Pharm Res (2017) 34:619–628

DOI 10.1007/s11095-016-2090-2

RESEARCH PAPER

Meal Effects Confound Attempts to Counteract

Rabeprazole-Induced Hypochlorhydria Decreases in Atazanavir
Absorption
Kathleen Panter Faber 1 & Hsin-Fang Wu 1 & Marc R. Yago 1 & Xiaohui Xu 2 & Pathanjali Kadiyala 2 &
Lynda A. Frassetto 3,4 & Leslie Z. Benet 1

Received: 16 September 2016 / Accepted: 19 December 2016 / Published online: 27 December 2016
# Springer Science+Business Media New York 2016

ABSTRACT mitigate the loss of ATV exposure observed during RAB-

Purpose Clinically relevant pharmacokinetic interactions ex- induced hypochlorhydria. Meal effects lead to a marked dif-
ist between gastric acid-reducing agents and certain weakly ference in the outcome of betaine HCl on atazanavir exposure
basic drugs that rely on acidic environments for optimal oral than we previously reported for dasatanib under fasting
absorption. In this study, we examine whether the administra- conditions.
tion of betaine hydrochloride under fed conditions can en-
hance the absorption of atazanavir, an HIV-1 protease inhib- KEY WORDS absorption . atazanavir . food effects . PPIs .
itor, during pharmacologically-induced hypochlorhydria. weakly basic drugs
Methods In this randomized, single-dose, 3 period, crossover
study healthy volunteers received ritonavir-boosted atazanavir
(atazanavir/ritonavir 300/100 mg) alone, following pretreat- ABBREVIATIONS
ment with the proton pump inhibitor rabeprazole (20 mg ARA Acid reducing agents
twice daily), and with 1500 mg of betaine HCl after ART Antiretroviral therapy
rabeprazole pretreatment. Atazanavir was administered with BDDCS Biopharmaceutics drug disposition classification
a light meal and gastric pH was monitored using the system
Heidelberg Capsule. BHCl Betaine hydrochloride
Results Pretreatment with rabeprazole resulted in significant BMI Body mass index
reductions in atazanavir C max (p < 0.01) and AUC 0-last HIV Human immune deficiency virus
(p < 0.001) (71 and 70%, respectively), and modest decreases LC/MS/ Liquid chromatography/tandem mass
in ritonavir Cmax and AUClast (p < 0.01) (40% and 41%, re- MS spectrometry
spectively). The addition of betaine HCl restored 13% of LLOQ Lower limit of quantitation
ATV Cmax and 12% of AUClast lost due to rabeprazole. PI Protease inhibitor
Conclusions The co-administration of rabeprazole with PPI Proton pump inhibitor
atazanavir resulted in significant decreases in atazanavir ex- RAB Rabeprazole
posure. The addition of betaine HCl did not sufficiently

* Leslie Z. Benet
INTRODUCTION
leslie.benet@ucsf.edu
A significant body of clinical data supports the use of human
1
Department of Bioengineering and Therapeutic Sciences, University of immunodeficiency virus (HIV) oral protease inhibitors (PIs) as
California San Francisco, 533 Parnassus Ave., Room U-68, San integral components in combination antiretroviral therapy
Francisco, CA 94143-0912, USA (ART; (1)). Since the approval of the first PI in 1995, advances
2
Bioanalytical Sciences, Bristol-Myers Squibb, Princeton, New Jersey, USA in dosing, tolerability and efficacy of drugs in the class, have
3
Department of Medicine University of California San Francisco, San contributed to their widespread use. Despite their overall ef-
Francisco, California, USA fectiveness, metabolic side effects and clinically relevant drug-
4
Clinical Research Center, University of California San Francisco, San drug-interactions continue to be limitations of use (1–3).
Francisco, California, USA Interactions arising from the concomitant use of gastric acid
620 Faber et al.

reducing agents (ARAs) with certain PIs have drawn consid- was co-dosed with a light meal. A lower dose of omeprazole
erable interest due to the potential for marked reductions in PI (20 mg once daily) resulted in more modest reductions in
exposure and compromised in vivo activity. boosted atazanavir exposures (39–46%; (12)). As such, the
Atazanavir (Reyataz®) is a potent and selective PI indicat- current FDA label for atazanavir allows for the use of low
ed for use in combination with other antiretrovirals for the doses of PPIs (comparable to omeprazole 20 mg once daily)
treatment of HIV-1 infection in adults and pediatric patients in treatment-naïve patients using a time-staggered approach
over 3 months of age. In adults, atazanavir is administered (administration of the two agents is separated by 12 h). The
once daily with food as a 300 mg tablet taken with 100 mg use of PPIs in treatment-experienced patients and/or those
of ritonavir (atazanavir/ritonavir 300/100 mg) or, more re- taking unboosted regimens is not recommended (4).
cently, as a fixed-dose combination with 150 mg of cobicistat Mitigation strategies targeting transient gastric
(Evotaz®; (4,5)). Coadministration of atazanavir with ritona- reacidification with oral acidic solutions to improve the extent
vir or cobicistat (potent CYP3A4 and P-glycoprotein (P-gp) of absorption of weakly basic drugs with pH-dependent solu-
inhibitors), exploits a favorable drug-drug-interaction, effec- bility have been previously reported. The systemic exposure of
tively enhancing the systemic exposure of atazanavir (6). the antifungal agent posaconazole was increased >70% when
Alternatively in treatment-naïve patients, atazanavir may be administered with Coca-Cola (pH = 2.5), compared to ad-
given at the 400 mg dose strength without a pharmacokinetic ministration with water (16). Ray et al. reported that
boosting agent. atazanavir concentrations 3 h post-dose increased in four of
Similar to the other drugs in this class, atazanavir is a weak six HIV patients following administration of 400 mg
base that exhibits pH-dependent aqueous solubility over the atazanavir with cola, relative to water (17). Acidic solutions
physiologic pH range. On the basis of its poor solubility and have also been investigated as a means to overcome exposure
extensive metabolism, it is classified as a Biopharmaceutical loss during PPI-induced hypochlorhydria (fasting gastric pH
Drug Disposition Classification System (BDDCS) class 2 drug >4). Studies investigating the use of Coca-Cola as a means to
(7,8). The pharmacokinetics of atazanavir is complex, highly increase the extent of atazanavir, posaconazle and ketocona-
variable, and its absorption is altered by the presence of food, zole absorption during PPI use have been reported with vary-
gastric ARAs, and substrates and/or inhibitors of drug metab- ing success. In the case of posaconazole and ketoconazole,
olizing enzymes and transporters (8–12). The in vitro solubility administration with cola improved the systemic absorption
profile of atazanavir and available clinical data suggest that of both agents in healthy volunteers, but was unable to
elevations in gastric pH by ARAs can alter drug absorption completely compensate for the loss of exposure during PPI-
resulting in lower systemic exposures. induced hypochlorhydria. Administration of boosted and
Interactions between ARAs and weakly basic drugs, partic- unboosted atazanavir with cola and a light meal to healthy
ularly in the therapeutic areas of HIV and oncology, have volunteers following pretreatment with omeprazole had no
been widely studied due to the prevalence of ARA use in these appreciable impact on atazanavir exposure (9,10,16,18).
populations, as well as the potential for undesirable conse- Another mitigation strategy involves the use of solid
quences (3,13–15). The clinical relevance of such an interac- reacidifying agents, such as betaine hydrochloride (BHCl).
tion depends on the level of impact to drug absorption as well BHCl is a nutraceutical available over-the-counter and is
as the therapeutic index of the victim drug. ARA class, dose commonly used as a digestive aid. When taken orally, BHCl
and time of administration relative to the victim drug can all rapidly dissociates into free betaine and hydrochloric acid.
alter the magnitude of the interaction. PPIs are the most po- Recently, our laboratory demonstrated that BHCl can rapidly
tent ARA class and serve as a Bworst-case-scenario^ in the and transiently reacidfy gastric pH in healthy volunteers
evaluation of an interaction as they exert their pharmacologic pretreated with the PPI rabeprazole (19). Furthermore, we
effect through irreversible binding to active gastric proton have also shown that a single 1500 mg oral dose of BHCl
pumps (H+/K+ -ATPase) at the secretory surface of gastric was sufficient to mitigate the reduced exposure of dasatinib,
parietal cells. another weakly basic drug with pH-dependent solubility, in
Previous pharmacokinetic analyses with atazanavir have healthy subjects pretreated with rabeprazole (20). Based on
shown that the concurrent use of high doses of PPIs result in these data, we hypothesize that the concomitant use of
substantial reductions to atazanavir absorption. In the absence BHCl with atazanavir may be a more effective mitigation
of a PK boosting agent, atazanavir exposures were reduced > strategy compared to Coca-Cola since in the same volume
90% in healthy volunteers pretreated with high doses of the of water (250 mL), BHCl has a greater buffering capacity
PPI omeprazole (40 mg once daily), compared to atazanavir and provides a higher equivalent of acid (in the form of H+
alone (400 mg once daily; (10)). Exposure reductions of 72- ions).
76% were observed when the same dose of omeprazole was Here we examine whether BHCl can mitigate marked re-
coadministered with ritonavir-boosted atazanavir, compared ductions in atazanavir exposure during rabeprazole-induced
to boosted-atazanavir alone (9). In both examples, atazanavir hypochlorhydria under fed conditions.
Meal Effects Confound Attempts to Counteract Hypochlorhydria 621

METHODS day until 3 h post-atazanavir dosing. Ten minutes after the light
meal participants first received treatment A, which was a single
A total of eight healthy, non-smoking volunteers between the oral dose of atazanavir/ritonavir 300/100 mg with 250 mL of
protocol allowed ages of 18–65 were enrolled in this three- water. For treatments B and C, participants were pretreated
period crossover study. Subject demographics are summarized with rabeprazole 20 mg twice daily with food for three days.
in Table I. The mean age and body mass index (BMI) ± SD On pharmacokinetic sampling days, an additional 20 mg dose
were 37 ± 16 years and 24 ± 2.5 kg/m2, respectively. Eligibility of rabeprazole was administered with four ounces of low fat
was determined by medical history review, physical examination, yogurt at least 2 h prior to atazanavir. When gastric pH
12-lead electrocardiogram and clinical laboratory evaluations. remained above pH 3.8 for at least 10 min, subjects received
Baseline gastric pH measurements were determined during their light meal and then 10 min later either atazanavir/
screening visits through the use of the Heidelberg pH ritonavir 300/100 mg alone (Treatment B), or 1500 mg of
Diagnostic System (Heidelberg Medical Inc., Mineral Bluff, BHCl followed by atazanavir/ritonavir 300/100 mg 5 min after
GA) to confirm normochlorhydria (fasting gastric pH <4), as BHCl administration (Treatment C; Fig. 1). Participants were
previously described (19). Female subjects who were surgically block randomized (4 subjects per block) for the order in which
sterile or post-menopausal (no history of menorrhea in the past they would receive treatments B and C. Treatment periods were
12 months) were eligible to participate. Individuals with a history separated by at least a seven-day washout.
of gastrointestinal disease and those taking concomitant medica- Venous blood samples for PK assessment were collected at 0,
tions 2 weeks prior to enrollment were excluded. 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, and 22 h, relative to atazanavir
Participants provided written informed consent to partici- dosing. An additional pre-rabeprazole sample was drawn during
pate in the study. The study was approved by the Committee treatments B and C. Blood samples were centrifuged within
on Human Research of the University of California, San 30 min of collection at 4°C. Plasma was subsequently separated
Francisco and registered on the US National Institutes of and stored in aliquots at −80°C until bioanalysis. Plasma con-
Health Clinical Trials Database (NCT01759875; centrations of atazanavir and ritonavir were measured using a
https://clinicaltrials.gov/ct2/show/NCT01759875). validated LC-MS/MS method (API 3000, MDS Sciex,
Thornhill, Ontario) by Tandem Labs, as described below.

Study Design Outcome Measures

This was an open-label, randomized, three-treatment, cross- The primary outcome measures for this study included the
over study conducted at the Clinical & Translational Science maximum plasma concentration (Cmax) and area under the
Institute’s Clinical Research Center at the University of plasma concentration-time curve from time zero to 22 h
California, San Francisco. All subjects fasted overnight prior to (AUClast) of atazanavir for each treatment. Secondary out-
receiving atazanavir in each treatment and consumed a stan- comes included ritonavir Cmax and AUC last, as well as
dardized light meal (336 kcal, 5.1 g fat, 9.3 g protein) 10 min atazanavir and ritonavir Tmax.
prior to each dose of atazanavir. Gastric pH was monitored
using the Heidelberg pH capsule on the morning of each study Pharmacokinetic Analysis

Table I Enrolled Subject Atazanavir and ritonavir pharmacokinetic parameters for

Demographics Total N 8
each period were calculated from plasma concentrations using
Male 7 non-compartmental analyses in Phoenix Winnonlin 5
Female 1 (Pharsight®, Sunnvale, CA). Cmax and Tmax were estimated
Race/Ethnicity from the observed data. Atazanavir and ritonavir AUCs were
Caucasian 4 calculated using the linear trapezoid method.
Asian 2
African American 1 Statistical Analysis
Hispanic 1
Age (years) A sample size of eight was calculated to detect a 40% change
Mean ± SD 37 ± 16 in atazanavir AUClast with 80% power and a two-sided
Range 22–59 α = 0.05, based on an observed standard deviation of the dif-
Body Mass Index ference of 75%. Statistical results were performed using
Mean ± SD 24 ± 2.5 GraphPad Prism 5.02 (La Jolla, CA). A repeated measures
Range 21–28 analysis of variance (ANOVA) with the Tukey’s test for mul-
tiple comparisons was used to determine statistical significance
622 Faber et al.

Fig. 1 Clinical study design. ATV atazanavir, RTV ritonavir, RAB rabeprazole, BHCl betaine hydrochloride, BID twice daily.

across all treatments for atazanavir and ritonavir pharmaco- One subject was excluded from pharmacokinetic analysis
kinetic parameters except Tmax. Logarithmic transformation because atazanavir concentration data suggested that
of all pharmacokinetic parameters (except Tmax) were per- plasma samples for treatments A and B were switched
formed prior to statistical analyses. Additionally, the geomet- prior to bioanalysis, but could not be confirmed. All study
ric mean ratios and 90% confidence intervals of atazanavir medications were well-tolerated and no adverse events
and ritonavir Cmax and AUClast were calculated for all treat- were reported due to study drugs or Heidelberg pH
ment comparisons. Diagnostic System use.

Quantitative Determination of Atazanavir Atazanavir Pharmacokinetics

and Ritonavir
Figure 2a shows the mean atazanavir plasma
Plasma samples were pretreated by a solid phase extraction concentration-time profile for each of the three treatment
procedure and analyzed by liquid chromatography-tandem periods. Following pretreatment with rabeprazole (treat-
mass spectrometry (LC-MS/MS) method (21). An API 3000 ment B), statistically significant reductions in atazanavir
was used to simultaneously detect atazanavir and ritonavir Cmax (p < 0.01) and AUClast (p < 0.001) of approximately
positive ions formed by Turboionspray™ ionization in the 70% were observed, compared to atazanavir/ritonavir
multiple reaction monitoring (MRM) mode. The lower limit 300/100 mg alone (treatment A; Table II). The addition
of quantitation (LLOQ) of the assay in human plasma were of BHCl to rabeprazole pretreated subjects during treat-
10.0 and 5.00 ng/mL for atazanavir and ritonavir, respective- ment C restored approximately 13% of atazanavir Cmax
ly, while the upper limit of quantitation were 10,000 and and 12% of atazanavir AUClast lost due to pretreatment
5000 ng/mL, respectively. The accuracy and precision of with rabeprazole, although the increases were not statisti-
the LC-MS/MS method for atazanavir and ritonavir in hu- cally significant. No significant differences in atazanavir
man plasma were determined by analyzing low, medium, high Tmax were observed across treatments.
and dilution quality control samples. The intra-assay precision To further quantify the effect of rabeprazole and betaine
for atazanavir was within 8.7% and the inter-assay precision HCl on atazanavir exposures on an individual basis, spaghetti
was within 6.5%. The mean % deviation was within ±10.0%. plots of individual atazanavir Cmax and AUClast values by
The intra-assay precision for ritonavir was within 9.1% and treatment were prepared (Fig. 3). Atazanavir Cmax and
the inter-assay precision was within 10.2%. The mean % de- AUClast were reduced in all subjects following pretreatment
viation was within ±3.8%. with rabeprazole (Treatment B), compared to the control
(Treatment A). Five of the seven subjects experienced greater
than a 55% decline in Cmax and AUClast following pretreat-
RESULTS ment with rabeprazole, with three experiencing greater than
80% declines in both parameters. The remaining two subjects
Study Demographics and Safety experienced more modest decreases (11–48%) in atazanavir
AUClast and Cmax. Interindividual differences in gastric pH
A total of eight subjects (seven males and one female) profiles did not appear to be correlated with Cmax within each
received study medication and completed all treatments. treatment.
Meal Effects Confound Attempts to Counteract Hypochlorhydria 623

Fig. 2 Mean (SD) atazanavir (a) and ritonavir (b) plasma concentration-time profiles plotted on logarithmic scale. ATV atazanavir, RTV ritonavir, RAB rabeprazole,
BHCl betaine hydrochloride.

Considerable overlap in individual atazanavir C max (90% CI, 26%–54%) and 40% (90% CI, 14%–76%), respec-
and AUClast parameters was observed between treatments tively (Table III). Similar to atazanavir, the addition of BHCl
B and C. With the addition of BHCl (Treatment C), three during rabeprazole-induced hypochlorhydria restored 12%
subjects experienced unexpected further declines in and 13% of ritonavir Cmax and AUClast, respectively, that
atazanavir exposure. Interestingly, two of these subjects was lost due to pretreatment with rabeprazole. Increases in
were also least sensitive to pretreatment with rabeprazole, exposure were not statistically significant.
as measured by their lower geometric mean ratios (B-to-
A) relative to the other subjects (Fig. 3). Conversely, the Intra-gastric pH Analysis
subject most sensitive to pretreatment with rabeprazole
also exhibited the largest gains in Cmax and AUClast of The mean gastric pH at the time of atazanavir dosing and
38% and 37%, respectively, with the addition of BHCl. immediately following a light meal (t = 0) did not differ signif-
Compared with treatment B (PPI), Cmax and AUC in- icantly across the three treatments: 2.76 ± 1.27, 2.36 ± 0.66
creased in 4 and 6 subjects, respectively, during and 2.89 ± 1.17 (mean ± SD) for treatments A, B and C, re-
Treatment C (PPI + BHCl). However, response to spectively. While there was considerable overlap in pH for all
BHCl, as measured by increases in atazanavir exposure, treatments throughout the measurement period, subjects
did not appear to correlate with responses to BHCl, as treated with BHCl were still able to achieve the lowest ob-
measured by reductions in gastric pH. served gastric pH in the study (0.838 ± 0.391), even exceeding
the lowest pH observed in the control group (1.39 ± 0.906).
Ritonavir Pharmacokinetics The time to achieve this pH after BHCl administration aver-
aged 67 ± 33 min, which suggests a delay in the onset of BHCl
Pretreatment with rabeprazole resulted in moderate decreases reacidification in the fed state, relative to the fasting state (20).
in mean ritonavir AUClast and Cmax (p < 0.01) values of 42% Moreover, the length of the reacidification period, as

Table II Atazanavir Pharmacokinetics with and Without Gastric pH Modulators

Geometric Mean Ratios (90% CIs)

PK Parameter ATV/RTV ATV/RTV + RAB ATV/RTV + RAB + B vs A C vs B C vs A

(Treatment A) (Treatment B) BHCl (Treatment C)

Cmax (ng/mL) 2920 ± 917 933 ± 545 1270 ± 524 0.291 (0.183–0.462) 1.44 (0.909–2.29) 0.420 (0.265–0.666)
AUC0-last 235,000 ± 6690 7410 ± 3580 10,100 ± 3790 0.300 (0.206–0.439) 1.41 (0.967–2.06) 0.424 (0.290–0.620)
(ng*hr/mL)
Tmax (hr) 3.0 (2.0–4.0) 3.0 (2.0–3.0) 3.0 ± (1.5–6.0) N/A N/A N/A

Pharmacokinetic parameter data presented as mean ± SD. Tmax is expressed as median (range)
CI confidence interval, PK pharmacokinetic, ATV atazanavir, RTV ritonavir, RAB rabeprazole, BHCl betaine hydrochloride, Cmax maximum plasma concentrationm,
Tmax time to maximum plasma concentration, AUC0-last area under the plasma concentration-time curve from time zero to 22 h
624 Faber et al.

Fig. 3 Individual atazanavir (ATV)

area under the concentration time
curve (AUC) and maximum
concentration (Cmax) values for
administration of ATV/RTV alone
(Treatment A), pretreatment with
rabeprazole (Treatment B) and the
addition of betaine HCl (Treatment
C).

measured by the time for gastric pH to rebound to pH >4, dasatinib. A closer examination of gastric pH-time profiles high-
averaged 76 ± 20 min. It should be noted however that the lights differences arising from the presence of food not previously
majority of enrolled subjects (n = 5 of the 8 subjects) did not observed in the fasted state, which may have contributed in part
achieve a gastric pH > 4 by 3 h post-ATV dosing. to the poor recovery. Before considering the impact of BHCl on
gastric pH and atazanavir bioavailability, it is important to un-
derstand the effect of rabeprazole.
DISCUSSION Reductions in atazanavir Cmax and AUC following pre-
treatment with rabeprazole in this study were comparable to
In this study, we confirmed previous findings that atazanavir those previously reported in healthy volunteers taking high
exposure is markedly reduced in the presence of high doses of doses of the PPI omeprazole (9). Interestingly, despite similar
PPIs, but BHCl was not able to reverse this affect. The interac- reductions in atazanavir exposure, gastric pH during pretreat-
tion between atazanavir and ARAs is believed to result from ment with omeprazole resulted in substantially higher gastric
reduced drug solubility in elevated gastric pH. Therefore, miti- pH around the time of atazanavir administration, compared
gation strategies aimed at transient gastric reacidification have to what we observed with rabeprazole (approximately pH 6
the potential to reverse this effect. Previously, our group has and pH 3, respectively). It is unclear why the differences in pH
shown that a single 1500 mg dose of BHCl was able to safely were observed as both PPIs, omeprazole and rabeprazole,
and significantly reverse the effect of rabeprazole-induced were administered for multiple days to reach maximum acid
hypochlorhydria on the absorption of dasatinib, a BCR-ABL suppression; however, one consideration is food. Subjects in
tyrosine kinase inhibitor, in healthy fasting volunteers (19). both studies completed a standardized low fat breakfast im-
Here, we investigated whether the use of BHCl could similarly mediately prior to receiving atazanavir. In healthy individuals
mitigate reductions in atazanavir exposure in healthy volunteers with normochlorhydria, the presence of food results in tran-
when administered after a light meal. Although atazanavir and sient elevations in gastric pH; however, the degree and dura-
dasatinib share similar physiochemical properties, recovery of tion of this effect is highly content-dependent. For example,
atazanavir exposure with BHCl was minimal, compared with meals taken with milk, which has a high buffering capacity,

Table III Ritonavir Pharmacokinetics with and Without Gastric pH Modulators

Geometric Mean Ratios (90% CIs)

PK Parameter ATV/RTV ATV/RTV + RAB ATV/RTV + RAB + B vs A C vs B C vs A

(Treatment A) (Treatment B) BHCl (Treatment C)

Cmax (ng/mL) 1210 ± 392 730 ± 254 860 ± 272 0.602 (0.241–0.861) 1.21 (0.849–1.74) 0.731 (0.511–1.05)
AUC0-last 7210 ± 1610 4280 ± 1180 5180 ± 1340 0.585 (0.460–0.743) 1.22 (0.960–1.55) 0.713 (0.562–0.905)
(ng*hr/mL)
Tmax (hr) 3.0 (1.0–8.0) 3.0 (2.0–4.0) 3.0 ± (1.5–6.0) N/A N/A N/A

Pharmacokinetic parameter data presented as mean ± SD. Tmax is expressed as median (range)
CI confidence interval, PK pharmacokinetic, ATV atazanavir, RTV ritonavir, RAB rabeprazole, BHCl betaine hydrochloride, Cmax maximum plasma concentration,
Tmax time to maximum plasma concentration, AUC0-last area under the plasma concentration-time curve from time zero to 22 h
Meal Effects Confound Attempts to Counteract Hypochlorhydria 625

result in higher postprandial gastric pH, compared to meals recordings below 4 in the hours following atazanavir dosing
taken with water, such as in this study (22,23). Therefore, it is (Fig. 4).
possible that differences in meal content contributed to differ- Although the pH effects on atazanavir absorption and sol-
ences in gastric pH. Despite the higher gastric pH noted with ubility have been well documented, drug metabolizing en-
omeprazole, the pH-time profiles for the control group (no zymes and transporters can also impact exposure and should
PPI) were similar between the two studies suggesting that the be considered here. Atazanavir is both a substrate and an
pH difference on PPI study days was real and unlikely due to inhibitor of CYP3A4 as well as a substrate of P-gp. Most
differences in data collection or analysis technique. PPIs are largely metabolized in the liver by the CYP2C19
In this study, a 3 h post-dose gastric pH observation and CYP3A4 isoforms, and as a result, can contribute to in-
window was selected as it was expected to be most critical teractions with drugs metabolized by the same CYPs (25).
period for atazanavir absorption based on a reported me- Rabeprazole, however, is unique in that it is metabolized pri-
dian Tmax of 2.7 h in healthy subjects (8). During the marily by non-P450-mediated mechanisms (26). The selection
post-dose monitoring period, gastric pH was slightly higher of rabeprazole in this study, therefore, was important for iso-
in treatment B (pretreatment with rabeprazole), compared lating the effect that pH plays on drug absorption by minimiz-
with the control (treatment A), although the difference ing the impact of CYP-related metabolism on study drug ex-
between the two treatments was smaller (Fig. 4) than what posure. Measuring systemic concentrations of rabeprazole
our group has previously observed in the fasted state (20). could have helped in confirming that levels were not impacted
Based on atazanavir’s in vitro solubility profile, the greatest by the other study drugs.
loss to solubility occurs between pH 2–4, where a decrease The loss of ritonavir exposure is another consideration for
in solubility from 5.2 mg/mL to <0.005 mg/mL is report- the observed reductions in atazanavir exposure. Similar to
ed (8). The gastric pH-time profiles for treatment A and B atazanavir, ritonavir is an HIV PI and a weakly basic drug
fall in the pH range of approximately 1.5 and 3 following (pKa = 2.8) with antiviral properties at therapeutically rele-
atazanavir dosing; therefore, relatively small changes in vant doses (300 mg twice daily). At low doses (100 mg once
gastric pH within this range have the potential to result daily), ritonavir exerts no antiviral activity but acts as a potent
in marked changes in drug solubility. inhibitor of CYP3A4 and P-gp. Coadministration of ritonavir
The relationship between intra-gastric pH and atazanavir with atazanavir, a substrate of CYP3A4 and P-gp, results in a
bioavailability has been previously explored by Eley et al. (24). favorable drug-drug interaction that reduces atazanavir me-
Using pooled data from two ARA studies, the authors con- tabolism and improves bioavailability. The interplay between
cluded that atazanavir exposures were largely unaffected at apical gut efflux transporters and metabolizing enzymes locat-
gastric pH < 4, but decreased substantially at pH > 4. ed within enterocytes has been well-characterized (27,28).
Although ritonavir-boosted atazanavir regimens were investi- Generally, drugs with high permeability rates and extensive
gated in both studies included in the ad hoc analysis, only data metabolism (BDDCS Class 1 and 2) are able to enter
from the atazanavir 400 mg arms (without ritonavir) were enterocytes through passive diffusion. However, for BDDCS
presented. Here, substantial reductions in atazanavir exposure Class 2 compounds such as atazanavir, poor aqueous solubil-
were noted in both treatments B and C, despite gastric pH ity increases their susceptibility to apical efflux transporter

Fig. 4 Mean pH-time profiles by 4

treatment prior to and following
oral administration of ATV/RTV
300/100 mg (t = 0 min). Individual 3
3-min median pH values were
averaged to construct each profile.
pH

A low fat meal was consumed 2

10 min prior to ATV/RTV dosing in
treatments A and B. In Treatment C
1
the meal was administered 10 min
prior to BHCl dosing, which was
followed 5 min later by ATV/RTV 0
dosing. In treatments B and C the -30 0 30 60 90 120 150
meal was administered 10 min after Time (min)
subjects exhibited a gastric pH of
3.8 or greater for 10 min.
Treatment A: ATV/RTV

Treatment B: ATV/RTV + RAB

Treatment C: ATV/RTV + RAB + BHCl

626 Faber et al.

effects in the gut (29). In this study, ritonavir Cmax and AUC are, however, other factors outside of the scope of this study
were both reduced by approximately 40% following pretreat- that may account in part to lower atazanavir exposures in the
ment with rabeprazole, likely a result of its own pH-dependent presence of ARAs.
solubility. Similar modest reductions in ritonavir exposure Recent in vitro and in situ studies have demonstrated that
were also observed in the presence of high doses of the PPI atazanavir permeability is sensitive to intestinal pH (31). Kis
omeprazole (9). The impact of a low dose of ritonavir on the et al. reported that atazanavir permeability rates (apical-to-
pharmacokinetics of atazanavir has been characterized previ- basolateral) increased with decreasing luminal pH over the
ously in a drug interaction study (8). In that study, the addition pH range of 4.5–7.4. In addition, the inhibitory effect of a
of 100 mg of ritonavir to 300 mg of atazanavir increased known P-gp inhibitor on atazanavir efflux was decreased at
atazanavir Cmax and AUC by 86 and 238%, respectively. pH 5.5 compared to 7.4 (78% and 245%, respectively), sug-
Therefore, reductions in ritonavir exposure in our study likely gesting that increases in atazanavir permeability at more acidic
contributed to a decrease in atazanavir bioavailability; how- pHs may be due to a decrease in P-gp mediated efflux. This was
ever, given that ARAs also reduce atazanavir levels in the a surprising observation as permeability rates for a weakly basic
absence of ritonavir, other factors were likely to be involved. drug typically increase at a higher pH due to a larger propor-
The effect of BHCl on the pharmacokinetics of atazanavir tion of the uncharged species (32). Median fasting duodenal pH
during rabeprazole-induced hypochlorhydria was examined in healthy volunteers is approximately 6.0 and has been report-
in treatment C. The addition of BHCl only restored 12– ed to increase in individuals taking a PPI (16,22). Additional
13% of the mean atazanavir exposure lost due to pretreat- clinical studies are needed to further explore the interplay be-
ment with rabeprazole and the increase was not statistically tween intestinal pH and atazanavir absorption in the presence
significant. Evaluation of pH-time profiles indicate that the and absence of food and ARAs; however, it is interesting to
buffering effect of food appeared to reduce the reacidification consider that changes in intestinal pH at the site of drug ab-
potential of BHCl, compared to the fasted state. Following sorption may contribute in part to alterations in bioavailability.
consumption of a light meal, the mean gastric pH values for Although our study aimed to examine the effects of gastric
all three treatments converged around pH 2.5-3 at the time of pH on ritonavir-boosted atazanavir pharmacokinetics, a num-
atazanavir administration (Fig. 4). During treatment A, pH ber of considerations must be addressed. The most telling, as
levels increased in all subjects at the start of the meal. For observed from the data presented herein, is the sheer com-
treatments B and C, where subjects were pretreated with plexity and impact of food on drug absorption. In addition to
rabeprazole and confirmed to have pH ≥4 prior to the start a meal’s immediate impact on gastric pH, the relative effects
of their meal, gastric pH fell in all but one subject immediately in the intestine, which include alteration of intestinal pH, sol-
following food intake, presumable due to the buffering effects ubilizing effects on oral dosage forms, and effects on intestinal
of the meal (Fig. 4.). Overall, the presence of food in this study drug metabolizing enzymes and transporters, may be of great-
appeared to narrow the difference in gastric pH between er significance to certain drug’s dissolution and absorptive
treatments containing the PPI versus the control, and delay processes. Furthermore, and specifically in the case of
the onset of BHCl, as evidenced by a shallower decline in atazanavir, the complexity surrounding bioavailability is fur-
gastric pH following administration of BHCl, relative to the ther raised due to pharmacokinetic interactions with ritonavir
fasted state. Both of the aforementioned observations may that occur upon first pass, as described above. Along these
help explain recovery differences in exposures between this lines, additional aspects of the clinical utility of BHCl in im-
study and our dasatinib study, which was conducted in a proving the absorption of weakly basic drugs given during
fasted state. drug-induced hypochlorhydria were also brought to light,
Oral drug absorption is complex and multifactorial. such as the timing of the administration of BHCl relative to
Determining the relative contribution of, and interplay be- administration of the victim drug.
tween, individual factors that alter bioavailability can be chal- In a previously conducted pilot study, we reported that the
lenging, particularly for a drug like atazanavir, which patients average onset of effect of BHCl, defined as the time to achieve a
are instructed to take with food and in combination with other pH < 3, was just over 5 min (19), which was subsequently used
medications. To our knowledge, no group has investigated the in the design of our study with dasatinib in the fasting state.
relationship between gastric pH as it relates to food and However, as observed in this study, this may not necessarily be
atazanavir pharmacokinetics, most likely because gastric pH ideal under conditions in which a drug must be administered
elevations are one of many ways food affects GI physiology with food as gastric pH effects may become secondary to the
and drug absorption (30). However, to minimize potential intestinal effects on drug absorption, as described above. In this
confounders in this study, all participants were administered study, gastric pH begins to rise around the 90 min time point in
the same standardized light meal during each treatment. Treatment C, suggesting that a second dose of BHCl at or
Alterations in gastric acid are thought to play a role in the before before this point may result in additional gains to expo-
clinical interaction between atazanavir and ARAs. There sure since the reported atazanavir Tmax in healthy volunteers is
Meal Effects Confound Attempts to Counteract Hypochlorhydria 627

2.7 h. This pH observation is consistent with the atazanavir contributions and through Dr. Benet’s consultations and ex-
pharmacokinetic profiles in Treatments B and C, where the pert witness fees and Board of Directors remunerations, all of
greatest differences in concentrations were observed at the which are made payable to the Regents of the University of
1.5 h time point (mean atazanavir concentration was 3.4 times California to support the research studies of the Benet
higher in Treatment C, compared with Treatment B). Laboratory. This clinical study was supported in part by the
Concentrations remain higher in Treatment C for subsequent National Center for Advancing Translational Sciences,
time points, however, the difference between the two treat- National Institutes of Health, through UCSF-CTSI Grant
ments decreases. While these observations were outside the Number UL1 TR000004.
scope of this current study, further dedicated studies to evaluate
these hypotheses are warranted to broaden our understanding
of the clinical utility of BHCl usage and the potential benefit it REFERENCES
may have on patient outcomes.
The interaction between ritonavir-boosted atazanavir and 1. Panel on antiretroviral guidelines for adults and adolescents.
PPIs has been studied previously in patients infected with Guidelines for the use of antiretroviral agents in HIV-1 infected
adults and adolescents. Department of Health and Human
HIV; however the clinical relevanceis not yet fully understood
Services [cited 2015 October 3]. Available from: http://www.
because results have been inconsistent (33,34). Although aidsinfo.nih.gov/ContentFiles/AdultandAdolescentGL.pdf.
conducting this study in HIV-infected patients would have 2. Josephson F. Drug-drug interactions in the treatment of HIV infec-
provided further insight into the clinical significance of this tion: focus on pharmacokinetic enhancement through CYP3A in-
interaction and mitigation strategy, the choice to test this hy- hibition. J Intern Med. 2010;268(6):530–9.
3. Falcon RW, Kakuda TN. Drug interactions between HIV protease
pothesis in healthy volunteers first was made in an effort to inhibitors and acid-reducing agents. Clin Pharmacokinet.
isolate the effects of pH on the drug interaction by minimizing 2008;47(2):75–89.
differences in GI physiology that have been noted in HIV 4. US Food and Drug Administration. Atazanavir sulfate capsule,
patients (35), and limiting the impact additional drugs may gelatin coated (Reyataz) prescribing information. 2015.
5. US Food and Drug Administration. Atazanavir sulfate and
have on pharmacokinetic outcomes. With a better under-
cobicistat (Evotaz) prescribing information. 2015.
standing of the role pH plays, future studies can assess these 6. Larson KB, Wang K, Delille C, Otofokun I, Acosta EP.
differences as well as their contribution to the overall clinical Pharmacokinetic enhancers in HIV therapeutics. Clin
significance of this interaction in a patient population. Pharmacokinet. 2014;53(10):865–72.
7. Benet LZ, Broccatelli F, Oprea TI. BDDCS applied to over 900
drugs. AAPS J. 2011;13(4):519–47.
8. US Food and Drug Administration. Atazanavir (Reyataz) summary
CONCLUSIONS basis of approval. 2003
9. Agarwala S, Gray K, Wang Y, Grasela D. Pharmacokinetic (PK)
effect of omeprazole (OMP) on atazanavir (ATV) with ritonavir
As expected, co-administration of the PPI rabeprazole with
(RTV) in healthy subjects. 12th Conference on Retroviruses and
atazanavir/ritonavir markedly decreased the systemic concen- Opportunistic Infections; 2005 Feb 22–25; Boston, MA.
trations of these poorly soluble, weakly basic HIV protease 10. Agarwala S, Gray K, Eley T, Wang Y, Hughes E, Grasela D.
inhibitors. Attempts to reverse this effect through administra- Pharmacokinetic interaction between atazanavir and omeprazole
tion of the gastric reacidification agent betaine hydrochloride, in healthy subjects. 3rd IAS Conference on HIV Pathogenesis and
Treatment; 2005 Jul 24–27; Rio de Janiero, Brazil.
previously shown to mitigate PPI-induced decreases with 11. Wang X, Boffito M, Zhang J, Chung E, Zhu L, Wu Y, et al. Effects
dasatinib, were unsuccessful here. In addition, the gastric pH of the H2-receptor antagonist famotidine on the pharmacokinetics
changes noted for betaine hydrochloride in fasted subjects of atazanavir-ritonavir with or without tenofovir in HIV-infected
were not repeated here in subjects receiving a concomitant patients. AIDS Patient Care STDs. 2011;25(9):509–15.
12. Zhu L, Persson A, Mahnke L, Eley T, Li T, Xu X, et al. Effect of
light meal with the parent drug. This study highlights the
low-dose omeprazole (20 mg daily) on the pharmacokinetics of
complexity of food effects on drug absorption and notes the multiple-dose atazanavir with ritonavir in healthy subjects. J Clin
necessity of further work required before the timing and ex- Pharmacol. 2011;51(3):368–77.
pected outcome of food effects and gastric pH may be better 13. Budha NR, Frymoyer A, Smelick GS, Jin JY, Yago MR, Dresser
understood. MJ, et al. Drug absorption interactions between oral targeted anti-
cancer agents and PPIs: is pH-dependent solubility the Achilles heel
of targeted therapy? Clin Pharmacol Ther. 2012;92(2):203–13.
14. Zhang L, Wu F, Lee SC, Zhao H, Zhang L. pH-dependent drug-
ACKNOWLEDGMENTS AND DISCLOSURES drug interactions for weak base drugs: potential implications for
new drug development. Clin Pharmacol Ther. 2014;96(2):266–77.
Systemic concentration measurements of atazanavir and rito- 15. Smelick GS, Heffron TP, Chu L, Dean B, West DA, Duvall SL,
et al. Prevalence of acid-reducing agents (ARA) in cancer popula-
navir carried out at Tandem Laboratories were supported by tions and ARA drug-drug interaction potential for molecular
Bristol Myers Squibb. All other aspects of the study were targeted agents in clinical development. Mol Pharm. 2013;10(11):
funded by the Benet Fund for Excellence through 4055–62.
628 Faber et al.

16. Walravens J, Brouwers J, Spriet I, Tack J, Annaert P, Augustijns P. 25. Meyer UA. Metabolic interactions of the proton-pump inhibitors
Effect of pH and comedication on gastrointestinal absorption of lansoprazole, omeprazole and pantoprazole with other drugs. Eur J
posaconazole: monitoring of intraluminal and plasma drug concen- Gastroenterol Hepatol. 1996;8 Suppl 1:S21–5.
trations. Clin Pharmacokinet. 2011;50(11):725–34. 26. Miura M, Satoh S, Tada H, Habuchi T, Suzuki T. Stereoselective
17. Ray JE, Marriott D, Bloch MT, McLachlan AJ. Therapeutic drug metabolism of rabeprazole-thioether to rabeprazole by human liver
monitoring of atazanavir: surveillance of pharmacotherapy in the microsomes. Eur J Clin Pharmacol. 2006;62(2):113–7.
clinic. Br J Clin Pharmacol. 2005;60(3):291–9. 27. Cummins CL, Jacobsen W, Benet LZ. Unmasking the dynamic
18. Chin TW, Loeb M, Fong IW. Effects of an acidic beverage (Coca- interplay between intestinal P-glycoprotein and CYP3A4. J
Cola) on absorption of ketoconazole. Antimicrob Agents Pharmacol Exp Ther. 2002;300(3):1036–45.
Chemother. 1995;39(8):1671–5. 28. Cummins CL, Salphati L, Reid MJ, Benet LZ. In vivo modulation of
19. Yago MR, Frymoyer AR, Smelick GS, Frassetto LA, Budha NR, intestinal CYP3A metabolism by P-glycoprotein: studies using the
Dresser MJ, et al. Gastric reacidification with betaine HCl in healthy rat single-pass intestinal perfusion model. J Pharmacol Exp Ther.
volunteers with rabeprazole-induced hypochlorhydria. Mol Pharm. 2003;305(1):306–14.
2013;10(11):4032–7. 29. Shugarts S, Benet LZ. The role of transporters in the pharmacoki-
20. Yago MR, Frymoyer A, Benet LZ, Smelick GS, Frassetto LA, Ding netics of orally administered drugs. Pharm Res. 2009;26(9):2039–
X, et al. The use of betaine HCl to enhance dasatinib absorption in 54.
healthy volunteers with rabeprazole-induced hypochlorhydria. 30. Winstanley PA, Orme ML. The effects of food on drug bioavail-
AAPS J. 2014;16(6):1358–65. ability. Br J Clin Pharmacol. 1989;28(6):621–8.
21. Schuster A, Burzawa S, Jemal M, Loizillon E, Couerbe P, Whigan 31. Kis O, Walmsley SL, Bendayan R. In Vitro and In Situ evaluation of
D. Quantitative determination of the HIV protease inhibitor pH-dependence of atazanavir intestinal permeability and interac-
atazanavir (BMS-232632) in human plasma by liquid tions with acid-reducing agents. Pharm Res. 2014;31(9):2404–19.
chromatography-tandem mass spectrometry following automated 32. Zheng Y, Benet LZ, Okochi H, Chen X. pH dependent but not P-
solid-phase extraction. J Chromatogr B Anal Technol Biomed Life gp dependent bidirectional transport study of S-propranolol: the
Sci. 2003;788(2):377–86. importance of passive diffusion. Pharm Res. 2015;32(8):2516–26.
22. Dressman JB, Berardi RR, Dermentzoglou LC, Russell TL, 33. Guiard-Schmid JB, Poirier JM, Bonnard P, Meynard JL. Lack of
Schmaltz SP, Barnett JL, et al. Upper gastrointestinal (GI) pH in interaction between atazanavir and proton pump inhibitors in
young, healthy men and women. Pharm Res. 1990;7(7):756–61. HIV-infected patients treated with ritonavir-boosted atazanavir. J
23. Malagelada JR, Longstreth GF, Summerskill WH, Go VL. Acquir Immune Defic Syndr. 2006;41(3):393–4. author reply 394.
Measurement of gastric functions during digestion of ordinary solid 34. Klein CE, Chiu YL, Cai Y, Beck K, King KR, Causemaker SJ, et al.
meals in man. Gastroenterology. 1976;70(2):203–10. Effects of acid-reducing agents on the pharmacokinetics of
24. Eley T, Agarwala S, Wang R, Gray K, Chung E, Wang Y, et al. lopinavir/ritonavir and ritonavir-boosted atazanavir. J Clin
Analysis of intra-gastric pH and atazanavir bioavailability and Pharmacol. 2008;48(5):553–62.
healthy subjects. 7th International Workshop on Clinical 35. Welage LS, Carver PL, Revankar S, Pierson C, Kauffman CA.
Pharmacology of HIV Therapy; 2006 Apr 20–22; Lisbon, Alterations in gastric acidity in patients infected with human immu-
Portugal. nodeficiency virus. Clin Infect Dis. 1995;21(6):1431–8.