Interactions between African herbal medicines and ARVs

Polly Clayden, HIV i-Base

A study published in the 3 January issue of AIDS evaluated the potential effects on antiretroviral metabolism of two herbal medicines commonly used in South African traditional medicine and identified the potential for clinically significant drug interactions for both H. hemerocallidea and Sutherlandia.

As part of national treatment rollout, the South African Department of Health has recently accredited 27 facilities, to provide HIV/AIDS care including ‘interventions that delayed the progression of the disease, including nutritional and micronutrient supplementations, and providing complementary and traditional medicines’

Some herbal medicines have been shown to affect the serum levels of antiretroviral medication through their effects on CYP3A4 metabolism and P-glycoprotein. Additionally some herbal medicines are also known to interact with nuclear receptors such as the pregnane X receptor (PXR), which modulates the expression of CYP3A4, and P-glycoprotein.

The authors analysed the effect of two herbs in common medical use for HIV treatment in Africa, Hypoxis hemerocallidea(African potato) and Sutherlandia, for their potential to cause drug interactions with common antiretroviral agent metabolising mechanisms in vitro.

Capsules, tablets or teas of each herb were water extracted and methanol extracted and tested for the ability to inhibit CYP3A4. The effects of the herbal medicines on PXR and P-glycoprotein were also investigated.

Water extracts of Hypoxis showed a significant 33.9% – 85.6% inhibition of CYP3A4 activity at an initial concentration of 100 mg/ml, with an increase in inhibition to 56.1% -79.4% seen with the methanol extract.

Sutherlandia showed even greater inhibition of CYP3A4 with both water and methanol extracts at 100 mg/ml, exhibiting almost complete inhibition with the methanol extract, 64.5% – 68.5% and 70.3% – 95.7% inhibition respectively.

Exposure to both Hypoxis and Sutherlandia resulted in significant activation of PXR (Hypoxis p= < 0.05 and Sutherlandia p= < 0.01). Additionally both herbal medicines showed moderate activity of the P-glycoprotein enzyme.

The authors write: “The combination of these findings suggest that the co-administration of these drugs with antiretroviral agents may result in the early inhibition of drug metabolism and transport followed by the induction of decreased drug exposure with more prolonged therapy.” And they continue [these findings] “…underscore the need for appropriately designed pharmacokinetic studies to unveil the true drug interaction potential of herbal drugs with antiretroviral agents.”

They conclude “…the South African Ministry of Health, along with member states’ and non-governmental organizations’ endorsement of traditional African herbs such as Hypoxis and Sutherlandia as HIV/AIDS remedies, policy makers and the research community should make investigations into the clinical ramifications of the use of these herbs on antiretroviral therapy an urgent priority.”


There are concerns about potential interactions in countries such as South Africa rolling out ARV programmes where use of traditional herbal medicines is widespread. The Foster study is of interest because it clearly shows the potential of the extracts of Hypoxis and Sutherlandia to modulate drug disposition.

David Back from Liverpool University advises, “…I would urge caution about over interpretation. These are in vitro studies with possibly high concentrations relative to what is seen in a patient. Also in vitro studies are notoriously difficult to correlate directly to changes in a patient. To me it shows the need for well designed small clinical studies of the interaction of herbals with ART.”

For information on HIV drug interactions, the Liverpool drug interactions website is an excellent resource – Liverpool drug interactions website


Mills E, Foster BC, van Heeswicjk et al Impact of African herbal medicines on antiretroviral metabolism. AIDS: Volume 19(1) 3 January 2005 p 95-97

Links to other websites are current at date of posting but not maintained.