Brand Name versus Generic Warfarin

A Systematic Review of the Literature

Francesco Dentali, M.D.; Marco P. Donadini, M.D.; Nathan Clark, Pharm.D.; Mark A. Crowther, M.D.; David Garcia, M.D.; Elaine Hylek, M.D.; Dan M. Witt, Pharm.D.; Walter Ageno, M.D.


Pharmacotherapy. 2011;31(4):386-393. 

In This Article

Selected Studies

The study identification and selection process is summarized in Figure 1. Eleven studies were included in this systematic review. All were English-language studies. Original studies ranged in size from 7–2299 patients, for a total of 3988 included patients. Baseline characteristics of the patients included in the studies are summarized in Table 1. Five studies were RCTs[6,7,12–14] and five were cohort studies;[3–5,8,15] one study was an ecologic study (in which the unit of analysis is a population rather than an individual) that evaluated the consequences of substituting generic for brand name warfarin in a cohort of 36,724 patients.[16]

Figure 1.

Schematic of the study identification and selection process. ASH = American Society of Hematology; ISTH = International Society of Thrombosis and Haemostasis; ACCP = American College of Clinical Pharmacy.

Quality assessment items for the cohort studies and RCTs are summarized in Table 2 and Table 3. Two[8,15] of the five cohort studies and one[14] of the five RCTs were considered of high quality.

Randomized Controlled Trials

Four of the included studies had a crossover design[6,7,12,13] and one had a multiple n-of-1 crossover design.[14] Three of the four RCTs with a crossover design measured the differences in the mean INR when brand name or generic warfarin was administered.[6,7,13] These studies included patients with atrial fibrillation or mechanical heart valves who were receiving long-term anti-coagulant therapy with stable anticoagulation. In two of these studies, brand name warfarin was compared with the warfarin sodium manufactured by Barr Laboratories (Pomona, NY);[6,7] in the other study, brand name warfarin was compared with warfarin sodium manufactured by Lennon Limited (Port Elizabeth, South Africa).[13] According to the randomized, two-sequence, three-period, crossover design of these three studies, patients were initially randomly assigned to receive either brand name or generic warfarin therapy. After the initial period of treatment (period A), patients were subsequently switched to the alternate formulation of warfarin (period B) at the same dose and regimen as during period A. After period B, patients continued the same formulation at the same dose and regimen as taken during treatment period B.

In each of these studies, there was no significant difference in the average INR between patients assigned to brand name and those assigned to generic warfarin, respectively (2.38 vs 2.43,[6] 2.45 vs 2.33,[7] and 2.27 vs 2.28[13]); the 90% CI of the difference in the pooled INRs was within the prespecified 80–125% range for bioequivalence (95–109%,[6] 90.5–99%,[7] and 96.4–109%[13]), and the intersubject variances for brand name or generic warfarin were similar. Furthermore, no significant differences were noted in all safety outcomes, including the number of thromboembolic and bleeding complications, which were similar during treatment with either of these drugs (3 of 27 patients vs 7 of 28 patients,[6] 2 of 27 patients vs 6 of 30 patients,[7] and 1 of 19 patients vs 1 of 16 patients[13] for brand name vs generic warfarin during period A).

In the fourth study with a crossover design, the authors evaluated the propensity for and the magnitude of a change in dosage among 113 patients with chronic atrial fibrillation who had been receiving a stable dosage of brand name warfarin and were randomly assigned to either brand name or generic warfarin (produced by Apothecon, Barneveld, The Netherlands) for 4 weeks.[12] Thereafter, patients were switched to the alternate warfarin product for 4 additional weeks. During the study period, patients' INR values were monitored weekly. Neither the propensity for nor the magnitude of a dosage change or INR change appeared related to a particular warfarin product (nonsignificant differences for each variable after each study period, p>0.5). After 4 weeks of treatment, the same number of patients (seven) in each group experienced a 20% or greater change in warfarin dose from their respective baseline doses. The number of patients with INRs outside the desired protocol range after 4 weeks of treatment was similar in both groups (18 of 104 vs 19 of 104 patients in the brand name vs generic warfarin groups). No major hemorrhagic or thromboembolic events occurred.

In the last trial, the authors used the multiple n-of-1 randomized crossover design to evaluate the equivalency of brand name and generic warfarin (Apo-warfarin; Apotex, Weston, Canada).[14] Seven patients were switched between brand name and generic warfarin over 30 weeks. Study patients took each drug for five 3-week periods, and the INR was measured 2 times/period. This study found no statistically significant differences between warfarin products in terms of mean INR results or number of dosage adjustments required. There was also no significant difference in INR variation based on warfarin formulation (p>0.69), nor was there a patient and warfarin interaction found (p>0.81).

Observational Studies

Three of the five cohort studies were retrospective and two were prospective. In the first of these studies, anticoagulation control in 40 patients receiving brand name warfarin and in 15 patients switched to a generic warfarin (Panwarfarin; Abbott Laboratories, North Chicago, IL) was compared.[3] In this study, prothrombin time was in range in only 39% of visits in patients in the generic group compared with 69% of visits in the brand name warfarin group (p<0.001). Furthermore, the percentage of patients requiring a dosage change and the mean dose change were higher in patients in the generic group compared with the brand name warfarin group (p<0.05).

A subsequent prospective cohort study evaluated anticoagulation control during a period of 8 weeks in a group of 105 patients who voluntary switched to a generic warfarin sodium product (Barr Laboratories, Pomona, NY) and in another group of 105 patients using brand name warfarin.[8] Although the two groups showed similar baseline characteristics, this was not a randomized study. At the visit at week 8 after enrollment, the mean INR values for the conversion group and the brand name group were similar (2.7 vs 2.8). Compared with the period before product switch, the variability of the INR in the conversion group was not significantly different after the switch. Furthermore, only a few patients had a change in INR greater than 1 after enrollment in the study, and dosage adjustments were required more frequently in the brand name warfarin group.

Three additional studies in patients receiving oral anticoagulant therapy compared anticoagulation control and the number of thromboembolic and hemorrhagic complications before and after switching to a generic warfarin product.[4,5,15] The first of these studies did not find significant differences in any end point in a group of 182 patients switched from brand name to generic warfarin (Barr Laboratories), who were followed for 18 months.[15] The second study evaluated the mean warfarin dose and the apparent warfarin sensitivity index (WSI, defined as the mean of all INR values/warfarin dose values) in 975 patients over two 6-month periods before and after switching from brand name to generic warfarin (Taro-warfarin; Taro Pharmaceutical Industries Ltd, Kibbutz Yakum, Israel).[4] Compared with the period before switching, the mean warfarin dose was 26.5% higher and the mean INR 4.2% lower after switching (p<0.0001). Warfarin sensitivity, measured by the WSI, was also found to be reduced after switching (p<0.0001). However, the ratio of the WSI (which was used to estimate altered bioavailability and was defined as the WSI after the switch divided by the WSI before the switch) fell within the range of bioequivalence. Furthermore, the number of hospital admissions was not significantly different between the two periods.

The third study compared time within the therapeutic range in 2299 patients receiving oral anticoagulant therapy during the 90 days before and after switching from brand name to generic warfarin (Barr Laboratories).[5] Although mean time within the therapeutic range was statistically significantly lower after switching (65.9% vs 63.3%, p=0.0002), this difference was not deemed to be clinically significant. Furthermore, the difference in the mean warfarin dose before and after conversion was only 0.05 mg, which represents an undetectable value in clinical practice and a dose impossible to obtain with marketed warfarin. At the individual patient level, however, a significant proportion of patients (72.0%) experienced a 10% or greater change in therapeutic INR control after the switch to generic warfarin. Slightly less than half of these patient had their INR control improved by more than 10%, and slightly more than half had a greater than 10% worsening of INR control. The observed rate of anticoagulation-related adverse events was not statistically significantly different between the two periods (2.65/100 patient-yrs before switching vs 4.06/100 patient-yrs after switching, p=0.21).

The final investigation was an ecologic study of trends in warfarin prescribing, INR testing, and hospitalization for major hemorrhage or stroke before and after the introduction of a policy of substitution of brand name warfarin with either one of two generic warfarin formulations (Apo-warfarin and Taro-warfarin) in Canada.[16] This study used administrative health data collected from April 1, 1997–March 31, 2002, in patients older than 66 years. The study examined the 40 months before a provincially mandated switch to generic warfarin, the month during which the switch occurred, and the 9 months thereafter. The number of patients treated with warfarin increased steadily, from 22,929 in April 1997 to 36,724 in March 2002. Three months after the policy implementation, 87% of all warfarin prescriptions were for generic warfarin. No significant changes in the rates of INR testing (p=0.93) or hospitalization for either major hemorrhage (p=0.97) or cerebral thromboembolism (p=0.89) in this study were noted.

The results of the 11 included studies are summarized in Table 4.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: