ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Mar. 2004, p. 824–831 0066-4804/04/$08.00⫹0 DOI: 10.1128/AAC.48.3.824–831.2004 Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Vol. 48, No. 3
Quality Assurance Program for Clinical Measurement of Antiretrovirals: AIDS Clinical Trials Group Proficiency Testing Program for Pediatric and Adult Pharmacology Laboratories Diane T. Holland, Robin DiFrancesco, Judith Stone, Fayez Hamzeh, James D. Connor, and Gene D. Morse* Adult and Pediatric AIDS Clinical Trials Group Pharmacology Laboratory Committees, Pediatric AIDS Clinical Trials Group, Division of AIDS, National Institute of Allergy and Infections Diseases, National Institutes of Health, Bethesda, Maryland Received 15 July 2003/Returned for modification 18 September 2003/Accepted 2 December 2003
Clinical trials designed to compare antiretroviral regimens, investigate therapeutic drug monitoring, or measure pharmacometrics often include protease inhibitors (PIs), nonnucleoside reverse transcriptase inhibitors (NNRTIs), and nucleoside reverse transcriptase inhibitors, requiring the measurement of these antiretrovirals in plasma. Within the adult and pediatric AIDS Clinical Trials Group (ACTG), a network of Pharmacology Support Laboratories (PSLs) is a component of the group laboratory infrastructure and conducts these types of pharmacologic assays. The adult ACTG has developed a comprehensive quality assurance program for the conduct of clinical pharmacology protocols, one component of which is the antiretroviral proficiency testing (PT) program that has been implemented between the adult and pediatric pharmacology laboratories of the ACTG. PT testing samples were prepared and distributed in July 2001, February 2002, and July 2002. High, medium, and low concentrations of PIs (indinavir, saquinavir, amprenavir, lopinavir, ritonavir, and nelfinavir) and NNRTIs (nevirapine and efavirenz) were added to drug-free EDTA plasma and distributed, on dry ice, to eight ACTG PSLs. One testing laboratory used liquid chromatography-tandem mass spectrometry, and seven used high-performance liquid chromatography-UV analysis. A result was considered acceptable if it was within 20% deviation of the assigned concentration. For all concentrations of PIs evaluated, 96% of samples tested (430 of 448 measurements) met the acceptance criteria. For both NNRTIs, 100% of samples tested (140 of 140 measurements) met the acceptance criteria. In conclusion, the PT program results presented demonstrate excellent interlaboratory agreement for all antiretrovirals tested and provide support for the merger of plasma concentration data among laboratories for large clinical trials. 2nd Int. Workshop Clin. Pharmacol. HIV Ther., abstr. 5.2, 2001). Clinical trials are often designed to compare the efficacy of antiretroviral regimens that include protease inhibitors (PIs), nonnucleoside reverse transcriptase inhibitors (NNRTIs), and nucleoside reverse transcriptase inhibitors while simultaneously evaluating drug-drug interactions. In addition, other types of clinical trials that include pharmacokinetic and pharmacodynamic assessment as well as therapeutic drug monitoring (TDM) also require the measurement of antiretroviral plasma concentrations (3, 5, 8, 12, 14; E. P. Acosta, D. V. Havlir, and D. D. Richman, Program Abstr. 7th Conf. Retrovir. Opportun. Infect., abstr. 455, 2000; D. Burger, M. Felderhof, P. Phanupak, et al., Program Abstr. 8th Conf. Retrovir. Opportun. Infect., abstr. 730, 2001; D. M. Burger, P. W. Hugen, J. Droste, and A. D. Huitema, for the Athena Group, Abstr. 2nd Int. Workshop Clin. Pharmacol. HIV Ther., abstr. 6.2a, 2001; D. M. Burger, P. W. Hugen, J. Droste, and A. D. Huitema, for the Athena Group, Abstr. 2nd Int. Workshop Clin. Pharmacol. HIV Ther., Abstr. 6.2b., 2001; P. Clevenbergh, J. Durant, R. Garraffo, et al., Program Abstr. 8th Conf. Retrovir. Opportun. Infect., abstr. 260B:119, 2001; C. V. Fletcher, P. Anderson, T. Kakuda, et al.,Program Abstr. 8th Conf. Retrovir. Opportun. Infect., Abstr. 259, 2001; R. M. Hoetelmans, R. P. Heeswijk, V. P. Meenhorst, et al., Abstr. 12th World AIDS Conf. Geneva, abstr. 12360, 1998; D. Mayers, T. Merigan, D. Wentworth, J. Neaton, M. Hoover, R. Hoetelmans, W. Verbiest, J. Baxter, and the CPCRA 046
The measurement of antiretrovirals in biological samples has progressed from the use of individual assays for each drug in a regimen to the more common application of assay methods that measure multiple drugs simultaneously. In addition, the progression has included a transition from earlier trials that utilized high-performance liquid chromatography (HPLC) and radioimmunoassay to measure nucleoside analogs such as zidovudine, didanosine and zalcitabine (6) to a predominance of HPLC methods, now incorporating liquid chromatographytandem mass spectrometry (LC-MS-MS) assays (9–11, 16, 18; C. Fletcher, Program Abstr. 7th Conf. Retrovir. Opportun. Infect., abstr. 101, 2000). Currently, the management of human immunodeficiency virus (HIV) infection is characterized by combination therapeutic regimens that include different classes of antiretrovirals with the potential for complex drug-drug interaction. This has led to the necessity for conducting numerous pharmacokinetic drug interaction studies in support of clinical trials that examine new combination regimens (4, 15; D. Mayers, T. Merigan, D. Wentworth, J. Neaton, M. Hoover, R. Hoetelmans, W. Verbiest, J. Baxter, and the CPCRA 046 Study Team, Abstr.
* Corresponding author. Mailing address: AACTG Pharmacology Support Laboratory, School of Pharmacy and Pharmaceutical Sciences, University at Buffalo, The State University of New York, Room 317, Hochstetter Hall, Buffalo, NY 14260. Phone: (716) 645-3635, ext. 252. Fax: (716) 645-2001. E-mail:
[email protected]. 824
VOL. 48, 2004
CLINICAL MEASUREMENT OF ANTIRETROVIRALS
Round
3a
NVP
200 272 363 250 1,200 270
2,000 2,702 3,422 2,500 4,000 2,500
NA NA 9,139 5,000 8,000 5,290
1a 1b 2a 2b 3a 3b
EFV
434 293 618 250 366 754
4,438 2,817 4,120 2,500 4,120 2,500
NA NA 7,621 5,000 7,416 5,000
1a 1b 2a 2b 3a 3b
a Target values in ng/ml, expressed as free base. a and b, prepared at UCSD and JUH, respectively; NA, not applicable, not prepared for this round.
EFV
I, S, N, R All I, N, R, A, L All I, N, R, A, L All All I, S, R, L
Plsc
None EFV NVP Both EFV Both Both Both
NNRTI(s)
Drug(s) assayed
H A (245), C, G H A (245), D, G A(212),C,G A (210), C, G A (250), C, F A (246), C, F
Study Team, Abstr. 2nd Int. Workshop Clin. Pharmacol. HIV Ther., abstr. 5. 2, 2001). The adult and pediatric AIDS Clinical Trials Groups (ACTGs) conduct multicenter clinical trials of antiretroviral therapies, many of which include pharmacologic substudies that require that the drug concentrations from the same trial be measured in multiple laboratories using different analytical methods. These trials may also involve TDM, requiring the pharmacology laboratory to be certified under the Clinical Laboratory Improvement Amendments of 1988 (CLIA) (17). Therefore, a comprehensive quality assurance-quality control program was established and integrated into the overall clinical pharmacology program of the adult and pediatric ACTGs. To
TABLE 2. Laboratory capabilities
9,610
Method(s) of analysis for drugb:
3,844
NVP
481
LPV
LPV
APV
2a 3a
RTV
7,714 8,054
NFV
2,611 4,643
SQV
264 155
IDV
APV
C, G
1a 1b 2a 2b 3a 3b
C, G
NA NA 11,700 5,000 9,360 4,696
E, G C, G
2,168 2,500 2,209 3,000 2,808 3,000
E, G C, G
217 250 117 1,500 157 1,500
E, G C, G
RTV
C, F C, G
1a 1b 2a 2b 3a 3b
D, F
NA NA 7,700 4,609 5,580 4,613
D, G
2,525 2,500 2,496 2,365 2,564 2,500
C, G E, F E, F
250 250 234 740 335 800
H H B, D, F A (280), H A (284), A (280), A (280),
NFV
D, G C, G C, G C, G C, G
1a 1b 2a 2b 3a 3b
H A (215), B, D, G A (205), A (212), A (210), A (212), A (220),
NA NA 5,220 5,000 6,960 4,664
D, G C, G C, G C, G
3,458 2,500 1,600 2,329 2,088 2,500
H A (265), B, D, G A (205), A (212), A (210), A (212), H
346 250 164 250 209 215
D, G C, G C, G C, G C, G
SQV
A (240), A (215), B, D, G A (245), A (212), A (210), A (212), A (234),
1a 1b 2a 2b 3a 3b
D, G C, G C, G C, G
NA NA 11,320 4,617 7,899 4,368
A (240), A (215), B, D, G A (205), A (212), A (210), A (212), H
2,830 2,500 2,830 2,500 3,396 2,302
C, G C, G C, G
257 250 113 250 209 263
A (240), A (235), H A (205), H A (210), A (212), A (234),
IDV
D, G C, G C, G C, G C, F
High
A (210), A (215), B, D, G A (205), A (212), A (210), A (212), A (210),
Middle
LAB IDa
Target value for drug concn Low
1 2 3 4 5 6 7 8
Drug
a Laboratory ID no. b A, HPLC-UV (wavelength) (47 of 53 [89.0%]); B, LC-MS-MS (6 of 53 [11.0%]); C, liquid/liquid extraction (34 of 53 [64.0%]); D, protein precipitation (14 of 53 [26.0%]); E, solid-phase extraction (5 of 53 [9.0%]); F, single drug/run (8 of 53 [15.0%]); G, multiple drugs/run (45 of 53 [85.0%]); H, no assay available or no results reported (10 of 53 [19.0%]). Total numbers of labs reporting results were as follows: for IDV, eight; for SQV, six; for NFV, seven; for RTV, eight; for APV, six; for LPV, seven; for NVP, five; and for EFV, six; for a total of 53 assays. c Single-letter abbreviations for drugs: I, IDV; S, SQV; N, NFV; R, RTV; A, APV; L, LPV; N, NPV; E, EFV.
TABLE 1. Antiretroviral drugs spiked into EDTA plasma for proficiency testinga
825
826
HOLLAND ET AL.
ANTIMICROB. AGENTS CHEMOTHER. TABLE 3. Summary of three rounds of proficiency testing resultsa Low target
Drug or round
%dev, min/max
No. (%) within 20% dev
Middle target %dev, min/max
No. (%) within 20% dev
High target %dev, min/max
IDV 1a 1b 2a 2b 3a 3b Total
⫺30 to 18 ⫺17 to 13 ⫺19 to 29 ⫺16 to 30 ⫺24 to 44 ⫺17 to 18
5/6 (83) 6/6 (100) 5/6* (83) 6/7 (86) 4/6* (67) 5/5** (100) 31/38 (86)
⫺22 to 20 ⫺9 to 15 ⫺13 to 5 ⫺17 to 11 ⫺8 to 6 ⫺8 to 6
4/6 (67) 6/6 (100) 7/7 (100) 7/7 (100) 7/7 (100) 5/5** (100) 36/38 (95)
N/A N/A ⫺16 to 12 ⫺11 to 15 ⫺5 to 7 ⫺17 to 14
SQV 1a 1b 2a 2b 3a 3b Total
⫺14 to 13 ⫺14 to 13 ⫺6 to 14 ⫺3 to 15 ⫺28 to 7 ⫺12 to 5
5/5 (100) 5/5 (100) 4/4** (100) 4/4** (100) 5/6 (83) 4/4** (100) 27/28 (96)
⫺30 to 13 ⫺5 to 11 ⫺9 to 5 ⫺13 to 7 ⫺8 to 7 ⫺12 to ⫺2
4/5 (80) 5/5 (100) 5/5 (100) 5/5 (100) 6/6 (100) 4/4** (100) 29/30 (97)
N/A N/A ⫺8 to ⫺1 ⫺10 to 2 ⫺7 to 2 ⫺19 to 14
NFV 1a 1b 2a 2b 3a 3b Total
⫺17 to 18 ⫺21 to 16 ⫺23 to 21 ⫺17 to 15 ⫺19 to 12 ⫺35 to 23
7/7 (100) 6/7 (86) 5/7 (71) 7/7 (100) 7/7 (100) 4/6 (67) 36/41 (88)
⫺13 to 19 ⫺19 to 17 ⫺13 to 17 ⫺12 to 14 ⫺14 to 14 ⫺15 to 24
7/7 (100) 7/7 (100) 7/7 (100) 7/7 (100) 7/7 (100) 5/6 (83) 40/41 (98)
N/A N/A ⫺13 to 17 ⫺18 to 17 ⫺14 to 11 ⫺7 to 7
RTV 1a 1b 2a 2b 3a 3b Total
⫺11 to 13 ⫺9 to 25 ⫺14 to 9 ⫺15 to 8 ⫺26 to 18 ⫺15 to 18
6/6* (100) 6/7 (86) 6/6* (100) 7/7 (100) 5/6*** (83) 6/6** (100) 36/38 (95)
⫺10 to 17 ⫺14 to 16 ⫺9 to 8 ⫺7 to 9.0 ⫺9 to 8 ⫺11 to 12
7/7 (100) 7/7 (100) 7/7 (100) 7/7 (100) 7/7** (100) 6/6** (100) 41/41 (100)
N/A N/A ⫺14 to 4 ⫺6 to 5 ⫺11 to 4 ⫺9 to 14
APV 2a 3a Total
⫺10 to 16 ⫺12 to 23
5/5 (100) 5/6 (83) 10/11 (91)
⫺13 to 11 13 to 12
5/5 (100) 6/6 (100) 11/11 (100)
⫺14 to 5 ⫺5 to 7
⫺19 to 11
7/7 (100)
⫺9 to 3
7/7 (100)
LPV 3a Total for PIs NVP 1a 1b 2a 2b 3a 3b EFV 1a 1b 2a 2b 3a 3b Total for NNRTIs
147/161 (91)
⫺8 to 15 ⫺13 to 11 ⫺8 to 14 ⫺8 to 8 ⫺11 to 8 ⫺16–15
5/5 (100) 4/4 (100) 4/4 (100) 3/3 (100) 4/4 (100) 4/4 (100)
⫺5 to 13 ⫺11 to 13 ⫺5 to 2 ⫺6 to 14 ⫺5 to 9 ⫺3 to 4
5/5 (100) 5/5 (100) 5/5 (100) 5/5 (100) 6/6 (100) 3/3** (100) 53/53 (100)
⫺4 to 12
164/168 (98)
⫺11 to 0.4 ⫺3 to ⫺0.8 ⫺15 to 10 ⫺12 to 1 ⫺9 to 6 ⫺3 to 13 ⫺8 to 11 ⫺7 to 9 ⫺9 to 3 ⫺1 to 4 ⫺11 to 4 ⫺14 to 6
No. (%) within 20% dev
Total no. (%) within 20% dev
7/7 (100) 7/7 (100) 7/7 (100) 5/5** (100) 26/26 (100)
93/100 (93)
5/5 (100) 5/5 (100) 7/7 (100) 4/4** (100) 21/21 (100)
77/79 (97)
7/7 (100) 7/7 (100) 7/7 (100) 7/7 (100) 27/27 (100)
103/109 (94)
7/7 (100) 7/7 (100) 7/7** (100) 6/6** (100) 27/27 (100)
104/106 (98)
5/5 (100) 6/6 (100) 11/11 (100) 7/7 (100)
32/33 (97) 21/21 (100) 430/448 (96)
119/119 (100)
5/5 (100) 4/4 (100) 4/4 (100) 3/3 (100) 4/4 (100) 4/4 (100)
N/A N/A ⫺10 to 9 ⫺5 to 0.9 ⫺3 to 9 ⫺11 to 9
5/5 (100) 5/5 (100) 5/5 (100) 5/5 (100) 6/6 (100) 3/3** (100) 53/53 (100)
N/A N/A ⫺3 to 5 ⫺8 to 4 ⫺8 to 16 ⫺13 to 14
4/4 (100) 3/3 (100) 4/4 (100) 4/4 (100)
5/5 (100) 5/5 (100) 6/6 (100) 3/3** (100) 34/34 (100)
*, 1 lab BQL; **, interfering peak for lab; ***, 1 lab interfering peak and 1 lab BQL; a, b, prepared at UCSD and JHU, respectively; N/A, not applicable, not spiked; No. within 20% deviation, number correct/number done; %dev, min/max, range of percent deviation.
VOL. 48, 2004
CLINICAL MEASUREMENT OF ANTIRETROVIRALS
827
TABLE 4. Proficient testing results (of all drugs) by laboratorya Category of target drug concn
LOW
Medium
High
FIG. 1. Proficiency testing summary by laboratory no. (x) versus percent deviation from assigned target value. Rounds represented by symbols: for low and medium, E, round 1; 〫, round 2; 䊐, round 3; for high, E, round 2; 〫, round 3. Dashed lines represent the ⫾20% acceptance range.
assure that drug assays performed at different pharmacology laboratories provide data suitable for integration into a single database for study analysis, an ongoing proficiency testing program was established that would be consistent with the clinical trials that the ACTG was planning to implement. Proficiency testing is also a CLIA requirement for laboratories when providing drug concentration results for clinical management. This report highlights the 2001-2002 results and findings of the ACTG proficiency testing program. MATERIALS AND METHODS Analytical powders. The following reagents were obtained through the AIDS Research and Reference Reagent Program, National Institute of Allergy and Infectious Diseases, National Institutes of Health: indinavir (IDV), saquinavir (SQV), nelfinavir (NFV), ritonavir (RTV), nevirapine (NVP), and efavirenz (EFV). All except efavirenz moved as a single spot in thin-layer chromatography, indicating very high purity. Efavirenz yielded a faint second spot in thin-layer chromatography. Amprenavir (APV) was obtained from GlaxoSmithKline; purity was not given and was assumed to be 100%. Lopinavir (LPV) was obtained from Abbott Laboratories; purity was not given and was assumed to be 100%. Drug powders were weighed on a precalibrated balance and dissolved in 100% methanol (HPLC grade; Fisher Scientific). Stock solutions were prepared in two ACTG pharmacology laboratories, at University of California, San Diego (UCSD) and John Hopkins University (JHU), and consisted of around 10 ml of
LAB
Mean % deviation (mean for all labs) for round: 1
1 2 3 4 5 6 7 8
1 2 3 4 5 6 7 8
1 2 3 4 5 6 7 8
2
3
6.9 12.0 11.7 13.1 7.0 11.1 3.1 ID
10.0 8.4 8.7 11.6 8.3 12.8 7.2 ID
13.1 19.1 12.3 11.1 8.6 14.0 6.3 ID
(9.3)
(9.6)
(12.1)
6.3 9.2 5.5 9.1 6.0 13.8 4.3 ID
6.9 3.9 6.5 8.8 5.3 4.7 6.5 ID
8.3 4.9 7.6 9.3 4.5 6.2 3.0 ID
(7.7)
(6.1)
(6.3)
ND ND ND ND ND ND ND ID
5.3 4.3 7.9 8.7 7.1 4.9 8.1 ID (6.6)
6.8 6.4 6.6 6.9 5.1 4.0 3.7 ID (5.7)
a Results are expressed as mean percent deviation minus absolute inaccuracies. ND, not spiked; ID, insufficient data; LAB, laboratory ID no.
drug at concentrations between 1 and 3 mg/ml. Stock solutions were stored at ⫺70°C or colder. Blinded specimens. From July 2001 to July 2002, the ACTG QA/QC program has distributed three rounds of proficiency testing samples. A laboratory involved in only pediatric work (UCSD) was chosen to prepare proficiency testing samples for the adult laboratories, and in a similar fashion, a laboratory involved in adult-only studies (JHU) prepared samples for the pediatric laboratories. (In this way UCSD and JHU each received one set of blinded samples.) EDTA plasma (purchased from Biological Specialty Corp., Colmar, Pa.) was used as the sample matrix, and target concentrations were chosen to cover the expected therapeutic ranges of the drugs (between 100 and 12,000 ng/ml depending on the drug) (2, 3). These ranges were chosen based on known drug values from previous pharmacokinetic studies. Samples were prepared at low, medium, and high concentrations of each drug, and target values are shown in Table 1. The first round was prepared in July 2001 and consisted of low and medium concentrations of IDV, SQV, NFV, RTV, EFV, and NVP. The second round was prepared in February 2002 and consisted of low, medium, and high concentrations of each drug, with the addition of APV by UCSD. The third round was prepared in July 2002 and was similar to the second round, with the addition of LPV in the UCSD samples. Samples were prepared so that all the PIs were together, and tubes were labeled A, B, or C, representing different concentrations for each drug. The NNRTIs were prepared in a similar way and labeled D, E, or F. Aliquots of 4 to 5 ml of each proficiency testing sample were stored at ⫺70°C in polypropylene tubes prior to distribution to testing laboratories. Target concentrations were entrusted to the ACTG Data Management Center, Frontier Science and Technology Research Foundation, to ensure blinding. Blinded sample shipment. Eight ACTG Pharmacology Support Laboratories are involved in either adult (six of eight) or pediatric (five of eight) studies or
828
HOLLAND ET AL.
ANTIMICROB. AGENTS CHEMOTHER. P value for significance was 0.05, and all pairwise comparisons were further performed using Bonferroni’s correction. SAS version 8 software was used. Results reporting. Results were tabulated, and individual results were expressed as percent deviation from the target values. Within 20% deviation was considered satisfactory. Participating laboratories received preliminary performance reports within 4 weeks, and this was followed by a formal report with comments on individual performances and a request to investigate results that were not satisfactory.
RESULTS
FIG. 2. Proficiency testing summary by round (x) and percent deviation from assigned target value (y). Dashed lines represent the ⫾20% acceptance range.
both (three of eight), and all laboratories take part in proficiency testing. Each laboratory received seven tubes, six spiked samples (A to F), and one blank plasma from both UCSD and JHU, and the tube source was identified. Samples were sent on dry ice within a few days of preparation. Laboratories were encouraged to measure both sets of samples and were instructed to process proficiency testing samples along with patient samples. They were asked to return results within 6 weeks of receipt. Data analysis and acceptance criteria. Weighed-in drug amounts were taken as preliminary target values. After results were received back from the testing laboratories, the preliminary target value was compared to the mean concentration determined from statistical analysis of the testing laboratories’ results, and if the difference was ⬎5% (after the removal of outliers), then the statistical mean drug concentration was used as the target value. If the difference was ⬍5%, the weighted-in concentration was taken as the target value. Target values are shown in Table 1. Laboratory performance was determined to be satisfactory if drug concentration results were within a 20% deviation from the target value. Twenty percent was chosen as an appropriate acceptance criterion to bring the ACTG program in line with the international antiretroviral proficiency testing program (1, 7) and CLIA fixed criteria for toxicology proficiency testing (17), although the 20% limit is more rigorous than that of CLIA, which is generally 25%. Our objective was also to establish performance criteria based on clinical requirements for good patient care (13). ANOVA. Multifactorial analysis of variance (ANOVA) was performed to evaluate three main effects, those of the drug being measured (APV, IDV, LPV, NFV, RTV, and SQV), the target concentration level (low, medium, or high), and the source of the drug (either from UCSD or JHU), on the absolute inaccuracy. All possible interactions of these main effects were also tested. The
Laboratory capabilities. Results were reported from all participating laboratories for various combinations of drugs depending on their assay capability (Table 2). In summary: only three laboratories reported results for all drugs (identification [ID] numbers 1, 6, and 7 in Table 2); all laboratories were able to measure IDV and RTV; laboratory ID no. 1 did not measure APV, LPV, or the NNRTIs: laboratory ID no. 2 did not measure NVP; laboratory ID no. 3 did not measure SQV and EFV; laboratory ID no. 5 did not measure SQV and NVP; and laboratory ID no. 8 did not measure NFV and APV. One laboratory (ID no. 3) used LC-MS-MS, while the other laboratories used reverse-phase-HPLC with UV detection. The majority of drug assays were multidrug assays (85%), and samples were prepared by either liquid-liquid extraction (64%), protein precipitation (26%), or solid-phase extraction (9%). Data results. Results are detailed in Table 3. The table summarizes the number of results reported for each drug concentration, the percent deviation range (minimum to maximum), the number of satisfactory results (within 20% deviation), and the percentage of satisfactory results. All NNRTIs (140 of 140) were 100% satisfactory. Overall, 96% of all results (430 of 448) for PI drugs met the acceptance criteria. Target concentrations had a dramatic influence on acceptable measurements which were 91, 98 and 100% for the low, medium, and high target concentrations, respectively. IDV and NFV proved to be the most difficult to measure at the low end, with satisfactory results of 86 (31 of 36) and 88% (36 of 41), respectively. Low target results meeting the acceptance criteria for SQV were 27 of 28 (96%); for RTV, 36 of 38 (95%); for APV, 10 of 11 (rounds two and three only); and for LPV, 7 of 7 (round three only). One laboratory could not measure low concentrations for IDV (for two out of three rounds) or RTV (three of three rounds) because the target concentrations (samples from UCSD, with means of 161 and 164 ng/ml, respectively) were below their assay sensitivity. Results improved with increasing target concentrations: medium target results for IDV rose to 95% (mean concentration of 2,726 ng/ml) from 86% at the low end; from 88 (low target) to 98% for NFV (mean concentration of 2,492 ng/ml); and from 96 (low target) to 97% for SQV. Medium target results for RTV, APV, and LPV were 100% satisfactory. High target results for all drugs were 100% satisfactory. Individual laboratory results for all three rounds for each target concentration are illustrated in Fig. 1. Results are tabulated in Table 4 to show trends for each laboratory over all three rounds (mean absolute inaccuracies for all PIs). The trends show no improvement for the low target and slight improvements for medium and high targets over time (Fig. 2). Most values at the low concentrations are within 30% of target,
VOL. 48, 2004
CLINICAL MEASUREMENT OF ANTIRETROVIRALS
829
FIG. 3. Proficiency testing results by laboratory (x) and percent deviation from the assigned target value (y). Each graph represents an individual drug analyte. Symbols for each concentration level (low, medium, or high) and round (1, 2, or 3) are specifically indicated with each drug analyte’s graph. Dashed lines represent the ⫾20% acceptance range.
and only two values are above 30%. Details of causes of proficiency testing failure, intervention, and remedial action will be the subject of a future paper. Figure 3 details the mean percent deviation laboratory result for each PI. In addition, the figure further delineates the concentration and round by marker. It is clear from this figure which antiretrovirals were measured and reported more often as well as which drugs measured experienced more or less deviation from target.
ANOVA indicated that significant predictors of performance were the drug being measured (P ⫽ 0.001) and the concentration of that drug (P ⬍ 0.001). Neither the drug source in the laboratory providing the proficiency samples nor any of the interactions was significant (P ⬎ 0.2). After correction by Bonferroni’s method, the absolute inaccuracy for IDV was significantly greater than that for RTV (P ⫽ 0.01) or SQV (P ⫽ 0.01). In addition, the low concentration had significantly
830
HOLLAND ET AL.
greater absolute inaccuracy than did the medium (P ⬍ 0.001) or high (P ⬍ 0.001) concentrations. For round three, the blank EDTA plasma used by JHU had an interfering peak that coeluted with the internal standard for one participating laboratory, which meant that that laboratory could only measure NVP in samples from JHU. The same interfering peak coeluted with EFV for another laboratory preventing quantitation of that drug. DISCUSSION Overall the performance of the ACTG testing laboratories for these three proficiency testing rounds was excellent for NNRTIs and very good for the PIs (results 100 and 96% acceptable, respectively). Interlaboratory variability is small, and accuracy is high. The low PI spikes were more troublesome for Pharmacology Support Laboratories to measure than the medium- or high-spike samples. These observations confirm data reported by the International Interlaboratory Quality Control Program (1), which had a similar number of participating laboratories (nine) and a lower percentage of satisfactory results for the low-PI-spike samples. This trend was seen again in a further study by the international program involving 30 laboratories (7). In this present report, the difficulty of achieving good sensitivity for IDV and NFV may be a reflection of the low wavelengths used to quantitate these drugs, an average of 211 and 216 nm, respectively. Complete UV spectra of PIs show high absorbance in the 200- to 220-nm wavelength range (2), and these low wavelengths are commonly used to measure PIs (2), often resulting in decreased sensitivity and specificity because of high background interference at these low wavelengths. This is a good argument in favor of moving to quantitation using mass spectrometry, coupled with HPLC, which would give greater sensitivity and specificity and has the added advantage of using smaller sample volumes. Figures 1 and 3 reveal that some laboratories trend to a negative or positive bias with their results. This indicates a problem with assay calibrators and provides a guide to these laboratories to recheck drug stock values. Inclusion of a blank sample matrix provided an important control for the proficiency testing program, enabling laboratories to further check their methods for endogenous interferences. Laboratories experiencing no interferences from blank samples were assured of method specificity, while laboratories that found interferences in their methods were made aware that the interference was in the sample matrix and could add an error to the accuracy of their results. There was a large interfering peak present in the blank plasma used by JHU to prepare round three samples. To help prevent this from occurring in future proficiency testing rounds, blank plasma will be screened in more than one laboratory before use. ACTG protocols are beginning to require TDM, which is regarded as “clinical testing” by Food and Drug Administration regulations. ACTG laboratories are required to be CLIA certified (13) for TDM and are also required to take part in proficiency testing (17). There is no CLIA-approved proficiency testing program for antiretroviral drugs, but the ACTG proficiency testing program is an acceptable alternative. Laboratories are challenged twice a year (six spikes for each drug),
ANTIMICROB. AGENTS CHEMOTHER.
which meets CLIA regulations, and samples are prepared according to CLIA guidelines, i.e., samples mimic actual patient samples as closely as possible (17). Acceptance criteria (within 20% deviation of the target value) are considered acceptable by CLIA for quantitative toxicology testing (17) and are also used by the International Quality Control Program (1, 7). However, the New York State Department of Health TDM toxicology proficiency testing program considers 15% allowable error as giving a better idea of analytical performance (13). The availability of the proficiency testing samples also aided laboratories in new method development and validation. Over the three-round testing period, some laboratories expanded their methods to include more drugs and used stored proficiency testing samples of known concentrations during the processes of method development and validation to provide proof that the new method was accurate and specific. It has long been recognized that the conducting of multicenter clinical trials that have pharmacokinetic aspects to their design requires that an infrastructure be integrated into the clinical trials network to assure sample integrity by proper collection, storage, and shipping and to insure data integrity by eliminating or reducing interlaboratory antiretroviral assay variability. The international interlaboratory quality control program for measurement of antiretroviral drugs in plasma has described efforts aimed at this type of program (1, 7). The National AIDS Clinical Trials Group Quality Assurance Program’s antiretroviral proficiency testing program provides important documentation that a network of pharmacology laboratories that can meet the clinical trial needs of ongoing HIV research protocols is in place. As the ACTG programs expands to international involvement, its pharmacology proficiency testing program will also include international pharmacology laboratories. ACKNOWLEDGMENTS Adult and Pediatric Pharmacology Support Laboratories were the following: Johns Hopkins University, Charles Flexner, PI, Fayez Hamzeh, Jin Lee; Stanford University, Terrence Blaschke and Patricia Burton; St. Jude’s Hospital, John Rodman and Brian Robbins; University of California, San Francisco, Fran Aweeka, Judith Stone, and Anura Jaywardene; University of Alabama, Edward Acosta and Michele Turner; University at Buffalo, Gene Morse, Robin DiFrancesco, and Kim Keil; University of Colorado, Courtney Fletcher and Lane Bushman; University of California, San Diego, James Connor, Diane Holland, and Rowena Espina-Quinto. The support of Alan Forrest in conducting and analyzing the ANOVA and the assistance of Melissa von Kerczek with the administration of the quality assurance program are appreciated. The work of the AIDS Clinical Trials Group Quality AssuranceQuality Control Program is supported by the following grants: University of Buffalo, National Institute of Allergy and Infectious Diseases, Adult Aids Clinical Trials Group Pharmacology Support Laboratory grant SSS200PC006 and Shared Instrumentation Grant no. S10RRR14572 from the National Center for Research Resources, National Institutes of Health; UCSF; UCSD SSS79455A; and JHU AACTG-L7-202QA. REFERENCES 1. Aarnoutse, R. E., C. P. W. G. M. Verweij-van Wissen, E. W. J. V. Kolmer, et al. 2002. International interlaboratory quality control program for measurement of antiretroviral drugs in plasma. Antimicrob. Agents Chemother. 46:884–886. 2. Aarnoutse, R. E., C. P. W. G. M. Verweij-van Wissen, W. J. M. Underberg, J. Kleinnijenhuis, Y. A. Hekster, and D. M. Burger. 2001. High-performance
VOL. 48, 2004
3. 4. 5.
6. 7.
8.
9. 10.
liquid chromatography of HIV protease inhibitors in human biological matrices. J. Chromatogr. B 764:363–384. Acosta, E. P., and J. G. Gerber. 2002. Position paper on therapeutic drug monitoring of antiretroviral agents. AIDS Res. Hum. Retrovir. 18:825–834. Barry, M., F. Mulcahy, C. Merry, S. Gibbons, and D. Back. 1999. Pharmacokinetics and potential interactions amongst antiretroviral agents used to treat patients with HIV infection. Clin. Pharmacokinet. 36:289–304. Burger, D. M., R. M. Hoetelmans, P. W. Hugen, J. W. Mulder, P. L. Meenhorst, P. P. Koopmans, K. Brinkman, M. Keuter, W. Dolmans, and Y. A. Hekster. 1998. Low plasma concentrations of indinavir are related to virological treatment failure in HIV-1-infected patients on indinavir-containing triple therapy. Antivir. Ther. 3:215–220. DeRemer, M., R. D’Ambrosio, L. Bartos, S. Cousins, and G. D. Morse. 1997. Radioimmunoassay of zidovudine: extended use and potential application. Ther. Drug Monit. 14:195–200. Droste, J. A. H., R. E. Aarnoutse, P. P. Koopmans, Y. A. Hekster, and D. M. Burger. 2003. Evaluation of antiretroviral drug measurements by an interlaboratory quality control program. J. Acquir. Immune Defic. Syndr. 32:287– 291. Durant, J., P. Clevenbergh, R. Garraffo, P. Halfon, S. Icard, P. Del Giudice, N. Montagne, J. Schapiro, and P. Dellamonica. 2000. Importance of protease inhibitor plasma levels in HIV-infected patients treated with genotypicguided therapy: pharmacological data from the Viradapt Study. AIDS 14: 1333–1339. Frerichs, V. A., R. DiFrancesco, and G. D. Morse. 2003. Determination of protease inhibitors using liquid chromatography-tandem mass spectrometry. J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 787:393–403. Gunawan, S., M. P. Griswold, and D. G. Kahn. 2001. Liquid chromatographic-tandem mass spectrometric determination of amprenavir (agenerase) in serum/plasma of human immunodeficiency virus type-1 infected patients receiving combination antiretroviral therapy. J. Chromatogr. A 914:1–4.
CLINICAL MEASUREMENT OF ANTIRETROVIRALS
831
11. Jayewardene, A. L., B. Kearney, J. A. Stone, J. G. Gambertoglio, and F. T. Aweeka. 2001. An LC-MS-MS method for the determination of indinavir, an HIV-1 protease inhibitor, in human plasma. J. Pharm. Biomed. Anal. 25: 309–317. 12. Kakuda, T., L. Page, P. Anderson, K. Henry, T. Schacker, F. Rhame, E. Acosta, R. Brundage, and C. Fletcher. 2001. Pharmacological basis for concentration-controlled therapy with zidovudine, lamivudine, and indinavir. Antimicrob. Agents Chemother. 45:236–242. 13. Jenny, R. W., and K. Y. Jackson. 1992. Evaluation of the rigor and appropriateness of CLIA ’88 toxicology proficiency testing standards. Clin. Chem. 38:496–500. 14. Marzolini, C., A. Telenti, G. Decosterd, J. Biollaz, and T. Buclin. 2001. Efavirenz plasma levels can predict treatment failure and central nervous system side effects in HIV-1 infected patients. AIDS 15:71–75. 15. Morse, G. D. 2000. Drug interactions with antiretrovirals. Curr. Infect. Dis. Rep. 2:257–266. 16. Pereira, A. S., K. B. Kenney, M. S. Cohen, J. E. Hall, J. J. Eron, R. R. Tidwell, and J. A. Dunn. 2000. Simultaneous determination of lamivudine and zidovudine concentrations in human seminal plasma using high-performance liquid chromatography and tandem mass spectrometry. J. Chromatogr. B Biomed.Sci. Appl. 742:173–183. 17. U.S. Department of Health and Human Services. 2002. Code of federal regulations, Title 42, part 493: Laboratory requirements. U.S. Department of Health and Human Services, Washington, D.C. 18. Villani, P., M. Feroggio, L. Gianelli, A. Bartoli, M. Montagna, R. Maserati, and M. B. Regazzi. 2001. Antiretrovirals: simultaneous determination of five PIs and three nonnucleoside transcriptase inhibitors in human plasma by a rapid high-performance liquid chromatography-mass spectrometry assay. Ther. Drug Monit. 23:380–388.
