Non-vitamin K antagonist oral anticoagulants (NOAC): considerations on once- vs. twice-daily regimens and their potential impact on medication adherence

We read with great interest the two letters1,2 concerning our paper entitled ‘Non-vitamin K antagonist oral anticoagulants: consideration on once- vs. twice-daily regimens and their potential impact on medication adherence’.3 We are pleased to see that our text stimulates more profound academic discussion on medication adherence. Many noteworthy points were raised, such as the fact that limiting the total pill burden in polymedicated atrial fibrillation patients by itself may have an impact on adherence; that adherence is only one of possible explanations to explain the PLATO results (although it is the factor on which we have the least information); that after forgetting a once-daily drug there is a longer interval to catch up (but the patient needs to be aware of the forgotten dose in the first place!); and that it is unclear what PK/PD parameters are most important for the antithrombotic non-vitamin K antagonist oral anticoagulant (NOAC) effect. It all underscores our ascertainment that many factors may play a role in the ultimate benefit/risk profile of once- vs. twice-daily drugs and that without reliable measurement and assessment, we will continue our guessing, and discuss based on opinion. Future clinical studies should address these issues by reliably measuring adherence and persistence, i.e. with electronic monitoring. Nevertheless, we would like to make some comments on other issues raised in both letters. Coleman et al.4 have shown indeed that when adherence to a dosing regimen is reported as an average percentage of prescribed pills taken, there is a 6.9% difference in favour of once-daily compared with twice-daily dosing regimens. However, such percentages are no sound pharmacometrical expression for concentration- and time-dependent drug actions.5 The key lesson from the Coleman et al. review4 is that even the use of once-daily regimens does not guarantee perfect adherence, with optimal adherence of once-daily dosing regimens ranging from 76.9 to 93.0%. In other words, there is evidence that among ambulatory patients prescribed once-daily dose cardiovascular drugs, 1 in ∼14 to even 1 in ∼4 doses have been omitted. Therefore, it is essential to understand the pharmacometric consequences of missing one or more (consecutive) doses, which was a key point of our paper. Timing adherence is certainly another important metric to consider, especially when estimating drug exposure as input to PK/PD models. However, we do not think that once- and twice-daily dosing regimens can be compared using the same metric (i.e. percentage of doses taken within an assigned interval). The definition of assigned intervals varies across studies, but, often, is defined as allowing a 25% deviation in timing (i.e. 6 h for once-daily and 3 h for twice-daily). This threshold is inappropriate as it assumes that for twice-daily dosing, the prescription requires dosing every 12 h. In fact, given the higher trough plasma levels, timing deviation may be less critical for twice-daily than that for once-daily NOAC with similar half-lifes (as we explained in our paper). Furthermore, giving less penalty (i.e. 6 h window) for once-daily dosing compared with only 3 h for twice-daily dosing does not make pharmacometric sense. Therefore, the claimed superiority of 22.9% for once-daily dosing compared with twice-daily is a mere arithmetic computation but is not necessarily medically relevant. Basic pharmacology principles teach us that the pharmacokinetic profiles of NOACs do differ between once- and twice-daily dosing in terms of Cmax and Ctrough, even in the study cited by Adler and Nagler: for a given total daily dose of rivaroxaban, Cmax was consistently higher, and Ctrough was consistently lower in patients receiving the full dose once-daily compared with patients receiving half the dose twice-daily.6 Therefore, the peak to trough ratio is smaller with twice-daily dosing compared with once-daily dosing. Whether this translates into more or less net clinical benefit remains a topic of discussion in the NOAC field. But it is a consideration to keep in mind, and it should spur more research. As long as better insights are lacking, we agree with the statement of Baker and Coleman that we should strive to replicate the dosing scheme as used in the randomized trials to achieve the proven results. Concerning the comment on using the HIV example is: the projections in Figures 2 and 3 in our paper are calculated as a pharmacokinetic model that is based on NOAC properties. This model is not derived from HIV or any other medical condition. The two examples of HIV and ACS were given for illustrational purpose only. We agree that the projections are theoretical. But again, it was the main aim of our paper to highlight the pharmacometric complexity beyond simple measures of ‘number of pills taken’ and even ‘timing of pills taken’. For the HIV example, the authors point out as an issue that the superiority in clinical outcome for twice-daily drug intake was shown in a subgroup only.7 It does not negate the finding, on the contrary. Dosing regimen advantages should not necessarily be expected in the average patient, or the very well adherent patient. But if there is increased effectiveness in those at higher risk or showing suboptimal adherence, this is an important consideration for the choice of treatment. We do not state that this is the factual case in NOAC-treated patients, but it certainly deserves study. On that last count, that clinical studies addressing the important issue of adherence to NOACs (and to cardiovascular drugs in general) are urgently needed, we all agree!