October 10th, 2017

Podcast 214: Drug-drug interactions and bleeding risks with NOACs

The non-vitamin-K oral anticoagulants (known familiarly as NOACs or DOACs) share metabolic pathways with other drugs, which can potentiate NOACs’ anticoagulant actions dangerously. Dr. Shang-Hung Chang and his group studied Taiwan’s national health insurance database, which records data on virtually all that nation’s citizens, to measure the actual risks of some of these drug – drug interactions. Their findings were published earlier this month in JAMA.


JAMA article (abstract)

Physician’s First Watch coverage

6 Responses to “Podcast 214: Drug-drug interactions and bleeding risks with NOACs”

  1. Chang, Bok Ryun says:

    How about Valproate, or Valproic acid, which inhibits coagulation of platelets? Also it suppress the production of platelets ,too.

  2. Joe Elia says:

    I’ve asked Dr. Chang to respond to the question, but in the meantime (and with the caveat that your host is not a clinician) I will say that the drugs studied all share common metabolic pathways with the NOAC group. Valproic acid apparently doesn’t.
    In addition, valproate/valproic acid doesn’t occur in the article or its references.

    I know I’ll be corrected if I’m wrong on this.

    Joe Elia

    • Chang, Bok Ryun says:

      Valproate is not directly related to coagulation cascade by interfering thrombotic factor. However, according to the Oxford Neurology Monograph, Epilepsy, valploate inhibits the coagulative ability of platelets even if the number of platelets are within standard value.

  3. Chang says:

    It’s plausible that most of platelet inhibitors might increase the risk of bleeding. The 12 medications checked in this study, however, were picked up for their shared metabolic pathways with NOACs, not for platelet inhibition.

  4. Christopher Giuliano says:

    The results of this article are surprising to the point where I believe some errors must have been made with the data analysis. The medications that were shown to have the highest bleeding risk were both inhibitors (amiodarone, fluconazole) AND inducers (phenytoin, rifampin) of CYP3A4 and/or P-glycoprotein. Inhibitors should increase NOAC concentrations and therefore increase the bleeding risk, which was seen in this article. Inducers should decrease concentrations of NOACS, therefore reduce the bleeding risk. Phenytoin has consistently shown in many studies to reduce drug levels, it doesn’t make sense this would increase bleeding risk. This suggests an error in data analysis, selection bias, or residual confounding is present. I struggle with using the results of this study without further confirmation of results. The authors do not comment on these contradictory findings and I am concerned these findings will be utilized extensively.

  5. Chang says:

    Thanks for the comments. Drug-drug interactions are complicated in real world and beyond the serum levels, which are unknown in this study. The paradox of amiodarone and atorvastatin observed in this analysis is a good example. The comorbidities and co-medications interact with bleeding risks in a complex way.

    Residual confounding and selection bias, however, should always be kept in mind when interpreting associations. This is more important in the cases with smaller sample size such as phenytoin here, although we’ve put many cofactors into the models.

Clinical Conversations

About the Podcast

Comments, suggestions, and story ideas welcome. Learn more about Clinical Conversations.

Follow us:
Follow on Apple Podcasts

Listen on Youtube

Follow on Spotify