Genetic Factors Influencing Severe Atazanavir-Associated Hyperbilirubinemia in a Population with Low UDP-Glucuronosyltransferase 1A1*28 Allele Frequency
Wan Beom Park
Pyoeng Gyun Choe
Jae Hyun Jeon
Sang Won Park
Hong Bin Kim
Nam Joong Kim
Kang Won Choe
Received 28 December 2009; accepted 20 March 2010; electronically published 26 May 2010. Presented in part: Annual Meeting of the Infectious Diseases Society of America
Philadelphia, PA, October 2009 (abstract 286).
Seoul National University College of Medicine
28 Yeongun-dong, Chongro-gu, Seoul
Department of Internal Medicine, Seoul National University College of Medicine
Republic of Korea
Background. High prevalence of severe atazanavir-associated hyperbilirubinemia in Asians with low prevalence of the UDP-glucuronosyltransferase (UGT)1A1*28 polymorphism suggests the importance of genetic factors other than UGT1A1*28 for atazanavir-associated hyperbilirubinemia in these populations. Methods. Serum bilirubin levels were measured in 129 Korean human immunodeficiency virus-infected patients 3 months after initiation of atazanavir (400 mg per day) with good adherence to medication. The multidrug resistance gene 1 (MDR1) C3435T and G2677T/A variations and UGT1A1*6 and *28 were examined by direct sequencing of DNA from peripheral whole blood samples. The associations between genetic polymorphisms and severe (grade 3-4) hyperbilirubinemia were evaluated using multivariate logistic regression analysis including demographic and clinical variables. Results. The median patient age was 39 years (interquartile range, 34-51 years), and 91% were men. At baseline, the median CD4 cell count was 261 cells/mL (interquartile range, 181-405 cells/mL). Severe hyperbilirubinemia was detected in 27 patients (21%). The independent risk factors for severe hyperbilirubinemia were low baseline CD4 cell count (adjusted odds ratio per 10 cells/mL increase, 0.97; 95% confidence interval, 0.940.99), UGT1A1*28 (adjusted odds ratio, 4.15; 95% confidence interval, 1.46-11.84), and MDR1 G2677T/A (adjusted odds ratio, 9.65; 95% confidence interval, 1.09-85.61). Of 19 patients with wild-type alleles for both MDR1 2677 and UGT1A1*28, none developed severe hyperbilirubinemia. Conclusion. The MDR1 G2677T/A variation and UGT1A1*28 are independent risk factors for severe atazanavirassociated hyperbilirubinemia in Korean human immunodeficiency virus-infected patients. Atazanavir is a human immunodeficiency virus (HIV) that atazanavir-associated hyperbilirubinemia is linked protease inhibitor associated with a clinically benign to a UDP-glucuronosyltransferase (UGT) promoter but potentially stigmatizing unconjugated hyperbilivariant that contains 7 thymine adenine (TA) nucleorubinemia; under conditions of severe hyperbilirubi-
tide repeats, A(TA)7TAA (UGT1A1*28), compared with
nemia, jaundice frequently develops, which often
rethe common promoter that contains only 6 TA repeats
sults in drug discontinuation . It is well documented
; the UGT1A1*28 promoter is less transcriptionally
active than the common promoter. This
pathophysiology is similar to that of Gilbert syndrome, which is
the most common inheritable condition leading to
transient unconjugated hyperbilirubinemia .
The frequency of the UGT1A1*28 allele is much
lower in Asians, including Koreans (13%), Chinese
(16%), and Japanese (11%), compared with Caucasians
(36%39%) and African Americans (43%) .
However, we previously found a similar prevalence of severe
atazanavir-associated hyperbilirubinemia in Koreans
and Caucasians . These findings suggest that genetic
factors other than UGT1A1*28 may play a role in the
development of atazanavir-associated hyperbilirubinemia in Asians.
The multidrug resistance gene 1 (MDR1) encodes a
transmembrane transporter, the P-glycoprotein (P-gp), that
functions in the absorption and distribution of most HIV-1 protease
inhibitors . Data on the association between MDR1
polymorphisms and atazanavir-associated hyperbilirubinemia
remain controversial . Rodriguez-No voa et al  suggested
that a polymorphism in MDR1 (C3435T) was associated with
lower concentrations of atazanavir and a reduced risk for
hyperbilirubinemia in Caucasians. In contrast, Ma et al  did
not find an association between MDR1 C3435T or G2677T and
concentration of atazanavir in several populations, including
African-Americans, Caucasians, and Hispanics. However, to
our knowledge, no study has investigated the relationship
between MDR1 polymorphisms and atazanavir-associated
hyperbilirubinemia in Asians.
In addition, other polymorphisms in the UGT gene may also
affect the development of atazanavir-associated
hyperbilirubinemia in Asians. The UGT1A1*6 polymorphism (G211A,
G71R), which is common in some Asians (13%23%) and rare
in Caucasians (0.1%) , results in an 70% reduction in
the rate of bilirubin glucuronidation in vitro  and is
associated with Gilbert syndrome in Asians, irrespective of the
UGT1A1*28 allele [6, 12]. In addition, Boyd et al  reported
that UGT1A1*6 was associated with indinavir-associated
hyperbilirubinemia. However, no study has examined the
association between UGT1A1*6 and atazanavir-associated
hyperbilirubinemia. In the present study, we examined potential
genetic factors for severe atazanavir-associated
hyperbilirubinemia in Korean HIV-infected patients with a low prevalence
of the UGT1A1*28 allele.
PARTICIPANTS, MATERIALS, AND METHODS
Study participants. This study included adult Korean
HIVinfected patients who initiated antiretroviral therapy with
unboosted atazanavir (400 mg per day) from May 2005 through
April 2007 at the Seoul National University Hospital. The
1600bed, university-affiliated teaching hospital is the largest referral
centre for HIV/AIDS in South Korea; one-quarter of all
HIVinfected patients in South Korea are seen at this hospital.
Patients who did not show viral suppression 3 months after
initiating atazanavir, who had active liver disease, or who did not
consent to participating in the study were excluded. Viral
suppression 3 months after initiating atazanavir was defined as an
HIV RNA level !40 copies/mL or at least a 2 log10 decrease in
HIV RNA copies/mL .
Three months after initiating atazanavir treatment, total
bilirubin levels were measured in serum samples.
Atazanavir-associated hyperbilirubinemia was defined as hyperbilirubinemia
(bilirubin level, 11.3 mg/dL) that developed after initiation of
atazanavir therapy in the absence of other causes of
hyperbilirubinemia. Hyperbilirubinemia was classified on the basis of
the AIDS Clinical Trials Group guidelines for total bilirubin
levels as follows: grade 1 (mild), 2332 mmol/L (1.31.9 mg/
dL); grade 2 (moderate), 3353 mmol/L (1.93.1 mg/dL); grade
3 (severe), 54105 mmol/L (3.16.1 mg/dL); and grade 4
(serious), 1105 mmol/L (16.1 mg/dL) . Severe
hyperbilirubinemia was defined as grade 3 and 4 hyperbilirubinemia.
Consent for genetic analyses was obtained from all participants,
and the study protocol was approved by the Institutional
Review Board of Seoul National University Hospital.
Genetic analyses. Genomic DNA was extracted from
peripheral whole blood with use of the AccuPower Genomic DNA
Extraction Kit (Bioneer). Primer sequences for polymerase
chain reaction (PCR) amplification of both the TATA box of
the UGT1A1 promoter and part of UGT1A1 exon 1 were 5
ATTAACTTGGTGTATCGATTGG-3 and 5
-AAGCATAGCAGAGTCCTTTTTTA-3 . PCR was performed under the
following thermal cycling conditions: 95 C for 15 min, followed by
35 cycles at 95 C for 20 s, 52 C for 40 s, and 72 C for 1 min.
PCR primers for amplification of the MDR1 gene, including
position 2677 (exon 21), were 5
-TCAGAAAATAGAAGCATGAGTTG-3 and 5 -AGCAGTAGGGAGTAACAAAATAAC-3 ;
the PCR conditions were as follows: 95 C for 5 min, followed
by 45 cycles at 95 C for 30 s, 56 C for 45 s, and 72 C for 1
min. PCR primers for amplification of the MDR1 gene,
including position 3435 (exon 26), were 5
-TCTTGTTTTCAGCTGCTTGATGG-3 and 5
-AGAGACTTACATTAGGCAGTGAC-3 ; the PCR conditions were as follows: 95 C for 15 min,
followed by 45 cycles at 95 C for 20 s, 56C for 30 s, and 68 C
for 30 s.
All PCR products were purified and directly sequenced
using the AMI 3730XL DNA Analyzer (PE Applied Biosystems);
samples were analyzed for the number of TA repetitions in
the UGT1A1 promoter, identity of position 211 in the UGT1A1
gene, and identity of positions 2677 and 3435 in the MDR1 gene.
Statistical analysis. All statistical analyses were performed
with SPSS software, version 17.0 (SPSS). Descriptive results of
continuous variables were expressed as median values and
interquartile range (IQR). Fishers exact test was used to compare
categorical variables. Logistic regression analysis was used to
determine risk factors for severe atazanavir-associated
hyperbilirubinemia. Variables in the models included age, sex,
hepatitis B or C infection, baseline CD4 cell count, and genetic
polymorphisms. Variables that were not significant in univariate
analyses (P 1 .10) were excluded from multivariate analyses. All
significance tests were 2-sided.
Study participants. Of a total of 190 patients who were
prescribed atazanavir during the study period, 12 patients (6%)
Table 1. Baseline Patient Characteristics
NOTE. Data are no (%) of patients, unless otherwise indicated. HBV,
hepatitis B virus; HCV, hepatitis C virus; HIV, human immunodeficiency virus; IQR,
discontinued treatment within 3 months after starting
atazanavir; six patients (3%) did so because of jaundice. One
hundred forty-four patients (75%) met the predefined criteria for
viral suppression 3 months after initiating atazanavir. By
excluding 12 patients who did not furnish informed consent for
the study and 3 patients who had active liver disease, a final
total of 129 patients were recruited for the study. All patients
were Korean, and the median age was 39 years (IQR, 3451
years); 91% of the patients were male (Table 1). Forty-eight
patients (37%) were naive to antiretroviral treatment, and 64
patients (50%) were taking other antiretroviral drugs within 1
month before beginning atazanavir treatment. Previous
antiretroviral drug regimens included nelfinavir (n p 24),
lopinavir/ritonavir (n p 24), efavirenz (n p 9), indinavir (n p 6),
and nevirapine (n p 1). Antiretroviral drugs that were
administered concurrently with atazanavir included
zidovudine-lamivudine (n p 103; 80%), lamivudine-stavudine (n p 14;
11%), lamivudine-abacavir (n p 7; 5%), and
lamivudine-didanosine (n p 5; 4%). No patients received any nonnucleoside
reverse-transcriptase inhibitor or tenofovir concurrently with
At baseline, the median plasma HIV RNA load was 4490
copies/mL (IQR, 2559,300 copies/mL), with 45 patients (35%)
having an HIV RNA level !40 copies/mL. The median CD4
cell count was 261 cells/mL (IQR, 181405 cells/mL). Hepatitis
B virus antigen was present in 8 (6%) patients, and hepatitis
C virus antibodies were present in 5 (4%) patients. The median
total bilirubin level was 0.75 mg/dL (IQR, 0.60.9 mg/dL).
Three patients were also receiving atorvastatin or rosuvastatin,
although none were taking any H2 blockers or proton pump
Severe hyperbilirubinemia according to UGT1A1 and MDR1
polymorphisms. The frequencies of the UGT1A1*6 and *28
alleles in the study participants were 19% and 11%, respectively.
The frequencies of the MDR1 2677 (T or A) and 3435 T alleles
were 54% and 31%, respectively (Table 2).
Three months after initiating atazanavir, 100 (78%) of the
129 patients developed hyperbilirubinemia and 27 (21%)
developed severe hyperbilirubinemia. Table 2 shows the
proportion of the patients with severe hyperbilirubinemia according
to genetic polymorphisms.
Risk factors for severe atazanavir-associated
hyperbilirubinemia. Univariate analyses showed old age, low baseline CD4
cell count, and presence of the UGT1A1*28 allele or MDR1
G2677T/A as potential risk factors for severe hyperbilirubinemia.
Multivariate analysis showed that independent risk factors for
severe hyperbilirubinemia were low baseline CD4 cell count
(adjusted odds ratio per 10 increase of CD4 cell count, 0.97; 95%
confidence interval, 0.940.99; P p .036), UGT1A1*28 allele
(adjusted odds ratio, 4.15; 95% confidence interval, 1.4611.84;
P p .008), and MDR1 G2677T/A (adjusted odds ratio, 9.65; 95%
confidence interval, 1.0985.61; P p .042) (Table 3).
no (%) of patients
a One patient in whom the genotype of UGT1A1*6 was not determined
Table 3. Clinical and Genetic Factors Associated with Severe Hyperbilirubinemia in Patients Treated
OR (95% CI)
aOR (95% CI)
NOTE. aOR, adjusted odds ratio; CI, confidence interval; OR, odds ratio.
In the 100 patients with hyperbilirubinemia, the severity
of hyperbilirubinemia was significantly associated with the
UGT1A1*28 and MDR1 G2677T/A polymorphisms (Figure 1).
Although none of the 15 patients who had wild-type alleles for
both UGT1A1*28 and MDR1 2677 developed severe
hyperbilirubinemia, we confirmed severe hyperbilirubinemia in 16
(26%) of the 62 patients who had wild-type alleles for
UGT1A1*28 and nonwild-type alleles for position MDR1 2677
(P p .031), as well as 10 (50%) of the 20 patients who did not
have wild-type alleles for both UGT1A1*28 and MDR1 2677
(P p .002).
To the best of our knowledge, ours is the first study to evaluate
the association between genetic polymorphisms and
hyperbilirubinemia in Asian HIV-infected patients receiving atazanavir.
Previous related studies have only evaluated Caucasian or
African-American HIV-infected patients. Furthermore, in
contrast to previous studies in which patient adherence to the drug
was not considered [2, 9], this study only included subjects
who had achieved viral suppression at the time of total bilirubin
measurement, indicating good adherence to atazanavir .
Our findings confirm that the UGT1A1*28 allele was strongly
associated with atazanavir-associated severe hyperbilirubinemia
in Korean HIV-infected patients, despite the lower prevalence
of the UGT1A1*28 allele in Korean patients, compared with
Caucasians. In the present study, the prevalence of the
UGT1A1*28 allele was 11%, which is comparable to that
previously reported in the general Korean population . In
particular, only 1 (0.8%) patient was homozygous for UGT1A1*28
(TA7/TA7), compared with 10% of Caucasians who exhibit the
homozygous genotype . However, the frequency of severe
hyperbilirubinemia in this study was comparable to that
previously reported in Caucasians .
The relatively high frequency of the UGT1A1*6 allele in
Asians may explain the prevalence of Gilbert syndrome in an
Asian population with a low prevalence of the UGT1A1*28
allele [11, 12, 19]. In the present study, the prevalence of
UGT1A1*6 was as high as 19%, whereas it was rarely found
in Caucasians . Our findings did not demonstrate an
association between UGT1A1*6 and severe atazanavir-associated
hyperbilirubinemia in Korean HIV-infected patients. However,
the lack of a statistically significant association could be
attributable to the small number of patients in each category who
developed severe hyperbilirubinemia.
Our study did identify MDR1 G2677T/A as another
independent risk factor for severe hyperbilirubinemia in Korean
HIV-infected patients taking atazanavir. Interestingly, among
19 patients who were wild-type for both MDR1 2677 and
UGT1A1*28, none experienced severe hyperbilirubinemia, and
8 (53%) of 15 patients who developed hyperbilirubinemia only
showed mild (grade 1) hyperbilirubinemia.
Figure 1. Proportion of severe (grade 34) hyperbilirubinemia according
to genetic polymorphisms in 100 human immunodeficiency virusinfected
patients with different grades of hyperbilirubinemia. G/G, homozygous
wild-type genotype; A or T, heterozygous or homozygous genotype for
polymorphism MDR1 G2677T/A; TA6/TA6, homozygous wild-type genotype;
TA6/TA7 or TA7/TA7, heterozygous or homozygous genotype for
The MDR1 G2677T/A polymorphism may be more common
in some Asians than in Caucasians or African-Americans.
Previous studies reported that 12%17% of Japanese, Chinese, and
Koreans carry the wild genotype (G/G) for MDR1 2677,
compared with 31%33% of Caucasians . This difference
may be partly attributed to the higher frequency of the MDR1
G2677A mutation in Asians (4%18%), compared with
Caucasians or African-Americans (0%2%) . In the present
study in a Korean population, the proportion of patients who
carried the wild genotype for MDR1 2677 was 17%, and the
frequency of the 2677A allele was 19%.
The 2 single-nucleotide polymorphisms in MDR1 2677 result
in distinct amino acid changes, Ala893Ser (G2677T) and
Ala893Thr (G2677A) . Most studies suggested that
polymorphisms in MDR1 G2677T/A were linked to an increased
level of the P-gp drug substrates , although some
contradictory data have been reported . Although the blood
levels of atazanavir were not evaluated in the present study, an
increased concentration of atazanavir, which was associated
with MDR1 G2677T/A, might cause severe hyperbilirubinemia.
Other studies suggested that the MDR1 C3435T
polymorphism was linked to a lower incidence of
atazanavir-associated hyperbilirubinemia in Caucasians , although the
association was not observed in a separate study by the same
group analyzing ritonavir-boosted atazanavir . However,
results from another study  and our study showed no
association between MDR1 C3435T and atazanavir-associated
Some possibilities may explain this discrepancy. First, the
difference may be attributable to differences among the study
populations, including ethnicity [31, 32]. The frequency of the
MDR1 C3435T allele was as low as 31% in this study, compared
with 48%54% in Caucasians [24, 31]. Second, because of
linkage disequilibrium between MDR1 C3435T and G2677T/A, the
previously observed findings that were attributed to the
synonymous C3435T single-nucleotide polymorphism, which does
not result in amino acid changes, might be due to the associated
nonsynonymous G2677T/A polymorphism .
In the present study, low CD4 cell count was significantly
associated with severe hyperbilirubinemia. In contrast, Torti et
al  demonstrated an association between high CD4 cell
count and the development of severe hyperbilirubinemia,
although their results could be influenced by adherence to
medication. In our study, poor medical conditions in patients with
low CD4 cell counts might increase the risk and severity of
hyperbilirubinemia. However, after adjusting this clinical
variable, the association of the genetic polymorphisms with severe
hyperbilirubinemia remained significant.
This study has several limitations. First, we excluded patients
with active liver disease, and there was no patient concomitantly
receiving tenofovir or a proton pump inhibitor, which could
significantly affect the concentration of atazanavir. However,
other confounding factors, such as medical conditions or
medications, might affect bilirubin levels 3 months after initiating
atazanavir. Second, other genetic polymorphisms that were not
evaluated in this study might affect the atazanavir-associated
development of hyperbilirubinemia. For example, atazanavir
concentration can be influenced by the single-nucleotide
polymorphisms in the pregnane X receptor (C63396T), which
regulates the expression of CYP3A4 and MDR1 , or in
SLCO1B1 (T521C), which codes for organic
anion-transporting polypeptide 1B1, an influx transporter that is
responsible for the uptake of protease inhibitors and unconjugated
bilirubin . Third, the outcome of the present study was
severe hyperbilirubinemia because this parameter could be
measured objectively. However, jaundice and possible treatment
interruption because of severe hyperbilirubinemia might be more
clinically relevant than hyperbilirubinemia per se. Finally, blood
concentrations of atazanavir were not evaluated in this study.
Although bilirubin levels after atazanavir initiation correlated well
with plasma drug levels in other studies , the bilirubin levels
in this study may not directly reflect the atazanavir concentration.
In summary, we identified MDR1 G2677T/A along with
UGT1A1*28 as independent risk factors for severe
atazanavirassociated hyperbilirubinemia in Korean HIV-infected patients.
The common prevalence of severe atazanavir-associated
hyperbilirubinemia in some Asians despite low prevalence of the
UGT1A1*28 allele may be attributable to the frequent MDR1
2677 polymorphisms in these populations.
Financial support. Korean Society for AIDS and the research fund of
Seoul National University Hospital (grant no 4-2007-006-0).
Potential conflicts of interest. All authors: no conflicts.