Efavirenz pharmacokinetics among pregnant women with and without tuberculosis coinfection
Polly Clayden, HIV i-Base
Pregnancy increased the risk of low efavirenz (EFV) concentrations, but vertical transmission was rare in a South African study of HIV treatment in pregnant women with and without tuberculosis (TB) coinfection. 
Four other studies have shown similar limited effects of pregnancy on EFV pharmacokinetics (PK).  Twelve studies have shown no significant effects during long-term treatment with EFV and rifampicin. 
EFV is recommended for treatment of HIV positive pregnant women. Pregnancy and TB treatment or prophylaxis can affect EFV PK, a woman’s own HIV treatment and risk of vertical transmission.
A paper first published ahead of print in the Journal of Infectious Diseases in July, and updated 5 September 2014, showed results from a PK substudy of TSHEPISO – a prospective cohort study looking at the effects of TB on maternal and infant outcomes in pregnant HIV positive women. We previously summarised preliminary results from this substudy – presented at AIDS 2012 – in HTB South. [2,3]
HIV positive pregnant women with active TB or matched controls, from the Chris Hari Baragwanath Hospital in Soweto, were enrolled in TSHEPISO at 13 to 34 weeks gestation between January 2011 and January 2013. Two controls were enrolled for each case – matched for maternal and gestational age, date of enrolment and site of planned delivery but not antiretroviral (ART) use or regimen. Women receiving EFV for 10 days or more at week 36 of gestation could participate in the substudy.
The EFV dose was 600 mg once daily, isoniazid 5 mg/kg (maximum 300 mg) and rifampicin 10 mg/kg (maximum 600 mg) – according to national and WHO guidelines.
In the substudy, sampling was performed at 37 weeks gestation or at delivery and then six weeks post-partum. EFV levels were also measured in cord blood at delivery and in infants at one week of age. Plasma concentrations were determined by liquid chromatography-tandem mass spectrometry. EFV trough concentrations (Cmin) were predicted using population PK models. Women had CYP286 genotyping and were categorised as extensive, intermediate, slow or very slow metabolisers. Maternal viral load was measured at delivery and that of the infants at six weeks of age.
There were 97 participants in the substudy, 44 were coinfected with TB and 53 had HIV only. Cases and controls were similar at enrolment with median values of approximately: 28 years of age, 30 weeks gestation and a CD4 count of 310 cells/mm3. A greater proportion of controls had a viral load less than 20 copies/mL (58% vs 28%). This difference was probably explained by the delay in starting ART among those needing treatment for TB; duration of EFV use was 12 and 21 weeks for the cases and controls respectively. The median gestational age at delivery was 39 weeks.
The distribution of women with extensive, intermediate, slow and very slow CYP2B6 metaboliser status was similar between the two groups, respectively 22%, 56%, 19% and 3%.
The investigators reported a median EFV Cmin during pregnancy of 1.35 ug/mL (IQR 0.90-2.07) compared with 2.00 ug/mL (IQR, 1.40-3.59) postpartum; 27% compared with 13% of women had an EFV Cmin of <1 ug/mL. The proportion of women with extensive CYP2B6 metaboliser status with Cmin of <1 ug/mL during pregnancy was 72%.
The median EFV Cmin for women receiving rifampicin- and isoniazid-containing TB treatment during pregnancy was 1.33 ug/mL (IQR 0.83-2.22) compared with 1.57 ug/mL (IQR 1.56-1.83) for those receiving isoniazid alone as TB prophylaxis and 1.28 ug/mL (IQR 1.04-1.92) for those receiving no TB drugs. The proportion of women with Cmin of <1 ug/mL was 36% for those receiving rifampicin- and isoniazid, 20% for those receiving isoniazid and 23% for those receiving no TB drugs. Use of isoniazid was associated with higher Cmin among slow CYP2B metabolisers.
Cord blood samples were available for 50 infants and the median EFV concentration was 1.30 ug/mL (IQR 0.75-2.33); 10% were below the limit of quantification (0.2 ug/mL). The investigators noted that cord and maternal pre-partum concentrations were highly correlated, r=0.95. EFV concentrations for one-week old infants were below the limit of quantification in 66% (44/67); quantifiable EFV correlated with higher cord blood concentrations, r=0.75.
There was one HIV infection among 91/93 live born infants. The mother enrolled in the study at 37 weeks gestation, having received EFV-based ART for 67 days and TB treatment for 51 days. The infant was exclusively breastfed and received nevirapine prophylaxis. The infant tested negative at 22 days by HIV DNA PCR and had a CD4 percentage of 38% but tested positive at 48 days and had 32% CD4. At the second test the infant’s mother had an HIV viral load of 218, 672 copies/mL and reported breastfeeding.
The investigators concluded that pregnancy increased EFV clearance and the proportion of women with EFV Cmin <1ug/mL was higher in pregnancy than after delivery. They noted that women with extensive CYP2B6 metaboliser status might be at particularly high risk of low EFV Cmin and merit special attention if an EFV dose of 400 mg is considered.
TB treatment that includes isoniazid and rifampicin does not reduce overall TB concentrations but women with TB and HIV coinfection were more likely to have detectable viral load at delivery than those with HIV alone, probably because ART was started later. Vertical transmission was rare among participants and the investigators considered the one case that did occur likely to be related to breastfeeding and maternal adherence post partum.
Other studies show similar findings
In a related article, published as a letter in the 19 June 2014 edition of AIDS, Andrew Hill and colleagues from Liverpool University, WHO, Chelsea and Westminster Hospital and Chaing Mai University, performed a systematic review of five studies of EFV PK in pregnancy – including TSHEPISO from 2012 conference abstract – which included 235 women treated with 600 mg EFV once daily. Results from the literature review updated with published data from TSHEPISO are summarised in Table 1.
|No. of participants||97||25||105||13||16|
|Ethnicity||100% African||21/25 Asian||100% African||100% African||No data|
|PK sampling||Single samples||Intensive 24 hour||Single samples||Single samples||Single samples|
|Effect of pregnancy on EFV PK||Modest increase in clearance||No effects on Cmax or AUC, 12% reduction in C24h||No effects on clearance||EFV in infant plasma||No change in Cmin by trimester|
|Viral load at delivery||<20 copies/mL 58% (no TB treatment) 28% (TB treatment)||<400 copies/mL 100%||No data||<200 copies/mL 100%||<200 copies/mL 100%|
|Vertical trans. rate||1/91 (with PCR results)||1/25||0/56||0/13||No data|
The authors concluded that these five studies reported that EFV drug concentrations were not significantly affected and high rates of viral load suppression in the mothers at the time of delivery. “The overall results suggest that pregnancy has limited, if any clinically important effect, on efavirenz pharmacokinetics.” they wrote. The additional data from women in TSHEPISO does not appear to alter their conclusion.
No effects longer-term with EFV and rifampicin co-treatment
Andrew Hill and colleagues from Liverpool University and Chelsea and Westminster Hospital have also performed a systematic review looking at EFV PK in the presence of rifampicin (either Cmin or mid-dose concentration). This review was presented as a poster at AIDS 2014.
The authors identified studies with a sequential or cross-over design. They found six that evaluated <8 days of EFV-rifampicin treatment, with reductions in EFV concentrations of 19% to 54% compared with EFV treatment alone. One study of 2 to 6 weeks of EFV-rifampicin treatment showed mean Cmin reductions of 34%. Five longer-term studies evaluated 4 to 24 weeks of EFV-rifampicin treatment with increases in EFV concentrations of 6% to 26%. See Table 2.
|Study||Number of participants||Population||Country||Duration of EFV||Effect of rifampicin on EFV||PK parameter|
|Short term studies|
|Lee 2013||33||HIV-||UK||Single -dose||-45%||C24h|
|Cho 2011||10||HIV-||USA||Single -dose||-54%||AUC|
|Yenny 2011||8||HIV-||Indonesia||Single -dose||-39%||AUC|
|Lopez 2002||8||HIV+||Spain||7 days||-22%||Cmin|
|Kwara 2011||11||HIV-||USA||8 days||-19%||C24h|
|Bienvenu 2014||21||HIV+||Rwanda||2-6 weeks||-34%||mid-dose conc.|
|Orrell 2011||34||HIV+||South Africa||4 weeks||+9%||mid-dose conc.|
|Semvua 2013||21||HIV+||Tanzania||8 weeks||+11%||Cmin|
|Luetkemeyer 2013||19||HIV+||International||24 weeks||+6%||Cmin|
|Freidland 2006||20||HIV+||South Africa||24 weeks||+26%||Cmin|
|Cohen 2009||17||HIV+||South Africa||24 weeks||+22%||mid-dose conc.|
The authors also noted two additional longitudinal studies. In one the differences in EFV Cmin between rifampicin treated and untreated patients were only seen in the first 1 to 4 weeks of co-treatment, with no significant effects during longer-term treatment (Mukonzo 2013). And in the other rifampicin co-treatment only reduced EFV concentrations significantly during the first week, but not afterwards (Ngiamisi 2011).
In the systematic review, they found that rifampicin only lowered EFV concentrations in the first 1 to 4 weeks of treatment. There were no reductions in EFV after levels 4 to 24 weeks of combined treatment.
They wrote: “The reason for the lack of long-term effect of rifampicin on EFV in several large studies, despite clear reductions in short-term studies, is unclear.” They suggested that the populations might differ between the studies, or rifampicin might only lower EFV concentrations in the short time before auto-induction of EFV metabolism has occurred.
Based on these results, the authors concluded that dose modification of EFV does not appear to be justified, when co-administered with rifampicin-based treatment.
Two PK studies will look at EFV concentrations with 400 mg in the presence of rifampicin and in the third trimester of pregnancy.
- Dooley KE et al. Phamacokinetics of efavirenz and treatment of HIV-1 among pregnant women with and without tuberculosis coinfection. J Infect Dis. Advance access published 5 September 2014.
- Hill A et al. Does pregnancy affect the pharmacokinetics of efavirenz? AIDS. 2014 Jun 2014 19;28(10):1542-3.
- Hill A et al. The drug interaction between rifampicin and efavirenz is time- dependent: systematic review of 12 pharmacokinetic studies. 20th International AIDS Conference. Melbourne. 20-24 July 2014. Poster abstract MOPE040.
- McIlleron H et al. Efavirenz (EFV) concentrations in pregnant women taking EFV-based antiretroviral therapy (ART) with and without rifampin-containing tuberculosis (TB) treatment: the TSHEPISO Study Team. 19th International AIDS Conference. Washington DC. 22-27 July 2012. Oral Abstract MOAB0303.
- Clayden P. Pharmacokinetics of old and new TB drugs. HTB South. 13 August 2012.