Table 2: Criteria for defining evidence reliabilities of an interaction [14, 16].

Reliabilities of evidenceCriteria required

I: highly probableA, B, and C plus any one or more of D to G
II: probableA, B plus one or more of C to G
III: possibleA plus one or more of B to G
IV: doubtfulAny combination of B to G or A alone

(A) Was the timing correct for an interaction to be pharmacologically plausible?
(B) Did laboratory test (INR, prothrombin time, and thrombotest) support the contention of an interaction?
(C) Were other potential factors affecting warfarin pharmacokinetic or pharmacodynamics ruled out?
(D) Was there other objective evidence?
(E) Was a dose-response relationship shown for the interacting herb?
(F) Was the subject rechallenged and, if so, did a similar response occur?
(G) Did the same thing happen on previous exposure to the herb?

A: in patient-based studies, warfarin must been taken at a stabilized dose before initiation of the interacting herbs. In addition, the potentially interacting herbs should be consumed long enough in usual doses to attain a substantial plasma level. For volunteer-based studies, subjects had to receive warfarin, both alone and with the interacting herbs, for similar periods.
B: in patient-based articles, the coagulation variable should be out of therapeutic range, whereas for volunteer studies, a change of at least 20% was required in coagulation parameters. For conclusion of “no interaction,” the absence of a statistically significant change in coagulation variables was required.
C: factors such as diet, other medications, or certain medical conditions, especially liver diseases, should be declared to be ruled out as possible causes of the outcome.
D: other objective evidences refer to changes in plasma level of warfarin or level of vitamin K dependent clotting factors (II, VII, IX, or X).
E: the alterations in the dose of the coadministrated interacting herbs correlated with subsequent changes in warfarin coagulation variables, inferring a dose-response relationship.
F: the interacting herb should be administered simultaneously with warfarin in two or more separate courses, with similar results for each course.
G: similar outcome should be found for the patient cousing the interacting herb with warfarin at a time prior to that reported.