recovery and linearity from a zidovudine-supplemented blood pool that was serially diluted with additional blood, then spotted onto filter paper. Zidovudine recovery ranged from 80% to 110%, and the coefficient of determination (R) of the regression line was 0.999.

The zidovudine DBS RIA working calibration range was 0.06 to 70 ng per 1/4" punch (6 µL of serum, which is equivalent to 11–13 000 µg/L serum), whereas the serum kit working range was 0.25–274 µg/L (3). For comparison, an EIA for serum had a detection range of 33.3 to 1038 µg/L, and a time-resolved fluoroimmunoassay for serum had a detection range of 1.34 to 1038 µg/L zidovudine (1).

Stability of zidovudine in DBSs.
Zidovudine was found to be stable in the dried-blood matrix when DBSs were stored with desiccant at either 4 or -20 °C over the 3-month study period. At ambient temperature with desiccant, 90% of zidovudine could be recovered at the end of the study period, whereas 85% was recovered in the absence of desiccant. At 45 °C, 75% of zidovudine was recovered at the end of the study period (data not shown). The presence of desiccant did not increase the amount of zidovudine recovered after storage at an increased temperature (45 °C).

Analysis of matched DBS and plasma patient pairs.
We examined the feasibility and accuracy of measuring zidovudine from DBSs by comparing measurements made on matched plasma and DBSs from adults on zidovudine therapy. Zidovudine concentrations were calculated against both a liquid calibration curve for plasma measurements and a DBS calibration curve used for the DBSs. Fig. 1 shows a bias plot, where the zidovudine concentration of plasma specimens was subtracted from the matched DBSs. Whenever a plasma specimen tested positive for zidovudine, its corresponding DBS also tested positive, and whenever a plasma specimen was negative for zidovudine, its matched DBS was also negative. Twenty four of 38 specimens had no detectable zidovudine concentrations. When corrected for dilution factors, DBS concentrations were lower than the corresponding plasma concentrations, as indicated by the negative bias compared with the plasma zidovudine concentrations.

View larger version (17K): [in this window] [in a new window]   Figure 1. Degree of agreement between matched DBS and plasma specimens for 38 adult patients receiving zidovudine (ZDV) therapy. Twenty four of the 38 specimens had no detectable zidovudine concentrations.

Assay cutoff value and limit of detection.
Zidovudine was measured from eluates remaining after HIV antibody testing for the HIV Seroprevalence Survey Among Childbearing Women. Because our samples consisted of eluates from previously diluted and tested specimens, an assay cutoff value for screening eluates for zidovudine was determined from 175 HIV-negative samples. The 99% upper confidence limit was used as the screening cutoff to estimate whether zidovudine was present in the eluates. Samples that fell above the screening cutoff value also had zidovudine concentrations greater than or equal to the assay limit of detection. The cutoff value provided a means of confirming the zidovudine status of eluates based on the limit of detection. Fig. 2 shows a population distribution of [log(response)] of HIV-positive and HIV-negative eluates and how they sorted with respect to the screening cutoff concentration.

View larger version (30K): [in this window] [in a new window]   Figure 2. Frequency of zidovudine-positive and zidovudine-negative responses for HIV-positive (P) and HIV-negative (N) specimens based on an eluate screening cutoff of 0.049. Results to the left of the cutoff line are zidovudine-negative; those to the right are zidovudine-positive. n = 317.

Analysis of DBS eluates.
Once we established the ability of the zidovudine RIA to determine zidovudine in DBSs, we examined the feasibility of using DBS eluates from the HIV Seroprevalence Survey Among Childbearing Women to measure zidovudine. We conducted several experiments to determine assay variables. Zidovudine-supplemented DBSs were treated with reagents from Genetic Systems and eluted according to the manufacturers' protocol for EIA HIV antibody screening. The projected amount of eluate remaining should have been approximately 80 µL after initial and repeat EIA testing and Western blot analysis. Because we were concerned that we would not be able to detect the presence of zidovudine from such a limited sample, we initially investigated whether we could determine total zidovudine from matched patient DBS and plasma specimens by converting the zidovudine glucuronide metabolite (G-zidovudine) back to zidovudine with glucuronidase (8). We developed QC DBS materials supplemented with both zidovudine and G-zidovudine to quantitatively measure total zidovudine recovery after enzymatic treatment of samples. We successfully recovered total zidovudine from our QC materials and saw an increase in zidovudine from patient DBS samples treated with glucuronidase; however, the procedure was time consuming and added much complexity to the assay. We decided that measuring free zidovudine from eluates was sufficient to screen for the drug without adding to the complexity of the assay scheme and sacrificing limited sample material. Nonetheless, the assay may not be sufficiently sensitive to give zidovudine concentrations above the cutoff for those samples where the ratio of metabolite to parent drug is high.

Because most of the residual eluates available for testing contained at least 50 µL in volume, we chose 40 µL as our routine sample volume and added 460 µL of the zidovudine kit diluent, giving a further 1:12.5 dilution of an already diluted DBS sample (total dilution of 250x for the original whole-blood sample). Although the small volume available did not allow us to test each sample more than once, diluting the samples did allow us to assay aliquots in duplicate by using 200 µL of the diluted sample per tube. The recovery of zidovudine from the zidovudine-supplemented DBS eluates matched the expected concentrations, indicating that small volumes of samples eluted for HIV antibody testing can be used in tests to screen for zidovudine. Samples containing slightly less than 40 µL of eluate were also screened and identified with a unique code for entry in the database for statistical analysis.

Determination of zidovudine from HIV-positive and HIV-negative DBS eluates.
A group of 180 HIV antibody-positive and 49 HIV antibody-negative eluate samples were screened for zidovudine. These samples were collected from infants born between October and December 1994, following the release of the US Public Health Service recommendations for zidovudine therapy, and were tested anonymously for the HIV Seroprevalence Survey Among Childbearing Women. No information was available to confirm that the mother or infant had undergone zidovudine therapy. Of the 180 HIV antibody-positive samples tested, 51% tested positive for zidovudine, and no HIV antibody-negative samples tested positive for zidovudine. Included in the analysis of these samples were blind-coded QC materials, up to 10% of the total number of specimens in the sample set. The zidovudine DBS screening assay correctly identified all blinded QC materials. A subsequent set of randomly selected HIV antibody-positive and -negative specimens collected from the same geographical area between January and March 1995 showed a similar percentage of HIV antibody-positive samples testing positive for zidovudine. However, after decoding specimens, there were two zidovudine-positive specimens among the HIV antibody-negative specimens (17). In addition, a set of 200 HIV antibody-positive and 100 HIV antibody-negative specimens collected between September 1993 and March 1994, before the US Public Health Service recommendations were issued, were tested for zidovudine and only 1% of the HIV antibody-positive samples tested positive for zidovudine.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
The zidovudine RIA for serum has been modified to measure the drug in DBSs. The modified assay performed similarly to the original assay, and they both gave comparable results for matched plasma and DBS pairs. To adjust for the DBS matrix, we prepared a separate DBS calibration curve and appropriate DBS QC materials. QC materials consisting of DBSs eluted in a similar manner to that of newborn specimens tested in HIV Seroprevalence Survey Among Childbearing Women for HIV antibodies were also prepared. The assay, which required an overnight elution of DBS calibration materials, took 2 days compared with 1 day for serum specimens. We are developing a mass spectrometric method to measure zidovudine from DBSs. To date, only our zidovudine calibrators have been verified by mass spectrometry. We plan to use this technology to verify and assess assay specificity, especially at low zidovudine concentrations.

The zidovudine RIA was successfully used to quantitatively determine zidovudine from DBSs and to screen for zidovudine in HIV antibody-positive and HIV antibody-negative DBS eluates with minimal volumes. Because we only had access to stored eluates and not the original DBSs, this application of the DBS RIA for zidovudine should be considered a screening assay. In some cases, the eluates were in storage (at -20 °C) 1 year or more before testing. However, zidovudine was shown to be stable in the DBS matrix when stored at -20 °C, and it has proven to be stable in DBS eluates that had been in long-term storage at -20 °C, as demonstrated by the stability of QC materials over the course of assay development and implementation.

Our study has shown that zidovudine could be measured in residual eluates saved from the HIV Seroprevalence Survey Among Childbearing Women or in DBSs collected from newborns and adults. This methodology is being applied to larger epidemiological studies of high-risk HIV populations (17). The zidovudine DBS assay could provide an important analytical tool for the assessment of zidovudine therapy among HIV-positive pregnant women and their newborns. In addition, this information could be useful to state and local public health professionals, who need population-based data concerning the prevalence of zidovudine therapy among HIV-positive pregnant women to target resources in a more cost-beneficial manner and to adjust HIV counseling and testing protocols. This assay is also being applied to studies investigating the association between zidovudine therapy and the reduction of viral load in women and newborns as assessed by reverse-transcriptase PCR.

In this study, we used residual specimens routinely collected from newborns that had been screened for inborn errors of metabolism and tested anonymously for HIV antibodies. DBSs are ideal collection devices (18) for the clinical monitoring of zidovudine therapy among newborns or adults with consent. The zidovudine RIA has sufficient sensitivity to assay very small blood volumes. Evaluating DBSs by RIA thus provides a means of monitoring drug therapy and dosing. DBSs can also be transported easily and inexpensively, allowing for remote collection of samples followed by zidovudine testing elsewhere. We plan to measure zidovudine in blood collected from HIV-positive pregnant women and newborns enrolled in several international clinical trials for the prevention of perinatal HIV transmission, and to assist ongoing epidemiologic studies designed to quantitatively assess zidovudine treatment protocols for diverse populations.

	Acknowledgments

 
We acknowledge Lanqing Wang and John T. Bernert, Jr. for their efforts in developing a mass spectrometric method for the measurement of zidovudine from DBSs.

	Footnotes

 
¹ Nonstandard abbreviations: DBS, dried blood spot; QC, quality control; EIA, enzyme immunoassay; and G-zidovudine, zidovudine glucuronide metabolite.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Tadepalli SM, Quinn RP. Determination of zidovudine concentration in serum by enzyme-linked immunosorbent assay and by time-resolved fluoroimmunoassay. J Acquir Immune Defic Syndr 1990;3:19-27.
2. Kupferschmidt R, Schmid RW. Specific routine determination of 3'-azido-3'-deoxythymidine (AZT) in plasma by partly automated liquid chromatography. Clin Chem 1989;35:1313-1317. [Abstract/Free Full Text]
3. Stretcher BN. State-of-the-art zidovudine monitoring. J Clin Lab Anal 1991;5:60-68. [Web of Science][Medline] [Order article via Infotrieve]
4. Borvak J, Kasanicka J, Mayer V. HPLC-monitoring of AZT in HIV-infected patients plasma: a critical study. Acta Virol 1992;36:428-434. [Web of Science][Medline] [Order article via Infotrieve]
5. Nattouf H, Davrinche C, Sinet M, Farinotti R. High-performance liquid chromatographic assay for 3'-amino-3'-deoxythymidine in plasma, with 9-fluorenyl methyl chloroformate as the derivatization agent. J Chromatogr B Biomed Appl 1995;664:365-371. [Web of Science][Medline] [Order article via Infotrieve]
6. Nebinger P, Koel M. Determination of serum zidovudine by ultrafiltration and high-performance liquid chromatography. J Pharm Biomed Anal 1994;12:141-143. [Web of Science][Medline] [Order article via Infotrieve]
7. Slusher JT, Kuwahara SK, Hamzeh FM, Lewis LD, Kornhauser DM, Lietman PS. Intracellular zidovudine (Zidovudine) and Zidovudine phosphates as measured by a validated combined high-pressure liquid chromatography–radioimmunoassay procedure. Antimicrob Agents Chemother 1992;36:2473-2477. [Abstract/Free Full Text]
8. Tadepalli SM, Puckett L, Jeal S, Kanics L, Quinn RP. Differential assay of zidovudine and its glucuronide metabolite in serum and urine with a radioimmunoassay kit. Clin Chem 1990;36:897-900. [Abstract/Free Full Text]
9. Quinn RP, Orban B, Tadepalli S. Radioimmunoassay for retrovir, an anti-human immunodeficiency virus drug. J Immunoassay 1989;10:177-189. [Web of Science][Medline] [Order article via Infotrieve]
10. Connor EM, Sperling RS, Gelber R, Kiselev P, Scott G, O'Sullivan MJ, et al. Reduction of maternal–infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. N Engl J Med 1994;331:1173-1180. [Abstract/Free Full Text]
11. . Centers for Disease Control and Prevention. Recommendations of the U.S. Public Health Service Task Force on the use of zidovudine to reduce perinatal transmission of human immunodeficiency virus. MMWR 1994;43(No. RR-11):1-20.
12. Gwinn M, Pappaioanou M, George JR, Hannon WH, Wasser SC, Redus MA, et al. Prevalence of HIV infection in childbearing women in the United States. Surveillance using newborn blood samples. JAMA 1991;265:1704-1708. [Abstract/Free Full Text]
13. O'Sullivan MJ, Boyer PJJ, Scott GB, Parks WP, Weller S, Blum R, et al. The pharmacokinetics and safety of zidovudine in the third trimester of pregnancy for women infected with human immunodeficiency virus and their infants: Phase I Acquired Immunodeficiency Syndrome Clinical Trials Group study (protocol 082). Am J Obstet Gynecol 1993;168:1510-1516. [Web of Science][Medline] [Order article via Infotrieve]
14. Boucher FD, Modlin JF, Weller S, Ruff A, Mirochnick M, Pelton S, et al. Phase I evaluation of zidovudine administered to infants exposed at birth to the human immunodeficiency virus. J Pediatr 1993;122:137-144. [Web of Science][Medline] [Order article via Infotrieve]
15. Taylor JK. Quality assurance of chemical measurements. Chelsea, MI: Lewis Publishers, 1987:328 pp..
16. Long GL, Winefordner JD. Limit of detection: a closer look at the IUPAC definition. Anal Chem 1983;55:712A-719A.
17. Gwinn M, Mei JV, Spruill CL, Dobbs TL, Hannon WH, LaLota M. Hospital variation in intrapartum use of zidovudine [Letter]. JAMA 1997;277:299.
18. National Committee for Clinical Laboratory Standards. Blood collection on filter paper for neonatal screening programs; approved standard. NCCLS publication LA4–A2. Wayne, PA: NCCLS, 1992:1–25..