The role of oral Vitamin K administration in the reversal of anticoagulation is not yet clear because of a paucity of data on the early effects of treatment, apparent differences in efficacy between preparations and a lack of data comparing oral with intravenous administration. We have compared the effects on the International Normalized Ratio (INR) and activities of the Vitamin K-dependent clotting factors II, VII, IX and X at 4 h and 24 h after administration of three oral Vitamin K preparations and of intravenous Vitamin K in 64 anticoagulated patients who required non-urgent partial correction of anticoagulation. Our data confirm that correction of anticoagulation is more rapid after intravenous administration of Vitamin K than after oral administration of similar or larger doses. At 24 h, satisfactory correction of INR can be achieved using low-dose Vitamin K given by either the intravenous or oral route. Our data, and that from previous studies, suggest that there may be differences in efficacy between orally administered products. Administration of Vitamin K by either route was accompanied by changes in the activities of the Vitamin K-dependent clotting factors that reflected their respective biological half-lives. In the 24 h after treatment, the relationship between the INR and the individual Vitamin K-dependent clotting factors was similar to that described previously in stable anticoagulated patients. We conclude that the reversal of anticoagulation with warfarin is achieved more rapidly by intravenous administration of Vitamin K. Satisfactory, but slower, reversal of anticoagulation can be effected using oral Vitamin K, but there may be differences in efficacy between the products tested in our study.
Despite innovations in the management of anticoagulation, even well controlled patients spend only 70% of the time in the therapeutic range and over-anticoagulation is a very common clinical problem (Poller et al, 1998). The presentation varies in severity from an asymptomatic attendance at an anticoagulant clinic to life-threatening haemorrhage. Although a significant proportion of major bleeding episodes occur with the International Normalized Ratio (INR) in the therapeutic range, the risk of haemorrhage, including spontaneous intracranial haemorrhage, rises markedly as INR values exceed 4·0–4·5 (Hylek & Singer, 1994; Palareti et al, 1996). Consequently, a range of effective strategies to reverse anticoagulation which take into account the severity of the clinical situation and therefore the urgency of reversal is required.
In this study, we examined the efficacy, including the rate of response, to three oral Vitamin K preparations currently available in the UK. The effects on the INRs and activities of the Vitamin K-dependent clotting factors in patients requiring non-urgent reversal or reduction of intensity of anticoagulation were established. The results were compared with those obtained in another group of patients, selected in an identical manner whose over-anticoagulation was reversed by 2 mg i.v. Konakion.
Patients and methods
Patients Sixty-four patients attending the Royal Hallamshire or Northern General Hospitals, Sheffield, Addenbrooke's Hospital, Cambridge, and Aberdeen Royal Infirmary, were studied. They were receiving warfarin for a variety of clinical indications, had INR values above the therapeutic range and were deemed by their clinicians to require some correction of anticoagulation with Vitamin K. All patients had no or minor bleeding only. Patients requiring immediate and complete correction of anticoagulation were excluded. One of three preparations of Vitamin K was administered orally. These were Orakay (BMS Laboratories, Hounslow, UK), 1 mg capsule; Menadiol sodium phosphate (Cambridge Laboratories, Newcastle-upon-Tyne) 5 mg (half 10 mg tablet) and Konakion (Roche, Welwyn Garden City, UK) 2 mg and 5 mg. Konakion was the preparation for injection (phytomenadione 10 mg/ml colloidal formulation) but administered orally. All of these preparations are available in UK hospitals for the reversal of oral anticoagulation, and the formulations and their respective doses used by participating hospitals were those routinely used for this purpose. This was the only selection criterion for the choice of Vitamin K preparation by the participating hospitals. Orakay was used exclusively at Cambridge and Konakion (5 mg) exclusively at Aberdeen. In addition, we studied an identically selected group of 12 patients who received a single intravenous injection of Konakion, 2 mg. These patients were recruited in one centre at an earlier time when administration of intravenous Vitamin K was the standard approach to partial revision of anticoagulation.
In all patients, a citrated blood sample was obtained immediately before and 4 and 24 h after administration of the Vitamin K. No warfarin was given in the 24-h after administration of Vitamin K. Prothrombin times for INR determination were performed on all samples using rabbit brain thromboplastins (Sheffield, Instrumentation Laboratory, ISI 1·1; Cambridge, Organon Teknika, ISI 1·26) and a recombinant human thromboplastin (Aberdeen, Ortho Diagnostic Systems, ISI 1·0). One-stage assays were used to measure factors II, VII, IX and X. All single factor assays were carried out in one laboratory. These were performed in all cases in which sufficient plasma was available from all three time points. All four participating centres perform satisfactorily in the UK National External Quality Assessment Scheme (UK NEQAS) for blood coagulation.
Six patients received 1 mg Orakay. Tables I, II and III and Fig 1) The median pre-treatment INR was 8·2 (range 5·1–14·4). At 4 h, the median INR was 5·5 (range 3·3–15·1) and 24 h after treatment the median INR was 4·1 (range 2·9–13·5). At 4 h, the INR was less than 4 in 1/6 (16·7%) of patients. At 24 h, the INR was less than 4·0 in 3/6 (50%) patients.
Table I. Median (range) INR values before treatment and 4 and 24 h after administration of Vitamin K.
Number of patients
4 h INR (range)
24 h INR (range)
Table II. Percentage of results in the specified ranges 4 and 24 h after the administration of different preparations of Vitamin K.
4 h INR
24 h INR
Table III. INR and individual coagulation factor levels after treatment with different Vitamin K preparations (median and range).
Orakay1 mg oral (n = 6)
Menadiol 5 mg oral (n = 4)
Konakion 2 mg oral (n = 4)
Konakion5 mg oral (n = 15)
Konakion 2 mg iv (n = 7)
Menadiol sodium phosphate, 5 mg
Eighteen patients received 5 mg menadiol sodium phosphate. The median pre-treatment INR was 9·3 (range 6·8–14·8). At 4 h, the median INR was 7·4 (range 4·1–16·6) and at 24 h, the median INR was 4·2 (range 1·5–8·4). At 4 h, none of the patients' INR was less than 4·0. At 24 h, the INR was less than 3·0 in 3/16 (19%) patients and less than 4·0 in 8/16 (50%) patients.
Konakion, 2 mg oral
Twelve patients received 2 mg Konakion. The median pre-treatment INR was 7·9 (range 3·6–20·0). At 4 h, the median INR was 7·5 (range 3·1–17·9) and at 24 h, 3·3 (range 1·5–11·1). At 4 h, 3/12 (25%) patients had an INR less than 4·0; at 24 h, 9/12 (75%) had an INR less than 4·0. At 24 h, four (33%) patients had an INR in the range 2·0–3·0 and four (33%) in the range 1·0–1·9.
Konakion, 5 mg oral
Sixteen patients received Konakion, 5 mg. The median pre-treatment INR was 10·6 (range 8·1–19·8). Four hours after treatment, the median INR was 10·0 (range 3·5–16·9) and at 24 h, 1·6 (range 1·1–6·8). At 4 h, one (7·7%) patient had an INR of less than 4·0. At 24 h, 14/15 (93%) had an INR less than 4·0; 2/15 had an INR in the range 2·0–3·0 and the remainder 12/15 (80%) in the range 1·1–1·9.
Konakion, 2 mg intravenous (i.v.)
Twelve patients received a single i.v. injection of 2 mg Konakion. The median pre-treatment INR was 12·4 (range 5·1–24·1). The median INR was 3·7 (range 2·7–12·0) at 4 h and 2·0 (range 1·4–2·8) at 24 h. At 4 h, the number of patients who had an INR less than 4·0 was 8/12 (67%) and at 24 h ,12/12 (100%). At 24 h, 6/12 (50%) patients had an INR within the range 2·0–3·0 and 6/12 (50%) within the range 1·4–1·9.
Comparison of correction of anticoagulation by Vitamin K by oral or i.v. route
The effect of route of administration of Vitamin K was assessed at 4 h and 24 h by comparing reversal of anticoagulation for all oral preparations versus intravenous Vitamin K. At 4 h, the chance of having an INR < 4·0 was 5/49 for all oral preparations and 8/12 for i.v. (x2 test P < 0·00001). At 24 h, 33/49 and 12/12, respectively, had INR values < 4·0 (x2 test P = 0·02). These data indicate that a significant difference existed between the different routes of administration.
Measurement of Vitamin K-dependent coagulation factors (Table III)
Levels of factor II, VII, IX and X activity were measured before treatment and 4 and 24 h after administration of Vitamin K in all patients for whom sufficient plasma was available. Twenty-nine patients treated with oral Vitamin K and seven who received intravenous Vitamin K were studied. Before treatment, the coagulation factor levels were as predicted, with slightly higher levels of factor IX compared with the other factors, as demonstrated in previous studies (Kumar et al, 1990). Four hours after administration of any oral preparation, in keeping with the minimal change seen in INR values, there was little or no increase in the activity of any of the Vitamin K-dependent factors. On the other hand, 4 h after administration of intravenous Vitamin K, there was a marked increase in the levels of all four in keeping with the observed fall in INR.
Twenty-four hours after administration of Vitamin K, a rise in coagulation factor levels was seen in all the treatment groups. Over the course of the first 24 h, the relationship between the coagulation factors was in the same pattern as that observed in patients who were stable on warfarin, with the factor IX and VII levels being higher than factor II and X levels (Kumar et al, 1990).
Clinical circumstances determine the rate at which correction of anticoagulation needs to be achieved. Rapid reversal is necessary for patients with serious bleeding, or before emergency surgery, whereas slower correction is acceptable for asymptomatic over-anticoagulated patients. In emergency situations, in view of the need for rapid correction of anticoagulation, the administration of coagulation factor concentrates is required in addition to Vitamin K (Makris et al, 1997). The degree of correction of the INR by Vitamin K is also important. Life-threatening haemorrhage demands complete reversal of anticoagulation, whereas the optimal reversal of anticoagulation in an asymptomatic patient brings the INR into the therapeutic range reasonably quickly and in a way that is safe and acceptable to the patient. Failure to correct over-anticoagulation prolongs the period of increased risk of haemorrhage, whereas over-correction results in temporary subtherapeutic anticoagulation with its attendant increased thrombotic risk (Hirsh et al, 1998). Clearly, these considerations must be borne in mind when selecting the route of administration, dose and preparation of Vitamin K appropriate to the clinical situation.
To date, most studies have recorded changes in INR 24–48 h after treatment with Vitamin K, and there is little available information on the important earlier effects. Comparison of published reports suggests that there may be differences between Vitamin K preparations and, that there is no consensus on the appropriate dose of oral Vitamin K for these indications (Weibert et al, 1997; Crowther et al, 1998; Penning-van Beest et al, 1999). Further, before we can safely recommend the use of oral Vitamin K in the emergency reversal of anticoagulation, it is appropriate to compare the early effects of administration by this route with the i.v. route.
In this study, we have demonstrated several points that are relevant to clinical practice. These are, first, that there is a marked difference in the rate at which anticoagulation is reversed when the i.v. route is compared with the oral route. Second, it is clear that anticoagulation can be reversed by low-dose i.v. or oral Vitamin K over a 24-h period, but that different oral Vitamin K preparations may have different efficacies. Finally, the observed changes in the INR after reversal are mirrored by the expected changes in the Vitamin K-dependent coagulation factors. This is in contrast to the effect of treatment with fresh-frozen plasma (FFP), which produces only a modest rise in factor IX levels (Makris et al, 1997).
The effects observed 4 h after receiving the different Vitamin K preparations indicate that, in the event of life-threatening haemorrhage, only administration by the i.v. route is appropriate (Fig 1). At 4 h, none of the oral preparations had satisfactorily reversed the INR, whereas in recipients of i.v. Vitamin K, the INR was less than 4·0 in 66% of cases. This difference was statistically significant. There was no relationship between the pre-treatment INR and the magnitude of the response (in fact the I.V. group had the highest starting INRs). Further, the fall in INR in this group was accompanied by a rise in the Vitamin K-dependent factors that was not seen in any of the groups who received oral Vitamin K (Table III).
Although this study was non-randomized and the number of subjects in some of the treatment arms was small, our results support the impression from previous reports that there may be differences in the effect on anticoagulation of different Vitamin K preparations. The effects of Menadiol sodium phosphate, 5 mg and Orakay, 1 mg were unpredictable in that, at 24 h, the INR was greater than 4·0 in at least 50% of patients treated with either drug. The response 24 h post-administration of 1 mg Orakay, was less than that described by Crowther et al (1998,2000) in patients receiving 1 mg of a different Vitamin K preparation. There was no difference in the pre-treatment INRs of patients treated with these preparations compared with those of the other groups. Consequently, if warfarin reversal by Vitamin K is required within 24 h we cannot recommend the use of these two drugs at the doses used in this study. Orally administered Konakion, on the other hand, effectively reversed over-anticoagulation at 24 h. Konakion (2 mg) reversed the INR to < 4·0 in 75% of patients, but 33% of the treated patients developed a subtherapeutic INR (INR < 2). Konakion (5 mg) was more effective in that at 24 h it reversed the INR to < 4·0 in 93% of patients but 73% became subtherapeutic. Effective reversal was thus achieved at 24 h by both doses of oral Konakion, irrespective of the initial pre-treatment INR. Although the small numbers of patients in each group limit the conclusions that can be made regarding efficacy of reversal between oral preparations, other studies have shown conclusively that the i.v. preparation of Vitamin K given orally does produce predictable correction of anticoagulation (Crowther et al, 1998, 2000). We have not explored the reasons for these apparent differences between preparations, but there may be differences in absorption or bioavailability of the products used.
The INR system of monitoring oral anticoagulant control is based on the prothrombin time. This test is sensitive to reductions of the Vitamin K-dependent clotting factors II, VII and X but not to a reduction of factor IX. The INR reporting system is validated on results obtained from stable anticoagulated patients. Consequently, we felt it was important to investigate the relationship between the post-treatment INR and levels of the individual Vitamin K-dependent coagulation factors after reversal of oral anticoagulation with Vitamin K. To our knowledge, this has not been addressed in previous studies of anticoagulant reversal by Vitamin K.
It is generally accepted that the Vitamin K-dependent factors are not all equally suppressed by administration of coumarins. Kumar et al (1990) reported that when Vitamin K-dependent factors were compared in patients anticoagulated stably with warfarin, factor IX levels were highest, factor VII levels intermediate and factors II and X levels lowest. Similar results were obtained by Makris et al (1997) in a group of anticoagulated patients requiring anticoagulant reversal with prothrombin complex concentrate. In the current study, the results obtained 24 h after treatment with all of the oral preparations and at 4 h and 24 h after the i.v. Vitamin K were entirely consistent with these observations (Table III). It is uncertain whether the minor differences observed in the relationship between factors VII and IX and the corresponding INR are of any clinical significance. Because the relationship between the individual Vitamin K-dependent clotting factors was largely maintained, it would seem to us to be unlikely.
A weakness of the current study is that it was not randomized. It was designed to compare the routine clinical practice in the four hospitals where different methods of reversing oral anticoagulation were used. The data were gathered prospectively, the patients included were similar and all assays were performed by laboratories that take part and perform satisfactorily in the UK NEQAS coagulation scheme. We therefore believe that the observations made are valid.
In conclusion, we have demonstrated that the route of administration of Vitamin K has an important impact on reversal of anticoagulation. In view of the more rapid onset of action, only i.v. administration can be recommended for those patients in whom a rapid correction is required. Even with the relatively low dose of 2 mg Konakion, the majority of patients showed a marked but incomplete reversal at 24 h. For patients in whom some reduction of intensity of anticoagulation is required without achieving complete or rapid reversal, oral Vitamin K may be prescribed. There may, however, be clinically important differences between different preparations of Vitamin K. Finally, our data on individual coagulation factor levels show that these all rose with correction of the INR after Vitamin K administration, and that the relative relationship between the levels of these factors in the first 24 h after Vitamin K administration is similar to that seen in stable anticoagulated patients.
The authors are very grateful to Dr Mark Vickers and Professor Mike Greaves for their comments and help.