Background:  Subtheraputic levels of protease inhibitors (PI) at standard doses in pregnancy have been reported for nelfinavir (NFV) and lopinavir/ritonavir (LPV/r). Favorable pharmacokinetics for saquinavir/ritonavir (SQV/r) in pregnancy make it a popular choice in the prevention of mother-to-child transmission (MTCT) of HIV. Previous studies in non-pregnant cohorts describe hepatotoxicity rates of 9 to 17% for this combination. This study prospectively evaluates rates of hepatotoxicity of SQV/r-based regimen in a pregnant cohort and assess pharmacokinetics of SQV in the third trimester of pregnancy.

Methods:  From February 2005 to September 2006, 49 pregnant women in their third trimester began taking zidovudine (AZT) + lamivudine (3TC) + SQV/r at standard doses (1000 mg/100 mg twice daily) to prevent MTCT of HIV. Baseline CD4, HIV viral load, full blood count, renal profile, liver function tests were recorded. Routine therapeutic drug monitoring in 18 women was taken 14 days after starting therapy in the third trimester and repeated if any dose adjustment was required. Transaminases were measured at 2, 4, and 8 weeks (30, 32, and 36 weeks’ gestation) on therapy. Hepatotoxicity was graded according to the AIDS Clinical Trials Group (ACTG) standards as grade 1-4.Women with baseline abnormal liver function (n = 4) and those co-infected with viral hepatitis (n = 3) were excluded from analysis. Drug monitoring samples were taken at time 0 and and 1 hour post-ingestion of drug for trough and peak levels. The estimated minimum trough concentration required for wild type HIV is 100 ng/mL.

Results:  Of 42 women, 13 (31%) developed abnormal transaminases on treatment:

 8 women developed grade 1; 4 developed grade 2; and 1 developed grade 3 transaminitis within 2 to 4 weeks of ART initiation. ART was changed in 5 patients as a result of hepatotoxicity. All transaminases normalized 6 weeks post-partum. At time of submission, drug monitoring in 18 women (42.8%) showed significant variability in SQV levels, with mean trough of 1561.42 ng/mL (range 278 to 5134 ng/mL), and mean peak of 2458.98 ng/mL (range 352 to 5763 ng/mL). These results precipitated no adjustments to any dosage. There was no correlation between SQV trough levels and hepatotoxicity.

Conclusions:  While on SQV/r, 31% (13/42) of HIV-seropositive pregnant women developed abnormal tranaminases. A switch in ART was required in 5 patients. There was no correlation between SQV trough and hepatoxicity. More frequent monitoring of liver function is recommended in all women prescribed protease inhibitors during pregnancy.