On 2014 Mar 06, David Keller commented:

Table 4 is a valuable reference for primary-care physicians

Clinicians who prescribe antibiotics to patients taking warfarin will want to review the results of this study, which are summarized in Table 4. The antibiotics with the highest rates of interaction resulting in INR > 5.0 are as follows (the name of each antibiotic is followed by its rate of adverse interaction):

1) Metronidazole 10.5%

2) Moxifloxacin 9.2%

3) Timethoprim-Sulfamethoxazole 7.1%

while the best-behaved antibiotic was cefuroxime axetil, with 0 interactions. See Table 4 for the complete list & confidence limits. Clinicians might wish to print it out for convenient reference.

An accompanying editorial advises, on the basis of the results of this study, that clinicians monitor INR more frequently while the patient is taking an antibiotic which interacts significantly with warfarin, but that pre-emptively lowering the dose of warfarin is not necessary. Other options may include bridging therapy with enoxaparin or fondaparinux; however, these parenteral anticoagulants are not approved for the prophylaxis of embolism due to atrial fibrillation. The prescriber could consider bridging therapy with one of the newer oral anticoagulants which does not interact as strongly with the antibiotic prescribed to the patient. However, the drawbacks to these medicines are that they are still subject to emesis by a nauseated patient, and they are costly, and they may cause fluctuations in anticoagulation activity when switching to or from warfarin. So, we are left with maintaining the patient's chronic warfarin dose unchanged during antibiotic therapy, but with a shorter interval between INR tests.

Perhaps a follow-up study can be designed to determine how short the INR testing interval should be for each antibiotic, such that no more than 1% of the patients taking that antibiotic plus warfarin ever experience an INR over 5.

I suspect that for antibiotics with the strongest interactions with warfarin, even the shortest practical INR testing intervals (daily? every-other-day?) will not be sufficient to drive down the rate of dangerous interactions low enough, and that preemptive lowering of the maintenance dose of warfarin will be found to be necessary in order to safely treat concomitantly with antibiotics such as metronidazole and moxifloxacin.

On 2014 Mar 06, David Keller commented:

Table 4 is a valuable reference for primary-care physicians

Clinicians who prescribe antibiotics to patients taking warfarin will want to review the results of this study, which are summarized in Table 4. The antibiotics with the highest rates of interaction resulting in INR > 5.0 are as follows (the name of each antibiotic is followed by its rate of adverse interaction):

1) Metronidazole 10.5%

2) Moxifloxacin 9.2%

3) Timethoprim-Sulfamethoxazole 7.1%

while the best-behaved antibiotic was cefuroxime axetil, with 0 interactions. See Table 4 for the complete list & confidence limits. Clinicians might wish to print it out for convenient reference.

An accompanying editorial advises, on the basis of the results of this study, that clinicians monitor INR more frequently while the patient is taking an antibiotic which interacts significantly with warfarin, but that pre-emptively lowering the dose of warfarin is not necessary. Other options may include bridging therapy with enoxaparin or fondaparinux; however, these parenteral anticoagulants are not approved for the prophylaxis of embolism due to atrial fibrillation. The prescriber could consider bridging therapy with one of the newer oral anticoagulants which does not interact as strongly with the antibiotic prescribed to the patient. However, the drawbacks to these medicines are that they are still subject to emesis by a nauseated patient, and they are costly, and they may cause fluctuations in anticoagulation activity when switching to or from warfarin. So, we are left with maintaining the patient's chronic warfarin dose unchanged during antibiotic therapy, but with a shorter interval between INR tests.

Perhaps a follow-up study can be designed to determine how short the INR testing interval should be for each antibiotic, such that no more than 1% of the patients taking that antibiotic plus warfarin ever experience an INR over 5.

I suspect that for antibiotics with the strongest interactions with warfarin, even the shortest practical INR testing intervals (daily? every-other-day?) will not be sufficient to drive down the rate of dangerous interactions low enough, and that preemptive lowering of the maintenance dose of warfarin will be found to be necessary in order to safely treat concomitantly with antibiotics such as metronidazole and moxifloxacin.