Efavirenz (EFV) is primarily metabolized by cytochrome P450 (CYP) 2B6 and to a lesser extent by CYP3A4. Drugs that significantly inhibit or induce these enzymes would then be expected to increase or lower the levels of EFV potentially resulting in toxicity or therapeutic failure, respectively. The constituents of Ginkgo biloba extract have been demonstrated to induce gene expression of the CYP450 enzymes. We report a case in which a potential drug–herb interaction may have led to virological breakthrough in a patient that was maintained on the same regimen for 10 years. Therefore, a drug–herbal interaction may be expected when these agents are taken concurrently, and we advise clinicians to avoid this combination when possible.
Case Report
A 41-year-old male, otherwise healthy, was diagnosed with human immunodeficiency virus (HIV) in 1999 and started antiretroviral therapy that same year with zidovudine (ZDV), lamivudine (3TC), and efavirenz (EFV). His baseline viral load and CD4 counts in May 1999 were 18 002 copies/mL (Chiron 3 assay) and 350 cells/mm3, respectively. Four months after starting therapy, his viral load was suppressed at less than 50 copies/mL. He continued on this same antiretroviral regimen for the next 10 years with excellent adherence and sustained virologic suppression. He developed no new medical illnesses and did not take any new medications on a regular basis during that time. In June 2010, his viral load became detectable at 1350 copies/mL (Chiron 3.0 assay), while his CD4 count was 1266 cells/mm3. On direct questioning, the patient denied any missed doses but revealed that he recently started the use of several herbal supplements. Two months prior, he started taking daily supplements of omega-3 fatty acid, calcium, magnesium, vitamin D, a multivitamin, flax oil, rutin, and 300 mg of Ginkgo biloba per day. Additionally, horse chestnut was being used periodically for hemorrhoidal flares for only a few days at a frequency of once or twice monthly.
Multiple attempts to genotype the June 2010 viral load sample were unsuccessful because of technical limitations of the polymerase chain reaction process as resistance testing may fail in up to 2% of specimens.
1
Attempts were also made to retrospectively measure EFV serum concentrations before, during, and after Ginkgo usage; however, adequate stored plasma samples were not available.
Recognizing that there had been a published report on the addition of G biloba extract (GBE) leading to subtherapeutic serum concentrations of EFV,
2
the patient was advised to stop using GBE because of concerns of a possible negative drug interaction. Having limited efficacy and drug interaction data, we also advised the discontinuation of the periodically used horse chestnut. The patient continued on the other supplements as they were not known or expected to be involved with CYP450 biotransformation. We did not measure EFV serum concentrations at this time because of the lack of therapeutic drug monitoring at our facility. Repeat blood work on our patient 1 month after discontinuation of the herbal supplements showed the viral load re-suppressed at less than 50 copies/mL. All subsequent viral loads have thus far remained undetectable, and the patient has not resumed taking either supplement.
Discussion
Our patient experienced virological breakthrough after 10 years of being on the same antiretroviral regimen, and we believe this occurred secondary to a drug interaction caused by a negative interaction of G biloba on EFV.
Efavirenz is a nonnucleoside reverse transcriptase inhibitor that has been demonstrated to have numerous pharmacokinetic interactions with other drug agents. Efavirenz itself is a substrate of CYP450 and may be the object drug of a pharmacokinetic interaction. It is primarily metabolized by the CYP2B6 isoform and to a lesser extent CYP3A4.
3
More recently, it has been recognized that an added complexity of its metabolism involves a unique secondary pathway with CYP2A6.
4
Drugs that significantly inhibit or induce these enzymes would then be expected to increase or decrease the exposure of EFV, respectively.
Ginkgo biloba extract is an herbal supplement that contains numerous constituents, namely flavonoids (quercetin, kaempferol, and isorhamnetin) and terpenoids (bilobalide and ginkgolides A/B/C/J).
5
In vitro, GBE has been demonstrated to activate the pregnane X receptor, leading to increased target gene expression, notably CYP3A4, CYP3A5, and P-glycoprotein.
5
Animal data have demonstrated the terpene trilactones and flavonoids to induce P450, and bilobalide to induce the activity, protein, and messenger RNA expression of CYP3A1 and 1A2, and increase CYP2E1 activity and CYP2B1/2 protein expression in a dose-dependent manner.
6,7
In humans, however, there are inconsistent data pertaining to the effect of GBE on the CYP450 enzymes, which may be related to study design, dose, intake period, and the varying amounts of GBE active constituents used.
8,9
However, in a more recent study of healthy volunteers where midazolam was used as a probe drug to represent CYP3A4 substrates, serum concentrations of midazolam were shown to be decreased after 4 weeks of exposure with GBE 240 mg daily, suggesting CYP3A4 induction.
10
With respect to the CYP2B6 isoenzyme, while there is in vitro data that demonstrates inhibition by G biloba,
11
pharmacokinetic studies in healthy adults showed that serum concentrations of the 2B6 probe drug bupropion remained unchanged after 14 days of concurrent GBE.
12
Given the ability of GBE to upregulate the specific enzymes responsible for the drug metabolism of EFV, we propose that the addition of Ginkgo would accelerate the hepatic clearance of EFV resulting in lower EFV serum concentrations and consequently lead to viral breakthrough. It is this mechanism that Wiegman et al also propose as the cause for breakthrough viremia in their patient who was on an EFV-based regimen for 2 years until the coadministration of GBE. Unfortunately, in their case, virological failure ensued forcing a change in antiretroviral regimen. Thus, they were unable to remeasure EFV levels following GBE withdrawal to confirm that a negative drug interaction with Ginkgo was the cause of the decreased EFV concentrations. Although we were unable to measure EFV serum concentrations in our case, we do believe that the time frame of events provides good support for this interaction. Our patient remained virologically suppressed on his original regimen for over 10 years which supports both a history of excellent adherence and a lack of baseline resistance. Only after the concomitant use of GBE, he developed viremia. One month after discontinuing the presumed interacting agent, his viral load then re-suppressed to undetectable levels. Since the constituents of G biloba have short pharmacokinetic half-lives, the stimulus for induction of P450 enzymes would be removed within the first day of stopping GBE,
13
which supports the observed offset of this interaction.
It has been recognized that HIV-positive patients on successful antiretroviral therapy may periodically experience episodes of viremia which are not explained by drug failure or nonadherence. It is postulated that these viremic episodes represent activation of the latent viral reservoir and are termed “blips” as repeat testing often results in an undetectable viral load. While this episode of virologic breakthrough could be argued to have been a blip, this magnitude of viremia would represent an uncharacteristically large blip as it is usually not greater than 1000 copies/mL.
14
–16
For example, Podsadecki et al studied 223 patients with viral load blips and report a median of 82 copies/mL with a range of 51 to 858 copies/mL.
Using the Drug Interaction Probability Scale proposed by Horn et al to evaluate drug interaction cases,
17
our proposed causal relationship of virological breakthrough due to a drug interaction between G biloba and EFV was calculated as probable. Positive findings such as the existence of a previous report, the interaction being consistent with known properties of both agents, the time line of events being reasonable, and the event resolving upon de-challenge, all support our case. In addition, because the horse chestnut supplement was also withheld during this time frame, we cannot rule out its influence on the biotransformation of EFV. However, in vitro data have shown horse chestnut to inhibit CYP3A4,
18
without having any inducing effect on CYP1A2, 2C19, 2D6, 2E1, and 3A4.
19,20
Therefore, an interaction with the infrequently used horse chestnut and EFV leading to our patient’s virologic breakthrough is doubtful.
Conclusion
There are limited drug interaction studies with G biloba and antiretroviral drugs in HIV-positive individuals. Although pharmacokinetic studies are needed to fully elucidate the effect that Ginkgo has on the various antiretroviral medications, we believe our case adds to the scant literature, supporting Wiegman et al, suggesting an inducing effect on EFV metabolism leading to virological breakthrough. In this case, recognizing the potential negative interaction and discontinuing the offending agent early may have avoided the development of drug resistance and therapeutic failure. As there are several antiretroviral agents that undergo a similar metabolic pathway to EFV, formal pharmacokinetic studies would be helpful to fully evaluate the effect Ginkgo may have on other HIV medications. Until additional data are available, we suggest caution be used if Ginkgo is coadministered with such agents and to consider therapeutic drug monitoring if available. Finally, given the numerous reports of G biloba interacting with other drugs,
21
clinicians must continue to be vigilant for the possibility of significant drug interactions whenever herbal products and medications are concurrently used.
