When Do You Start Warfarin Treatment and How Do You Monitor It?

By Eric A. Dietrich, PharmD, and Louis Kuritzky, MD 

In our last discussion, we covered the basics of antithrombotic therapy with warfarin and presented a typical case of an 85-year-old hypertensive woman with noted atrial fibrillation 3 weeks prior. Her arrhythmia (heart rate of 110 bpm) causes minimal symptoms. In discussions between the doctor, the patient, and her sister, warfarin was decided as the course of treatment. How do we start, how and when do we monitor, and why?

Warfarin and Atrial Fibrillation
Warfarin Initiation and Monitoring


Clinical trials have shown that 5 mg/d is generally an acceptable starting dose, and that checking an international normalized ratio (INR) 5 to 7 days after initiation is reasonable in most uncomplicated situations (no interacting drugs, patients younger than age 65). Because there is no “one-size-fits all” dose for warfarin, one should take into account other factors that increase bleeding risk. 

In the past, some clinicians have favored a “loading dose” of warfarin in an effort to hasten the attainment of a therapeutic level of anticoagulation. Although this can indeed be effective, it also increases the risk of a supratherpeutic INR. In our experience, when clinicians have chosen the loading dose method the INR measured at day 7 of dosing and beyond often increases substantially beyond what was sought—necessitating dose reduction and placing the person at bleeding risk until the INR is successfully reduced down into the therapeutic range. Hence, the standard 5 mg/d dosing is preferred. When urgency of anticoagulation is needed, parenteral heparinoids such as a low-molecular weight heparin (discussed in future cases) are appropriately employed for essentially immediate therapeutic effect and continued until a therapeutic INR is attained.

Genetic markers may one day soon help physicians predict which patients should start on higher (or lower) doses of warfarin. Until such testing has become firmly established in clinical care, it will be more sensible to use experience-derived administration dosing pathways of initiating 4 or 5 mg daily. Recently published trials employing genetic testing methodologies have not shown convincing evidence of superior outcome, hence routine use in clinical care is not yet recommended.1

A Case Study

The fact that Agatha is older (85 years old) likely increases her risk for bleeding. A younger person (age <65) could probably wait until 1 week to check the full impact of warfarin. It is wise to be cautious in older patients, especially those with a CHADS of 4 or greater, and/or other factors (ie, interacting medications, changing medications [stopped and/or started], higher baseline CHADS2 score) that may increase the risk for bleeding. In such situations, it may be prudent to check the INR after the first 3 to 4 days. 

On Monday, our patient was started on 5 mg/d, to be taken at night. 

On Thursday morning, day 4 of warfarin administration, Agatha has returned for her INR check. She is experiencing no bleeding or adverse effects. Her INR is 1.4 today. That is a very common result at this point. She is neither excessively sensitive nor insensitive to warfarin, so we recommend that she return at day 7 to measure the full impact of the drug on her INR to determine if any dose adjustments are required to attain a therapeutic INR. Remember, the full therapeutic effects of warfarin may not become apparent until 7 days after warfarin administration so we would not expect to see a significant increase in her INR after only 3 doses.

Common Scenarios

By far, the most common scenario is that Agatha reports an INR <1.6. However, let’s take a minute to reassess what we would do if Agatha reported a higher INR. 

1. INR 1.7-1.9. We have some choices to make here, depending upon how easy it will be for Agatha to access INR monitoring over the next few days (given the upcoming weekend). Since at day 4 she is already establishing a substantial antithrombotic response, and we know that it is likely to grow still greater by day 7 (maximal therapeutic effect at this dose), it may be wise to slightly reduce her warfarin dose to 4 mg/d. 

Warfarin is inexpensive ($10/month or less), and comes in a diverse dose range with each dose identified by a different color. Agatha has already purchased warfarin 5 mg tablets, but we now ask her to obtain some 4 mg tablets to slightly decrease her dose. Note: The AT9 guidelines account for this possibility by including a 4 mg/d dose as an alternative initiating dose for persons of advanced years due to increased risk of bleeding.

Since Agatha is taking her pills at night per our recommendation, we can give her 4 mg/d from Thursday through Sunday. When we retest her levels on Monday morning, the lower dose is less likely to have caused a supratherapeutic INR than the 5 mg/d dosage.

Keep in mind that we would not have known that had the day 4 INR been checked. If day 4 INR had occurred on a Monday (as opposed to Thursday mentioned above), it will probably not be difficult to obtain another INR on day 7 (which would be a weekday) when maximal effect is anticipated. In that circumstance, it is clinician preference about whether to lower the warfarin dose or simply check INR again on day 7. 

The Take Away: Availability of monitoring motivates dose choice.

2. INR 2.0-3.0. The choice here is very similar to the scenario discussed earlier (robust INR response to warfarin). The rationale is the same as in scenario 1 but given the greater INR excursion, it is probably best to lower the dose to 4 mg/d rather than continue at 5 mg/d. This option will be irrespective of the day of the week and the INR should be rechecked on day 7 (or as close to day 7 as possible) to ensure it has not continued to rise. Again, as Agatha takes her medication at night the new lower dose can be started the same night as the INR was measured.

3. INR >3.0. The INR is supratherapeutic. This patient is very sensitive to warfarin, and likely needs dose adjustment. Years of experience with new-start patients on warfarin have found very few patients with an INR much above 3.5 within the first week of therapy. This patient has thus far received a total dose of 15 mg, and is supratherapeutic. It is likely that she may require a weekly dose totaling less than 15 mg. 

There was no way to predict this response in advance in the absence of genetic testing. In any case, omitting 1 dose and reducing subsequent dosing to 2 mg/d (which would lead to a total weekly dose of 14 mg) is reasonable. Remember, the patient has not yet taken her dose today, so we omit today’s dose of medication. 

Patients with this marked sensitivity to warfarin require very frequent monitoring until a stable dose is achieved. Fortunately, these patients are a distinct minority of cases. 

What’s the Take Home?

1. There is no “one-size-fits-all” dose for warfarin. Nonetheless, clinical trials have shown consistent success using the recommended dosing for warfarin is 5 mg/d without a loading dose.1 

2. A loading dose of warfarin does accelerate the time to attainment of a therapeutic INR, but unfortunately also increases the risk for a supratherapeutic level—placing the patient at risk of bleeding and requiring downward dose adjustment.

3.  Young (age <65) otherwise healthy persons with lower CHADS scores (<4) can likely have their initial INR check safely performed at 1-week post initiation.

4. When a patient has characteristics that predict increased risk of bleeding complications (age greater than 65 years, CHADS2 score greater than 4), we recommend an early INR check in most situations (day 3-4) to begin to establish the level of the patients’ sensitivity to warfarin in an effort to limit supratherapeutic values measured at the first INR check. ■

Eric A. Dietrich, MD, graduated from UF College of Pharmacy in 2011 and completed a 2-year fellowship in family medicine where he was in charge of a coumadin clinic. He now works for the UF Colleges of Pharmacy and Medicine. 

Louis Kurtizky, MD, is a family physician affiliated with the University of Florida Family Medicine Residency Program, where he commonly co-manages warfarin cases with his colleagues.


1.  Antithrombotic Therapy and Prevention of Thrombosis. 9th ed. CHEST Evidence-Based Clinical Practice Guidelines. 2012;141(2 Suppl).