Abstract
The pharmacokinetics of raltegravir (RAL) in HIV patients is characterized by high interpatient/intrapatient variability. We investigated the potential contribution of the drug pharmaceutical formulation to RAL pharmacokinetics. We first compared in vivo the pharmacokinetics of RAL for 67 patients to whom the drug was administered by swallowing the intact tablet with those obtained from 13 HIV-infected patients who chewed the RAL tablet due to swallowing difficulties. Subsequently, we evaluated in vitro the dissolution of RAL tablets under different conditions. In the in vivo study, we found that patients given RAL by chewing the tablets presented pharmacokinetic profiles characterized by significantly higher RAL absorption than did patients receiving the drug by swallowing. The in vitro studies showed that when the whole tablets were exposed to an acidic medium, the release of RAL was very low, whereas when the tablets were crushed, the profiles presented significantly higher concentrations of RAL. Crushed tablets tested in water or in a pH 6.8 buffer exhibited prompt and complete dissolution of RAL. HIV-infected patients receiving RAL by chewing the tablet showed higher drug absorption and reduced pharmacokinetic variability compared with patients swallowing the intact tablet. This is related to problems in tablet disintegration and to erratic drug absorption. The amelioration of the RAL pharmaceutical formulation could improve drug pharmacokinetics.
INTRODUCTION
Raltegravir (RAL) is the first approved integrase inhibitor targeting the strand transfer step of HIV integration (14). Studies in HIV treatment-experienced as well as in naive patients have shown that RAL-containing regimens have potent antiretroviral activity (2). Moreover, the favorable side effect profile in comparison to all other antiretrovirals, as well as the minimal impact on lipid homeostasis, has made RAL a strong option for the treatment of an array of HIV-infected patients (2). Unexpected negative results from recent clinical trials in which RAL-based arms were associated with higher than expected virologic failure rates compared with protease inhibitor-based therapies, however, have raised some concerns about the clinical efficacy of RAL (7, 12, 15).
A clear relationship between clinical efficacy and RAL plasma concentrations has not been identified yet, and the pharmacokinetics of the drug appears to have less influence on treatment outcome than other covariates (e.g., baseline HIV RNA and use of other active agents in optimized background therapy) (2). Nevertheless, this concept has been challenged by recent findings. Indeed, a prospective study involving 106 HIV-infected patients experiencing treatment failure under RAL-containing regimens has shown that plasma RAL exposure significantly influences the drug antiviral activity and the eventual selection of resistance mutations (11). This has also been confirmed by the results of the QDMRK trial (7), showing that plasma RAL concentrations correlate significantly with the likelihood of virologic response in patients given the drug once daily. Despite the lack of establishment of a well-defined therapeutic range for RAL, these studies indicate that a correlation may exist between the individual RAL plasma concentrations and the clinical outcome for HIV-infected patients, suggesting potential relevance of therapeutic drug monitoring (TDM) in selected situations.
Limited studies in healthy volunteers and in HIV-infected patients have shown that RAL pharmacokinetics is characterized by high interpatient variability (3, 5, 11, 13). Recently, we have shown that the pharmacokinetics of RAL in HIV-infected subjects is also characterized by high intrapatient variability (4). Interestingly, in the previously published studies, it was shown that gender, race, age, body mass index, food intake, and renal or hepatic insufficiency explain only in part the observed variability in RAL pharmacokinetics (2). More recently, it was also shown that the presence of allelic variants in the UGT1A1 gene, which encodes the enzyme involved in the metabolism of the drug, only minimally affects plasma RAL concentrations (1, 17). Therefore, at present, the majority of RAL pharmacokinetic variability remains unexplained by genetic and nongenetic factors.
In the present study, we investigated whether the observed wide inter- and intrapatient distribution of RAL plasma concentrations is related to the release of the drug from its pharmaceutical formulation.
MATERIALS AND METHODS
Patients.
All HIV-infected patients given RAL at 400 mg twice daily as part of maintenance highly active antiretroviral therapy (HAART) and undergoing TDM as routine clinical practice were included in the present study. HIV patients given concomitant administration of antacids and/or other drugs known to significantly influence the absorption of RAL were excluded from the present study.
Study design.
We first compared in vivo the pharmacokinetics of RAL from patients receiving the drug by swallowing the whole tablet with those obtained from HIV-infected patients who chewed the tablet due to swallowing difficulties (no specific contraindications for the administration of the drug by chewing are given in the RAL [Isentress] full prescribing information sheet).
Subsequently, we evaluated in vitro the dissolution of RAL tablets under different conditions, namely, in water, in acidic solution, and in pH 6.8 buffer. Dissolution tests were performed comparing whole RAL tablets (which mimics swallowing administration) with tablets crushed by grinding with a mortar and pestle (which mimics administration of the tablet by chewing).
In vivo pharmacokinetic evaluations.
The present study is based on a retrospective analysis of routine pharmacokinetic evaluations carried out as day-to-day clinical practice for the optimization of drug dosing in HIV-infected patients. For most antiretrovirals, TDM is usually performed by assessing the single plasma trough concentration (Ctrough). However, this approach is not feasible for RAL, which requires routine assessment of the predicted area under the concentration-time curve (AUC) (5). The assessment of RAL plasma concentrations was therefore carried out as routine TDM based on the collection of blood samples within the first 4 h after the morning RAL dose.
On the morning of the pharmacokinetic studies, blood samples were collected for the measurement of RAL plasma trough concentrations. Then, the patients took the morning RAL dose according to their practice in the presence of the nursing staff. Subsequently, additional blood samples were collected at 1, 2, 3, and 4 h after the morning drug dose. A maximum deviation from the scheduled sampling times of ±5 min was considered acceptable. A light standardized breakfast (coffee and semiskim milk with biscuits or rusks, often with orange juice) was served to all patients 90 to 120 min after drug intake. The patients had free access to water.
The palatability of the chewed RAL tablet was assessed on the day of the pharmacokinetic evaluation by the patients, using a 3-point scale (0, poor; 1, fair; 2, good) when responding to the following question: “How palatable do you feel the study medication is?” The safety of the chewed RAL tablet was assessed by the recording of adverse events.
The study was performed in accordance with the Declaration of Helsinki and in compliance with guidelines of good clinical practice.
Assessment of RAL plasma concentrations.
RAL concentrations in plasma samples were determined by a high-performance liquid chromatography (HPLC) method coupled with tandem mass spectrometry (HPLC–MS-MS) based on the assay originally developed by Fayet et al. (9). The method was validated in agreement with the consensus guidelines on bioanalytical method validation (10). Our laboratory participates in an external quality control program for the continuous monitoring of the method's performance (the Dutch KKGT). The method was linear over the RAL concentration ranges of 10 to 10,000 ng/ml. Between- and within-day imprecision and inaccuracy were less than 15%.
Drug release studies.
Dissolution tests were conducted according to USP 30 Apparatus 2 guidelines (paddle method; model AT7; Sotax, Basel, Switzerland) with 900 ml dissolution medium maintained at 37 ± 0.5°C and agitated at 50 rpm (n = 6). The media studied included purified water and pH 1.0 (0.1 N HCl) and pH 6.8 (50 mM phosphate buffer). At each sampling time, 1.5 ml of dissolution medium was withdrawn, filtered (0.22-μm syringe filter; Millipore), and stored in polyethylene test tubes at 4°C until the analysis. Crushed tablets were prepared by grinding with a mortar and pestle for 5 min, and the powders obtained were transferred into a dissolution vessel with wax paper.
Samples were analyzed for drug content using an HPLC system with a photodiode array detector (Model Alliance; Waters, Milford, MA) set at a wavelength of 260 nm. For the analysis, 50 μl of the samples was diluted with 450 μl of purified water in HPLC glass vials, vortexed for 30 s, and injected. Concentrations were confirmed by HPLC–MS-MS after appropriate dilution of the sample.
Statistical analyses.
Normal distribution of the continuous variables was confirmed by the Kolmogorov-Smirnov test (and eventually retested after logarithmic transformation). Nonnormally distributed variables were expressed as median (interquartile range[IQR]) and normally distributed variables as means ± standard deviations (SD).
Between-patient variability in the main RAL pharmacokinetic parameters was expressed as the coefficient of variation (CV%). The RAL AUC from 0 to 4 h (AUC0-4) was estimated using the trapezoidal rule. The RAL AUC0-12 was estimated using the following algorithm (5): AUC0-12 = 2.082 C0 + 0.821 C1 + 1.238 C2 − 0.210 C3 + 4.280 C4 + 783.1, where C is the concentration and the subscript number is the time at which the concentration is collected (for instance, C2 means the concentration measured 2 hours after drug intake).
Comparisons of the main RAL pharmacokinetic parameters between patients given the drug by swallowing or by chewing were performed using an unpaired t test or the corresponding nonparametric test as deemed appropriate according to the data distribution.
RESULTS
In vivo RAL pharmacokinetics.
A total of 106 RAL pharmacokinetics profiles from 0 to 4 h after the morning drug dose were collected from 80 HIV-1-infected patients. Ninety-three out of the 106 pharmacokinetics evaluations were assessed in 67 patients given the drug by swallowing the intact tablet. The remaining 13 pharmacokinetics evaluations were obtained from 13 HIV-infected patients who chewed the RAL tablet due to swallowing difficulties. Baseline demographic characteristics of the patients included in the present study are given in Table 1.
Table 1.
Demographic characteristics of HIV-1 patients given raltegravir by swallowing the intact tablet or by chewing the tablet due to swallowing difficulties
| Parametera | Value for raltegravir: |
|
|---|---|---|
| Chewed (n = 13) | Swallowed (n = 67) | |
| Male (%) | 62 | 68 |
| Caucasian (%) | 85 | 88 |
| Age (yr) | 49 ± 13 | 45 ± 9 |
| Wt (kg) | 69 ± 11 | 68 ± 15 |
| AST (IU/liter) | 39 ± 23 | 30 ± 9 |
| ALT (IU/liter) | 36 ± 30 | 38 ± 27 |
| Serum creatinine (mg/dl) | 1.0 ± 0.3 | 0.9 ± 0.2 |
| CD4 cell count (no./μl) | 599 ± 447 | 580 ± 424 |
| Concomitant HAART (%) | ||
| Protease inhibitor based | 46 | 64 |
| Tenofovir based | 38 | 22 |
| Other | 16 | 14 |
AST, aspartate aminotransferase: ALT, alanine transaminase.
As shown in Fig. 1, patients taking RAL by chewing presented regular pharmacokinetic profiles characterized by a single sharp drug peak and significantly higher RAL absorption—assessed by RAL Cmax (maximum concentration of drug in serum) and AUC0-4 (Table 1)—compared with patients taking the drug by swallowing the intact tablet (RAL Cmax, 5,404 ± 3,032 versus 3,128 ± 2,588 ng/ml, P = 0.004; RAL AUC0-4, 11,634 ± 7,288 versus 7,007 ± 5,803 ng · h/ml, P = 0.011). The higher RAL absorption also determined a nonsignificant trend for higher daily drug exposure in patients taking the drug by chewing versus by swallowing (RAL AUC0-12, 15,024 ± 9,693 versus 11,803 ± 9,544 ng · h/ml; P = 0.084).
Fig 1.
Raltegravir time-concentration profiles in 80 HIV patients given the drug by swallowing the whole tablet (n = 67) or by chewing the tablet before swallowing (n = 13). The error bars indicate SD.
Large interpatient variability in the main RAL pharmacokinetic parameters was observed in both groups of patients (Table 2). However, with the exception of RAL trough concentrations, patients taking RAL by chewing had a 30 to 60% reduction in the coefficient of variation associated with the main RAL pharmacokinetic parameters (Table 2).
Table 2.
Main raltegravir pharmacokinetic parameters measured in HIV-1 patients to whom the drug was administered by swallowing and in those who chewed the RAL tablet due to swallowing difficulties
| Pharmacokinetic parameter | Value |
P valuec | |||
|---|---|---|---|---|---|
| Raltegravir chewed (n = 13) | CV% | Raltegravir swallowed (n = 93) | CV% | ||
| Ctrough (ng/ml) | 350 ± 414 | 118 | 544 ± 655 | 120 | |
| C1 (ng/ml) | 5,397 ± 3,043 | 56 | 2,110 ± 2,132 | 101 | <0.0001 |
| C2 (ng/ml) | 3,480 ± 2,415 | 69 | 2,256 ± 2,342 | 104 | 0.044 |
| C3 (ng/ml) | 1,975 ± 1,490 | 75 | 1,872 ± 1,858 | 99 | |
| C4 (ng/ml) | 1,212 ± 1,004 | 82 | 1,360 ± 1,537 | 113 | |
| Cmax (ng/ml) | 5,404 ± 3,032 | 56 | 3,128 ± 2,588 | 83 | 0.004 |
| Tmax (min)a | 60 (60–120)b | 25 | 180 (0–240)b | 63 | 0.028 |
| AUC0-4 (ng · h/ml) | 11,634 ± 7,288 | 62 | 7,007 ± 5,803 | 83 | 0.011 |
| AUC0-12 (ng · h/ml) | 15,024 ± 9,639 | 64 | 11,803 ± 9,544 | 83 | 0.084 |
Tmax, time to maximum concentration of drug in serum.
Median (interquartile range).
The P value refers to the comparison of the raltegravir pharmacokinetic parameters between the two groups.
Most of the patients taking RAL by chewing found the palatability fair (n = 12), while the remainder rated palatability as poor (n = 1). No adverse events were reported during the study.
In vitro studies.
The in vitro studies showed that the intact tablets presented relatively slow release profiles, probably due to a lack of disintegration compared with crushed tablets in all the media considered (Fig. 2A to C). In particular, for intact tablets tested in an acidic medium, the RAL concentrations were very low, reaching less than 10% of the dose after 2 h (Fig. 2B), owing to the well-known poor solubility of RAL at low pH (RAL [Isentress] product monograph; Merck). Crushed tablets tested in water and in a pH 6.8 buffer exhibited prompt and complete dissolution of RAL, with the total amount of the drug dissolved in 15 min (Fig. 2A and C). When crushed tablets were tested in acid, the powder tended to float on the fluid at the beginning of the test. After a few seconds however, the powder was totally wetted, and a white dispersion formed in the dissolution vessel. The dissolution profiles obtained from crushed tablets in the acidic medium presented concentrations of RAL higher than those from intact tablets, with an evident supersaturation peak possibly explained by RAL potassium salt, which almost immediately recrystallized as free acid.
Fig 2.
In vitro dissolution profiles of whole tablets versus crushed tablets of raltegravir in water (A), at pH 1 (B), and at pH 6.8 (C).
DISCUSSION
The present study documents for the first time that HIV-infected patients taking RAL by chewing show higher drug absorption and less interpatient pharmacokinetics variability than patients taking the drug by swallowing. These results may explain the previously observed large inter- and intrapatient variability in the daily concentrations of RAL (3–5, 11, 13), whose pharmacokinetics is not influenced significantly by genetic and nongenetic characteristics of patients (1, 2, 17). Our results point to a significant role of the drug pharmaceutical formulation per se in determining the variable and irregular pharmacokinetics of RAL. This conclusion is sustained, not only by the pharmacokinetic profile of RAL release in vivo, but also by in vitro analyses.
The in vitro dissolution of RAL tablets was tested under different conditions (pH 1 and pH 6.8 buffer and water) and by comparing the dissolution of intact RAL tablets (which mimics drug administration by swallowing) with tablets crushed by grinding in a mortar and pestle (which mimics drug administration by chewing the tablet). Using this approach, we documented that in vitro, the dissolution of RAL from whole tablets is very low under acid conditions, while it improved in a neutral environment (water and pH 6.8 buffer). Conversely, crushed tablets tested in water and pH 6.8 buffer exhibited prompt and complete dissolution of RAL, with the total amount of the drug dissolved in 15 min. Also, when tested under acid conditions, crushed tablets showed higher concentrations of RAL than whole tablets.
Our in vivo and in vitro findings suggest that tablets administered to patients by swallowing do not fully disintegrate in the stomach, leading to negligible drug release from the tablets for the entire duration of gastric residence. In order to release the active ingredients, tablets need to reach the duodenum, where RAL presumably can find an environment (neutral pH) that can facilitate its dissolution. The gastric residence times of the nondisintegrating dosage form with dimensions greater than 10 mm are extremely variable, mainly linked to the presence or absence of food and to the type of food ingested (6). In the case of an empty stomach (fasting state), the transit time of pharmaceutical dosage forms up to 12 to 14 mm in size is relatively rapid and depends on the recurrence of strong contractions of the gastric muscular tissues approximately every 2 h (also known as housekeeper waves), which transfer the stomach contents to the duodenum. In a fed state, the gastric residence time of relatively big solid dosage forms is linked to the duration of the digestive phase, which in turn depends on the amount and type of food ingested. In contrast, solutions or suspensions of solids and small units (<4 to 7 mm) can freely pass the pylorus, which is partially open even during the digestion process, and are thus less influenced by the variability in gastric emptying times. Thus, the lack of disintegration of RAL tablets and the relatively low solubility of the active ingredient in an acidic environment, which are highly affected by the gastric residence times, are consistent with the lower absorption of RAL observed in patients who swallowed the intact tablets than in patients who chewed the tablets. Besides the higher reproducibility of the stomach transit, dispersion obtained by chewing the tablet can also promote the dissolution of the active ingredient in acidic medium by exploiting the supersaturation of RAL potassium salt in vivo and consequently can enhance RAL absorption.
The present study was not designed and powered to assess whether the observed improvement in RAL pharmacokinetics eventually impacts patient outcome. The potential contribution of RAL pharmacokinetics to the HIV virologic response is still a matter of debate. Indeed, phase II/III trials have shown that RAL plasma concentrations and drug exposure were correlated with the efficacy outcome, whereas no relevant clinical associations were found with the RAL Ctrough (11, 13, 16). The latter finding was not unexpected, given the failure of the RAL Ctrough to reliably predict daily drug exposure, expressed as RAL AUC0-12 (5). In the present study, we have shown that patients taking RAL by swallowing whole tablets experienced a wide distribution in the main drug pharmacokinetic parameters. However, this variability is greatly reduced in patients who chewed the RAL tablets. Accordingly, it can be reasonably speculated that the administration of RAL by chewing the tablet—by providing higher RAL absorption with reduced variability—may improve the way the drug is used in clinical practice. The possibility that the pharmacokinetics of RAL could be further improved by dissolving the pulverized tablets in water before taking the drug is an intriguing hypothesis that needs to be investigated.
The use of RAL once daily has been considered an attractive option to promote patient adherence to the therapeutic regimen. However, a recent phase III trial involving antiretroviral-naive patients treated with RAL at 800 mg once daily or 400 mg twice daily, both in combination with tenofovir/emtricitabine, has shown that RAL once daily was associated with significantly lower virological response than twice-daily administration (8). Nevertheless, considering that (i) inhibition by RAL of the integrase enzyme is irreversible (2), (ii) administration of RAL by chewing the tablet produces significantly higher drug absorption, and (iii) more drug available for binding to the integrase may result in a greater extent of target enzyme inhibition, the possibility that the administration of RAL at 800 mg/once daily by chewing the tablet may improve the virologic response compared with RAL at 800 mg/once daily by swallowing the whole tablet and may allow better patient compliance than RAL at 400 mg twice daily cannot be excluded. In our opinion, this is an intriguing hypothesis that needs further investigation.
In conclusion, we have shown that HIV-infected patients taking RAL by chewing the tablets have higher drug absorption and lower drug intersubject pharmacokinetic variability than patients taking the drug by swallowing whole tablets. We also show that this may depend on issues with appropriate tablet disintegration in the gastric environment, leading to erratic drug release. Improvement of the drug pharmaceutical formulation could potentially increase the response of HIV-infected patients to RAL-based regimens.
ACKNOWLEDGMENTS
D.C. has received educational grants from Merck Sharp & Dohme (MSD) and from Janssen-Cilag. C.G. has received educational grants from Abbott, Boehringer Ingelheim, Bristol Myers Squibb, MSD, and Janssen-Cilag. M.G. and G.R. have received educational grants from Abbott, Boehringer Ingelheim, Bristol Myers Squibb, Gilead, MSD, VIIV, and Janssen-Cilag. E.C. has received educational grants from Zambon Italia and Nicox.
No specific funding was received for this study. The study was carried out as part of our routine work.
Footnotes
Published ahead of print 10 September 2012
REFERENCES
- 1. Arab-Alameddine M, et al. 2012. Population pharmacokinetic analysis and pharmacogenetics of raltegravir In HIV-positive and healthy individuals. Antimicrob. Agents Chemother. 56:2959–2966 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Brainard DM, Wenning LA, Stone JA, Wagner JA, Iwamoto M. 2011. Clinical pharmacology profile of raltegravir, an HIV-1 integrase strand transfer inhibitor. J. Clin. Pharmacol. 51:1376–1402 [DOI] [PubMed] [Google Scholar]
- 3. Carey D, et al. 2012. A randomized study of pharmacokinetics, efficacy and safety of two raltegravir plus atazanavir strategies in ART-treated adults. J. Acquir. Immune Defic. Syndr. 60:143–149 [DOI] [PubMed] [Google Scholar]
- 4. Cattaneo D, et al. 2012. Inter- and intra-patient variability of raltegravir pharmacokinetics in HIV-1-infected subjects. J. Antimicrob. Chemother. 67:460–464 [DOI] [PubMed] [Google Scholar]
- 5. Cattaneo D, et al. 2012. Limited sampling strategies for the estimation of raltegravir daily exposure in HIV-Infected patients. J. Clin. Pharmacol. 52:440–445 [DOI] [PubMed] [Google Scholar]
- 6. Davis SS. 1985. The design and evaluation of controlled release systems for the gastrointestinal tract. J. Control Release 2:27–38 [Google Scholar]
- 7. Eron JJ, et al. 2010. Switch to a raltegravir-based regimen versus continuation of a lopinavir-ritonavir-based regimen in stable HIV-infected patients with suppressed viraemia (SWITCHMRK 1 and 2): two multicentre, double-blind, randomised controlled trials. Lancet 375:396–407 [DOI] [PubMed] [Google Scholar]
- 8. Eron JJ, Jr, et al. 2011. Raltegravir once daily or twice daily in previously untreated patients with HIV-1: a randomised, active-controlled, phase 3 non-inferiority trial. Lancet Infect. Dis. 11:907–915 [DOI] [PubMed] [Google Scholar]
- 9. Fayet A, et al. 2009. A LC-tandem MS assay for the simultaneous measurement of new antiretroviral agents: raltegravir, maraviroc, darunavir, and etravirine. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 877:1057–1069 [DOI] [PubMed] [Google Scholar]
- 10. FDA 5 July 2011. access date FDA guidance for industry 2001—bioanalytical method validation. http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm070107.pdf
- 11. Garrido C, et al. 2012. Plasma raltegravir exposure influences the antiviral activity and selection of resistance mutations. AIDS Res. Hum. Retroviruses 28:156–164 [DOI] [PubMed] [Google Scholar]
- 12. Kozal MJ, et al. 2010. The SPARTAN Study: a pilot study to assess the safety and efficacy of an investigational NRTI and RTV-sparing regimen of atazanavir experimental dose of 300 mg bid plus raltegravir 400 mg bid in treatment naïve HIV-infected patients, abstr THL BB204, p 286 Abstr. 18th Int. AIDS Conf., Vienna, Austria International AIDS Society, Geneva, Switzerland [Google Scholar]
- 13. Rizk ML, et al. 2012. Pharmacokinetics and pharmacodynamics of once-daily versus twice-daily raltegravir in treatment-naive HIV-infected patients. Antimicrob. Agents Chemother. 56:3101–3106 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Steigbigel RT, et al. 2008. Raltegravir with optimized background therapy for resistant HIV-1 infection. N. Engl. J. Med. 359:339–354 [DOI] [PubMed] [Google Scholar]
- 15. Taiwo B, et al. 2011. Efficacy of a nucleoside-sparing regimen of darunavir/ritonavir plus raltegravir in treatment-naïve HIV-1-infected patients (ACTG A5262). AIDS 25:2113–2122 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Wenning L, et al. 2008. Pharmacokinetic/pharmacodynamic analyses for raltegravir in phase II and III studies in treatment experienced HIV-infected patients, abstr. OA25, p. 56 Abstr. 9th Int. Workshop Clin. Pharmacol. HIV Ther., New Orleans, LA Virology Education, Utrecht, The Netherlands [Google Scholar]
- 17. Wenning LA, et al. 2009. Pharmacokinetics of raltegravir in individuals with UGT1A1 polymorphisms. Clin. Pharmacol. Ther. 85:623–627 [DOI] [PubMed] [Google Scholar]