Therapeutic drug monitoring of atazanavir/ritonavir in pregnancy




Pregnant women experience physiological changes during pregnancy that can have a significant impact on antiretroviral pharmacokinetics. Ensuring optimal plasma concentrations of antiretrovirals is essential for maternal health and to minimize the risk of vertical transmission. Here we describe atazanavir/ritonavir (ATV/r) plasma concentrations in a cohort of pregnant women undergoing routine therapeutic drug monitoring (TDM).


Pregnant HIV-positive women received ATV/r as part of their routine pre-natal care. Demographic and clinical data were collected. ATV plasma concentrations ([ATV]) were determined in the first (T1), second (T2) and third (T3) trimesters and at postpartum (PP) using liquid chromatography−tandem mass spectrometry (LC-MS/MS).


From January 2007, 44 women (37 black African) were enrolled in the study. All received ATV/r at a dose of 300/100 mg once a day. Twenty-four had received antiretroviral therapy (ART) prior to pregnancy, and 20 initiated ATV/r in pregnancy. At the time nearest to delivery, 36 patients had undetectable plasma viral loads. [ATV] values were determined in 11 (T1), 25 (T2), 34 (T3) and 28 (PP) patients. [ATV] at 24 hours post-dose (C24) values significantly lower at T2/T3 relative to PP.


This study was carried out in one of the larger cohorts of women undergoing TDM for ATV in pregnancy. Lower [ATV] values were seen in T2/T3 compared with T1/PP. However, [ATV] were not associated with a lack of virologic suppression at delivery. Nonetheless, careful monitoring of women in pregnancy is required, and dose adjustment of ATV to 400 mg may be an option.


Combination antiretroviral therapy (cART) is recommended during pregnancy for the benefit of maternal health and for prevention of vertical transmission [mother-to-child transmission (MTCT)] of HIV to the baby. Pregnancy is associated with clinically significant alterations in drug absorption, distribution, metabolism and elimination which can impact on the choice of dosing regimens and may compromise treatment efficacy for both mother and baby. Changes include alterations in gastrointestinal motility and absorption, plasma volume, protein binding, cardiac output and the level of cytochrome P450 (CYP450) enzyme activity, most of which are believed to be regulated under hormonal control [1]. Therefore, study of the pharmacokinetics of antiretroviral drugs in pregnant women is crucial as achieving and maintaining optimal plasma concentrations of antiretroviral drugs is essential for maternal health and to minimize the risk of MTCT of HIV.

Atazanavir/ritonavir (ATV/r) is increasingly prescribed in pregnancy as it is potent and well tolerated. ATV has been assigned to pregnancy category B by the US Food and Drug Administration (FDA), as animal studies failed to reveal any evidence of teratogenicity at maternal doses that produced systemic drug exposures equivalent to that observed in humans at the recommended ATV/r dose of 300/100 mg per day [2].

Pharmacokinetic data on ATV/r exposure during pregnancy conflict with the findings of some studies reporting significantly lower ATV concentrations antepartum compared with postpartum [3-5], whereas others have reported no differences [6-9]. Ripamonti et al. reported comparable ATV concentrations during the third trimester and at postpartum in women receiving standard ATV/r dosing (300/100 mg once per day) [6]. However, in an intensive pharmacokinetic study conducted in the USA, ATV exposures were approximately 28% lower during the third trimester compared with postpartum. Nonetheless, despite the concerns surrounding ATV pharmacokinetics in pregnancy, there are available data to suggest that ATV-based cART remains efficacious throughout pregnancy. In a retrospective case report on 155 women of whom 96% received ATV at the standard dose of 300/100 mg once daily, of 130 women with viral load data available, 104 (80%) had viral loads of < 50 HIV-1 RNA copies/mL at delivery. In addition, for women who commenced ATV in pregnancy, the proportion with an undetectable viral load increased with time on treatment (29% and 72% with < 12 and > 12 weeks of ART, respectively). Therapeutic drug monitoring (TDM) was performed in a subset of 28 patients (routine in 17; for detectable viral load in 11), of whom only four subjects had subtherapeutic values (< 150 ng/mL) [10].

In view of the limited data and discrepancies concerning dosing, further pharmacokinetic studies are warranted to ensure the safe and effective use of ATV/r in pregnancy. The objectives of the current study were to determine ATV plasma concentrations in HIV-infected pregnant women receiving standard dosing of ATV/r (300/100 mg once daily) undergoing routine TDM during pregnancy and postpartum.



HIV-infected pregnant women attending the Infectious Diseases Clinic of the Mater Misericordiae University Hospital, Dublin, Ireland were recruited between January 2007 and July 2011. Pregnant women receiving a protease inhibitor (PI)-based regimen were eligible for inclusion and written informed consent was obtained. Patients received a triple-drug ART regimen containing the oral ATV capsule boosted with ritonavir (RTV) at the standard dose of 300/100 mg once daily as part of their antiretroviral regimen. Patients who had decompensated liver disease or, in the investigators' opinion, were likely to deliver within 2 weeks of study entry were excluded. There were no exclusionary medications and informed consent was obtained prior to enrolment in the study. Patient adherence was self-reported at the time of TDM.

Study design

Blood sampling was undertaken in the first (weeks 1−13), second (weeks 14−25) and third (weeks 26−40) trimesters in women already stable on ART at conception. For women who commenced ART in pregnancy, steady-state plasma concentrations were measured at 2 weeks following initiation of ART and during the third trimester. Additionally, women who remained on ATV/r after delivery had drug concentrations determined postpartum. Demographic and clinical parameters were collected. HIV plasma viral load (pVL) and CD4 cell counts were determined at baseline and at the time of TDM sampling (antepartum and postpartum) and at delivery. Throughout the study period, total plasma ATV concentrations were acquired in real time and the predetermined ATV minimum effective concentration (MEC) was 150 ng/mL [11]. Where the concentration fell below the MEC, patients were initially counselled and medication adherence was reinforced, regardless of their self-reported compliance. If no virological response was noted on their subsequent antenatal visit, dose adjustment was considered on a case-by-case basis.

Analytical and pharmacokinetic methods

Blood samples were taken by venipunture the morning after the previous dose of ATV/r (approximately 14−20 h post-dose). The time of drug intake was recorded. Blood was collected in heparin tubes and centrifuged immediately (1000 g for 10 min at 4°C) and the plasma was removed and stored at −30°C. Prior to analysis, the plasma was heat-inactivated (at 58°C for 40 min).

Total plasma ATV concentrations were determined in real time at the Liverpool Pharmacology Research Laboratories using a validated high-performance liquid chromatography−tandem mass spectrometry (HPLC-MS/MS) methodology. The laboratory complies with the International Conference on Harmonisation Good Clinical Practice (ICH-GCP) guidelines with standards and participates in an external quality assurance programme (KKGT, Radboud University Medical Centre, Nijmegen, The Netherlands). The assay lower limit of quantification (LLQ) for ATV was 11 ng/mL [12].

Data analysis

All demographic and clinical characteristics were given as the median (range). Observed ATV plasma concentrations were expressed in terms of the geometric mean with 95% confidence intervals (CIs). In addition, observed concentrations were extrapolated to 24 h post-dose (C24; ng/mL) using a population-derived ATV half-life of 8.6 h (in the presence of RTV) [13], as a means of normalizing the effect of variable sampling times post drug intake. Values with missing drug intake data were excluded. Inter-subject variation in plasma concentrations was estimated using a coefficient of variation, expressed as a percentage (CV%) [%CV = (standard deviation/mean) × 100]. ATV plasma concentrations observed antepartum and postpartum were related to the drug's proposed MEC of 150 ng/mL.

Differences in ATV plasma concentration data antepartum vs. postpartum were assessed independently using a one-way analysis of variance (ANOVA), with a Bonferroni correction to test for multiple comparisons. Normality of data was assessed using a Shapiro−Wilk test, and nonnormally distributed data were log-transformed. All statistics were performed and analysed using statsdirect (Version 2.6.8©2008; Biomedical Software, Statsdirect Ltd, Cheshire, UK). P-values were two-sided at the 0.05 significance level.


A total of 44 women were enrolled in the study (37 black African, six Caucasian and one other). Patient baseline characteristics are summarized in Table 1. Twenty-four women were receiving ART prior to pregnancy. Seventeen of the 24 women had an undetectable pVL at enrolment (one patient had a pVL of 103 copies/mL; data were unavailable for the other six women). Twenty women (eight treatment-naïve, 10 treatment-experienced, and two with an unknown treatment history) initiated ATV/r therapy in pregnancy. All took therapy for at least 2 weeks prior to the first TDM sampling. The median (range) gestational age at the time of ATV/r treatment initiation in these patients was 23.5 weeks (7–35 weeks). All patients were prescribed ATV/r at the standard dose of 300/100 mg once daily. The nucleoside reverse transcriptase inhibitor (NRTI) backbone was emtricitabine plus tenofovir (Truvada, Gilead Sciences Inc., Foster City, CA, USA) in the majority of cases (25 patients; 57%).

Table 1. Baseline demographics and clinical characteristics
  1. ART, antiretroviral therapy; bid, twice daily; qd, once daily; TDM, therapeutic drug monitoring.
  2. aValues are given as median (range; n).
  3. bValues are given as n (%).
  4. cIn patients initiating ART during pregnancy.
Number of patients44
Age (years)a30.0 (20.0–40.0; 44)
CD4 count (cells/μL)a391 (153–869; 26)
HIV RNA (copies/mL)a10 857 (<50–107 7434; 24)
HIV RNA (< 50 copies/mL)b18 (75%; 24)
Black37 (84)
Caucasian6 (14)
Other1 (2)
Patients initiating ART in pregnancyb20 (45)
Naïve to ART at baselinebc 
Yes8 (40)
No10 (50)
Unknown2 (10)
Gestational age at ART initiation (weeks)ac23.5 (7.0–35.0; 20)
Gestational age at first TDM (weeks)a26 (9.0–39.0; 20)
Combivir12 (27) (10 bid; 2 qd)
Kivexa (qd)4 (9)
Trizivir (bid)1 (2)
Truvada (qd)25 (57)
Tenofovir (qd)1 (2)

ATV concentrations were determined in 11 patients in the first trimester, 25 in the second and 34 in the third. The median (range) gestational age at the time of pharmacokinetic sampling was 9 weeks (6–12 weeks) in the first trimester, 20 weeks (14–25 weeks) in the second trimester and 32 weeks (26–39 weeks) in the third trimester. In addition, 28 patients had measurements taken at postpartum. The median (range) follow-up time after delivery was 10.5 weeks (5–50 weeks).

At the time nearest to delivery, 36 patients had undetectable pVL, six patients had detectable pVL [65, 101, 108, 162, 291, 633 copies/mL] and two did not have data available. One patient (pVL 633 copies/mL) had her ATV dose increased to 600 mg at 37 weeks, which was then reduced to 400 mg a week later. She received single-dose nevirapine in labour and discontinued treatment post-delivery. Five of the six patients with detectable pVL had therapeutic concentrations in either the second or third trimester. A single subject (291 copies/mL) had undetectable plasma concentrations in the third trimester, which suggested issues with adherence. The majority were newly diagnosed at the antenatal clinic and commenced treatment at ≥ 24 weeks of gestation. Thirty-three patients (76%), including four of the six patients with detectable pVL at delivery, remained on ART postpartum for their own health. There were 43 live births (including two sets of twins), one miscarriage and one stillbirth in the cohort. One patient was transferred to another maternity unit. Delivery method was recorded in 33 of the 43 live births: nine (27.3%) were born by spontaneous vaginal delivery (SVD), 20 (60.6%) were delivered by caesarean section (nine elective; 11 emergency) and four (12.1%) were instrumental deliveries. All emergency caesarean sections were for obstetric reasons. The median (range) gestation age at delivery was 38.5 weeks (31.8−41.4 weeks) and the median (range) birthweight was 3.1 kg (1.52–4.72 kg). There were no HIV-positive infants.

The observed and extrapolated C24 ATV concentration data for the 44 patients during antepartum and postpartum are summarized in Table 2. Geometric mean ATV concentrations were comparable during the second [723 (95% CI 427−1019) ng/mL] and third [617 (95% CI 450−784) ng/mL] trimesters, but were 32–42% lower (albeit nonsignificantly) relative to ATV concentrations observed during the first trimester [1070 (95% CI 150–1989) ng/mL] and 41–50% lower than in the postpartum period [1226 (95% CI 962–1489) ng/mL; significant only for the third trimester, after adjusting for multiple comparisons; P < 0.001]. Furthermore, when the values were normalized to 24 h post-dose, ATV C24 concentrations achieved in both the second and third trimesters remained significantly lower compared with C24 postpartum (P = 0.003). Inter-subject variation in ATV plasma concentrations was high throughout pregnancy (63–102%) and moderately high at postpartum (49%). The time of post-dose sampling was consistent during the second and third trimesters of pregnancy and postpartum, at approximately 18–19 h. However, on average, sampling was undertaken earlier during the first trimester (∼12 h; P < 0.004), which may have accounted somewhat for the higher ATV plasma concentrations observed during this period. As a consequence, extrapolation of the first trimester concentrations to 24 h post-dose reduced the difference between the first and second/third trimesters to < 25% (Table 2). ATV C24 concentrations were not significantly different in the third trimester or at postpartum when comparing women who were or were not receiving tenofovir.

Table 2. Atazanavir (ATV) plasma concentrations (observed and extrapolated C24) and virological/immunological responses in 44 HIV-infected pregnant women receiving atazanavir/ritonavir (ATV/r) 300/100 mg once daily
 Trimester 1 (n = 11)Trimester 2 (n = 25)Trimester 3 (n = 34)Postpartum (n = 28)P-value
  1. [ATV], atazanavir plasma concentration; C24, 24 hour trough concentration; CV, coefficient of variation; 150 ng/mL = ATV minimum effective concentration (MEC); T1, first trimester; T2, second trimester; T3, third trimester; PP, postpartum; TDM, therapeutic drug monitoring.
  2. Bonferroni critical P for six comparisons = 0.0083. ATV concentrations below the assay lower limit of quantification are excluded.
  3. 1One missing sample; 2two missing samples; 3three missing samples.
  4. aValues are given as geometric mean (95% confidence interval).
  5. bValues are given as n (%).
  6. cValues are given as median (range; n).
Observed [ATV] (ng/mL)a1070 (150–1989)723 (427–1019)617 (450–784)1226 (962–1489)0.003, all comparisons;
0.01, T2 vs. PP;
0.0006, T3 vs. PP
CV (%)102786349
< 150 ng/mLb0 (0)2 (8)2 (6)2 (7)
Time post-dose (h)a12.0 (8.8–15.3)117.9 (15.5–20.3)319.0 (17.1–20.9)218.9 (16.5–21.3)1

0.006, all comparisons;

< 0.004, T1 vs. T2/T3/PP

Extrapolated [ATV] C24 (ng/mL)a557 (166–947)1477 (326–628)3431 (329–535)2893 (592–1193)10.003, all comparisons;
0.005, T2 vs. PP;
0.0005, T3 vs. PP
CV (%)84615671
C24 < 150 ng/mLb0 (0)3 (12)4 (11)3 (11)
Gestation/postpartum (weeks)c at time of TDM9 (6–12)20 (14–25)32 (26–39)10.5 (5–50)
pVL < 50 copies/mlb10/11 (91)19/23 (83)27/34 (79)23/28 (82)
CD4 (cells/μL)c386 (230–415; 10)394 (191–783; 23)413 (200–1032; 34)441 (105–875; 27)0.506
Weight (kg)c77.2 (59.8–91.1; 10)75.0 (58.6–99.3; 20)81.0 (52.0–127.0; 28)80.2 (45.0–121.2; 27)

Overall, three of 44 patients (7%) had ATV concentrations that were below the proposed ATV MEC (< 150 ng/mL) in both the second and third trimesters of pregnancy. The time of TDM sampling was within the 24-h dosing interval; however, one patient had complained of vomiting which may have resulted in reduced ATV concentrations. A further three patients had ATV concentrations approaching the MEC during the second (n = 1; 163 ng/mL; time post-dose = 11.6 h) and third (n = 2; 154 ng/mL; time post-dose = 26.8 h; 189 ng/mL; time post-dose = 12.0 h) trimesters, respectively. All six patients had been receiving ATV prior to pregnancy, no ATV/r dose adjustments were made throughout gestation and patients remained on treatment post-delivery. Of the six patients who had values that were either below or approaching the MEC during pregnancy, five had undetectable values (pVL < 50 copies/mL) at the time of TDM sampling, and one patient had a pVL of 31 692 copies/mL in the third trimester. Four achieved undetectable pVL at delivery (one had a pVL of 291 copies/mL; one did not have data available).

At postpartum, two of 28 patients, previously with therapeutic concentrations antepartum, had both undetectable ATV concentrations (time post-dose = 15.5 and 16.5 h, respectively) and detectable pVL (159 and 8273 copies/mL, respectively). One patient was approaching the MEC (183 ng/mL); this patient had previously had undetectable values during pregnancy. However, these subjects did not disclose that they were nonadherent to treatment, nor were they suspected by the study personnel of being so.

In a paired analysis of 28 patients with matched second and/or third trimester and postpartum samples, geometric mean ATV C24 concentrations were significantly (44–54%) reduced antepartum compared with postpartum (P = 0.002), as shown in Table 3. Twenty-three of the 28 patients (82%) experienced an increase in ATV C24 concentration at postpartum, compared with either the second or third trimester, as shown in Figure 1. Of the five patients with a decreased ATV C24 at postpartum, two had a history of noncompliance during pregnancy.

Table 3. Atazanavir (ATV) plasma concentrations (observed and extrapolated 24 hour trough concentration) and virological/immunological responses in 28 patients with samples taken both antepartum and postpartum
 Trimester 2 (n = 19)Trimester 3 (n = 23)Postpartum (n = 28)P-value
  1. Bonferroni critical P for three comparisons = 0.0166.
  2. One patient had antepartum measurements taken in the second trimester (22 weeks) as she delivered prematurely (27 weeks).
  3. [ATV], atazanavir plasma concentration; CV, coefficient of variation; T1, first trimester; T2, second trimester; T3, third trimester; PP, postpartum; pVL, plasma viral load; TDM, therapeutic drug monitoring.
  4. aValues are given as geometric mean (95% confidence interval).
  5. bValues are given as n (%).
  6. cValues are given as median (range).
  7. dTwo patients had undetectable values.
  8. eOne measurement unavailable.
  9. fTwo measurements unavailable.
Observed [ATV] (ng/mL)a749 (386–1111)d569 (376–762)d1226 (963–1490)d0.002, all comparisons;
CV (%)8265490.03, T2 vs. PP;
0.001, T3 vs. PP
< 150 ng/mL (%)b2 (10)2 (9)2 (7)
Time post-dose (h)a17.9 (15.1–20.7)19.6 (17.5–21.7)18.9 (16.6–21.2)0.913
Extrapolated [ATV] C24 (ng/mL)a500 (353–647)414 (286–543)893 (592–1193)0.002, all comparisons;
0.02, T2 vs. PP;
0.001, T3 vs. PP
CV (%)556171
C24 < 150 ng/mL3 (16)4 (17)3 (11)
Gestation/postpartum (weeks)c at time of TDM20.0 (14.0–25.0)34.0 (26.0–35.5)10.5 (5.0–50.0)
pVL < 50 copies/mlb15 (88)e21 (91)23 (82)
CD4 (cells/μL)c380 (211–668)f389 (200–656)441 (105–875)e0.290
Figure 1.

Change in atazanavir (ATV) twenty-four hour plasma concentrations (C24) in patients with paired pharmacokinetic samples taken throughout pregnancy and at postpartum (n = 28). Crosses indicate patients with a detectable plasma viral load at the time of therapeutic drug monitoring. The dotted line indicates the ATV minimum effective concentration (150 ng/mL).


This study has been carried out in one of the larger European (Irish) cohorts of women undergoing TDM for ATV in pregnancy, with the majority of women being of African descent. These data are vital given that previous studies investigating ATV/r pharmacokinetics in pregnancy have, thus far, yielded inconsistent results. A recent systematic review by Eley et al. provides a comprehensive overview of the key data from intensive pharmacokinetic trials of ATV in pregnancy [14]. In the BMS (Bristol-Myers Squibb) A1424182 trial, ATV area under the curve (AUC) and maximum concentration (Cmax) values were 21% and 27%, respectively, lower in pregnant subjects during the third trimester when compared with data from historical controls, although C24 values were comparable [9]. Ripamonti et al. observed no significant differences in ATV twenty-four hour area under the concentration-time curve (AUC0–24) and C24 during pregnancy and postpartum, whereas in the current study, and in other studies, ATV AUC0–24 and/or C24 increased significantly at postpartum compared with the second and third trimesters, respectively [6]. One explanation is that ATV concentrations at postpartum in the Italian cohort were lower than reported in other studies and compared with those seen in nonpregnant adults, which may have accounted for the lack of significance. Such differences in pharmacokinetic outcomes between international sites could be attributed to the relatively small sample sizes and large inter-patient variability in PI disposition, as well as physiological, genetic and environmental factors. Ethnicity may also have played a role, as 37 (84%) of the 44 women in our study were of African origin, compared with nine of 17 (53%) in the Italian cohort. This highlights the importance of acquiring pharmacokinetic data within different pregnancy populations.

In this study, lower ATV concentrations were observed in the second and third trimesters when compared with the first trimester and postpartum. However, such findings were not associated with significant viral breakthrough. Although six of 42 patients had detectable pVL at the time nearest delivery, only one patient had a detectable pVL associated with a subtherapeutic ATV concentration. Furthermore, concentrations in this subject were undetectable, suggesting that she may not have been fully adherent to treatment. Of the remaining five patients, reasons for not achieving virological suppression at delivery may relate to their late presentation, high baseline pVL (one subject had a pre-pregnancy pVL of 107 454 copies/mL) and timing of ART initiation (all initiated ART at ≥ 24 weeks) as opposed to ATV exposures per se, as all had therapeutic concentrations throughout pregnancy.

Most of the efficacy and safety data for cART in pregnancy are based on three/four drug combinations, with the NRTI backbone commonly consisting of zidovudine and lamivudine. However, use of tenofovir and emtricitabine (Truvada) in combination with a boosted PI is becoming increasingly common. This is because the most recent adult prescribing guidelines now recommend tenofovir/emtricitabine in first-line therapy, particularly in patients with pVL > 100 000 copies/mL [15]. Co-administration of ATV/r with tenofovir in nonpregnant adults has been reported to give a reduction of approximately 25% in ATV plasma concentrations [16], and a similar magnitude of reduction has been reported in pregnant women [4]. However, the data are inconclusive, with other studies showing an apparent lack of an interaction, both in the general HIV-infected population [17, 18] and, more recently, in pregnancy [19]. The most recent (2012) guidelines for the management of HIV infection in pregnancy recommend that TDM is undertaken during the third trimester in pregnant women receiving this combination [20]. In our cohort, concomitant tenofovir did not significantly impact on ATV concentrations, when compared with those not receiving tenofovir during the third trimester and postpartum. The current study, therefore, provides useful ‘real-life’ ATV concentration data and reassurance regarding the use of this combination in pregnancy. However, in line with the British HIV Association (BHIVA) guidelines, TDM should be considered on a case-by-case basis.

In conclusion, in this study lower ATV concentrations were observed in the second and third trimesters when compared with the first trimester and postpartum. However, changes in ATV plasma concentration did not appear to be detrimental to virological control, with the majority of women achieving virological suppression by the time of delivery. Nonetheless, careful monitoring of women in pregnancy is required, especially if there is concern about inadequate ATV concentrations. Dose adjustment of ATV upwards from 300 to 400 mg may therefore be an option in certain cases.


The authors would like to thank the medical staff of the maternity hospitals in Ireland for referring patients to The Mater Misericordiae University Hospital and The Rotunda Hospital for their HIV care in pregnancy.

Conflicts of interest:

SK and DB have received research grants and travel bursaries from Merck, Bristol Myers Squibb, GlaxoSmithKline, Pfizer, Abbott, Boehringer Ingelheim and Tibotec. JSL received an unrestricted grant from Bristol Myers Squibb which was used to pay for the costs of TDM testing for this project. LJE has received travel bursaries from Boehringer Ingelheim and Janssen Pharmaceuticals. VJ, MB and SCS have no conflicts of interest to declare.