Influence of Kidney Transplant Status on Warfarin Dose, Anticoagulation Control, and Risk of Hemorrhage

Abstract

Study Design

To assess whether warfarin dose requirement, anticoagulation control, and risk of hemorrhage differ in kidney transplant recipients (KTRs) compared with patients without kidney transplants (non-KTRs).

Design

Analysis of data from the Warfarin Pharmacogenetics Cohort, a prospective cohort of first-time warfarin users followed at two anticoagulation clinics.

Setting

Two outpatient anticoagulation clinics at two large, academic, tertiary care hospitals.

Patients

Adults aged 20 years or older starting warfarin for anticoagulation with a therapeutic international normalized ratio (INR) goal of 2–3 who were kidney transplant recipients (n=65) or patients without kidney transplants (n=1630).

Measurements and Main Results

Warfarin dose requirement, anticoagulation control, and risk of hemorrhage were assessed in each group. KTRs required an approximately 20% lower warfarin dose (4.7 vs 5.6 mg/day, p=0.0005) compared with non-KTRs. Genetic variants had similar effects on dose in both groups. Mean percentage of time in therapeutic range (PTTR) was similar among KTRs and non-KTRs. Although the proportion of patients achieving good anticoagulation control (PTTR ≥ 60%) was low in both groups, it was similar among KTRs and non-KTRs. KTRs had a higher risk of major hemorrhage (hazard ratio 2.1, p=0.0081), but this difference was not statistically significant after controlling for kidney function, clinical, and genetic factors.

Conclusion

KTRs initiating warfarin require lower doses and closer monitoring to optimize anticoagulation therapy. Additional studies are needed to confirm these findings.

Patients with chronic kidney disease (CKD) are known to be at higher risk of venous thromboembolism (VTE), with the risk increasing as estimated glomerular filtration rate (eGFR) decreases^1,2. Increased risk of VTE appears to continue after kidney transplantation, not only in the immediate perioperative period as might be expected, but persisting up to 10 years posttransplantation³. Patients with CKD who develop VTE are also at increased risk of serious complications, with one study finding creatinine clearance < 30 ml/minute to be an independent risk factor for fatal pulmonary embolism and fatal bleeding⁴.

The prevalence of atrial fibrillation (AF) in patients receiving chronic hemodialysis is estimated to be 13–27%, at least 10-fold higher than that in the general population⁵. New-onset AF is also common after kidney transplantation, with one study observing a 7.3% incidence at 36 months posttransplantation⁶. AF is associated with poor clinical outcomes in kidney transplant recipients (KTRs), as it has been shown to be an independent risk factor for mortality and allograft loss^6,7.

Warfarin is a commonly used oral anticoagulant for treatment of clotting disorders and prevention of AF-related stroke. Recent evidence suggests that although warfarin is metabolized in the liver, renal function significantly affects warfarin dosing and toxicity^8,9. Patients with CKD have a higher frequency of supratherapeutic international normalized ratios (INRs), and the risk of warfarin-associated hemorrhage is increased in patients with lower eGFRs^8,9. Warfarin dosing in KTRs is complicated, not only due to decreased eGFR, but also by concurrent use of medications that interact with the cytochrome P450 (CYP) system such as corticosteroids, tacrolimus, and sulfamethoxazole-trimethoprim. Therefore, KTRs who are prescribed warfarin may be at higher risk of serious complications from suboptimal anticoagulation control, and, to our knowledge, there are no clinical guidelines on how to adjust warfarin dosing in this subgroup. Despite many of these patients having a clinical indication for anticoagulation, there is little information available on warfarin dosing and safety in KTRs^10,11. A recent study using Medicare and United States Renal Data System (USRDS) data to evaluate KTRs with AF suggested that there was no significant difference in outcomes of death or allograft failure between those prescribed and not prescribed warfarin¹². This study was limited in that diagnoses and medication information were derived from International Statistical Classification of Diseases and Related Health Problems codes and Medicare claims.

The objective of this study was to assess whether warfarin dose requirement, anticoagulation control, and risk of hemorrhage differ in kidney transplant recipients compared with patients without kidney transplants (non-KTRs).

Methods

Warfarin Pharmacogenetics Cohort

We analyzed data from the Warfarin Pharmacogenetics Cohort (WPC), a prospective cohort of first-time warfarin users who were followed at two anticoagulation clinics—University of Alabama at Birmingham (Birmingham, AL) and Emory University (Atlanta, GA). The study was approved by the institutional review boards at both centers (IRB protocol # X131122001), and the study was conducted with ethical principles in accordance with the Declaration of Helsinki. The characteristics of this cohort have been described in previous publications^13–17. Briefly, inclusion criteria were adults aged 20 years or older starting warfarin for anticoagulation with a target INR of 2–3. Patients had a variety of indications for anticoagulation including VTE, AF, stroke, and myocardial infarction. The cohort included patients who had undergone kidney transplantation as well as those who had CKD (estimated GFR < 60 mL/min/1.73 m²). Patients who required target INR ranges above or below 2–3 were excluded, as were patients with hepatic impairment (elevated bilirubin or transaminase levels, or INR > 1.5 at baseline).

Patient Monitoring and Data Assessments

Patients were monitored at monthly clinic visits after initiation of warfarin for a period of up to 2 years. KTRs and non-KTRs followed the same protocol for monitoring and warfarin dose titration at the anticoagulation clinic. KTRs who received warfarin prior to transplantation were excluded from our analysis. Following enrollment in the WPC, detailed demographics, medical history, and social history were obtained as well as baseline laboratory test results, as detailed in previous publications^13–17. At subsequent clinic visits, data on warfarin dose, INR, active medications, and dietary history were collected. Use of medications known to interact with warfarin was verified through review of the participant’s medical and pharmacy records. Single nucleotide polymorphisms in the following genes known to influence warfarin metabolism were assessed: VKORC1 (rs9923231), CYP2C9 (*2 [rs1799853], *3 [rs1057910]), CYP4F2*3 (rs2108622), CYP2C9 (*5 [rs28371686], *6 [rs9332131], *11 [rs28371685]), and CYP2C (rs12777823)^13–17. Hardy-Weinberg equilibrium was verified for all single nucleotide polymorphisms (p > 0.20).

eGFR was estimated by using the four-variable Modification of Diet in Renal Disease (MDRD) study equation for KTRs and non-KTRs¹⁸. Patients were categorized into three groups based on eGFR at time of enrollment: ≥ 60, 30–59, and < 30 mL/min/1.73 m².

Outcome Definitions

Warfarin dose (mg/day) was defined as the average dose required to maintain INR within the therapeutic range (2–3) and was log transformed to attain normality of residuals. All patients had at least one INR within the target range.

Anticoagulation control was based on percentage of time in therapeutic range (PTTR) that was calculated for each patient using the Rosendaal linear interpolation method¹⁹. Patients with a PTTR ≥ 60% were considered to have attained good anticoagulation control²⁰. As low PTTR is a recognized risk factor of hemorrhage, and PTTR ≤ 60% is a predictor for hemorrhage among warfarin users, we accounted for its influence when evaluating the association of transplant status (KTRs and non-KTRs) and hemorrhage. We also reported the percentage of time above therapeutic range (PTAR) and percentage of time below therapeutic range (PTBR).

Overanticoagulation was defined as INR ≥ 4. Only serious, life-threatening or fatal bleeding episodes were classified as hemorrhagic events as per Schulman et al²¹. INRs above the upper limit of the therapeutic range that were not associated with clinical bleeding were not classified as hemorrhagic events. Hemorrhagic episodes were collected through review of hospital admissions, emergency department visits, and death records. All complications were adjudicated independently by the medical director of the anticoagulation clinic. Only adjudicated events were included in the analyses.

Statistical Analysis

Group differences were tested by using analysis of variance for continuous variables and χ² test of independence for categorical variables. The assumption of Hardy-Weinberg equilibrium was assessed by using the χ² test. We evaluated the association between transplant status and warfarin dose by using multivariable linear regression to account for demographics, clinical characteristics, concurrent medications, and genetic variants. Regression coefficients were used to calculate a percent dose change for relevant clinical and genetic coefficients compared to the referent patient. Similar models for PTTR were implemented.

We used a proportional hazard (PH) model to assess the influence of transplant status on the risk of overanticoagulation (INR ≥ 4) and major hemorrhage. Dependence among multiple events for each individual was corrected by using robust variance estimation, and 95% confidence intervals (CIs) for the hazard ratios (HRs) were assessed. Deviations from the PH expectations were evaluated by investigating the interactions between predictor variables and survival time. All statistical analyses were performed by using SAS statistical analysis software, version 9.3, at a significance level of 0.05.

Results

Baseline characteristics of the study cohort are described in Table 1. The mean ± SD age of participants was 61.0 ± 15.8 years, with African-Americans constituting 44% of the cohort and women 48%. A total of 65 (3.8%) participants were KTRs and were younger than non-KTRs (p < 0.001). KTRs were more likely to be taking warfarin to prevent VTE (p = 0.004) and have a history of hypertension (p = 0.029). KTRs were more likely to be on concurrent statin therapy (80% vs 54.7%, p < 0.001) and more likely to be on amiodarone (20% vs 9.9%, p = 0.018) with coexistent atrial fibrillation. As expected, KTRs had significantly lower eGFRs and required lower warfarin doses to maintain therapeutic INR (4.7 vs 5.6 mg/day, p = 0.0005). A majority of KTRs were treated with a calcineurin inhibitor (80% [52/65 patients]), mycophenolate (71%), and corticosteroids (78%). Possession of pharmacogenetic variants known to influence warfarin response did not significantly differ by transplant status.

Table 1.

Clinical, Genetic and Sociodemographic Characteristics of the Study Patients by Kidney Transplant Status

Characteristic	Patients without Kidney Transplants (n=1630)	Kidney Transplant Recipients (n=65)	P Value
Age (years)	61.6 ± 15.8	53.8 ± 13.9	< 0.001
Height (inches)	67.6 ± 4.1	67.5 ± 4.7	0.92
Weight (pounds)	196.9 ± 50.9	192.2 ± 46.7	0.47
Body mass index (kg/m2)	30.3 ± 7.7	29.5 ± 6.4	0.36
Warfarin dose (mg/day)	5.6 ± 5.5	4.7 ± 4.2	0.0005
eGFR (ml/min/1.73 m2)	69.6 ± 29.4	44.3 ± 33.4	< 0.001
Female	804 (49.3%)	25 (38.5%)	0.086
African-American	711 (43.6%)	32 (49.2%)	0.37
Current smoker	200 (12.3%)	5 (7.7%)	0.27
Indication for warfarin therapya
Venous thromboembolism	690 (43.1%)	41 (63.1%)	0.004
Atrial fibrillation	673 (42.1%)	18 (27.7%)
Stroke, transient ischemic attack, or other	237 (14.6%)	6 (9.2%)
Comorbid conditions
Hypertension	1002 (67.0%)	52 (80.0%)	0.029
Hyperlipidemia	727 (48.6%)	39 (60.0%)	0.073
Diabetes mellitus	472 (31.6%)	28 (43.08%)	0.052
Congestive heart failure	343 (22.9%)	14 (21.5%)	0.79
Kidney function
eGFR ≥ 60 ml/min/1.73 m2	962 (64.2%)	19 (29.7%)	< 0.001
eGFR ≥ 30–59 ml/min/1.73 m2	427 (28.5%)	19 (29.7%)
eGFR < 30 ml/min/1.73 m2	109 (7.3%)	26 (40.6%)
Concurrent medications
Statinsb	819 (54.7%)	52 (80.0%)	< 0.001
Antiplateletc	895 (59.8%)	43 (66.2%)	0.37
Amiodarone	149 (9.9%)	13 (20.0%)	0.019
Patients possessing ≥ 1 minor allele
CYP2C9*2	200 (16.4%)	8 (16.7%)	0.75
CYP2C9*3	98 (8.0%)	5 (10.4%)	0.73
CYP2C9 *5, *6, *11	18 (1.5%)	1 (2.1%)	0.73
VKORC1	523 (41.2%)	22 (44.0%)	0.74
CYP4F2	439 (36.8%)	15 (32.6%)	0.56
rs12777823	429 (36.0%)	11 (23.9%)	0.09

Data are mean ± SD values or no. (%) of patients.

eGFR = estimated glomerular filtration rate.

^aPatients may have had more than one indication for anticoagulation.

^bStatins included any of the 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors.

^cAntiplatelet agents included aspirin, clopidogrel, and dipyridamole as monotherapy or dual therapy.

The association of transplant status with warfarin dose needed for the maintenance of therapeutic anticoagulation is presented in Table 2. Clinical and demographic variables along with transplant status explained 21.1% of the variance in warfarin dose. After adjustment for these variables, transplant status was associated with 18.3% reduction in warfarin dose requirements (model 1; p = 0.0004). After further adjustment for known warfarin pharmacogenetic variants, transplant status was similarly associated with 18.9% reduction in warfarin dose requirements (model 2; p = 0.0004). Clinical, genetic, and demographic variables along with transplant status explained 42.2% of the variance in warfarin dose. Concurrent use of immunosuppressants was included as dependent variables to evaluate their influence on warfarin dose. Among KTRs, concurrent use of immunosuppressants (calcineurin inhibitors [p = 0.37], mycophenolate [p = 0.92], and prednisone [p = 0.91]), did not influence warfarin dose requirements.

Table 2.

Effect of Clinical and Genetic Predictors on Warfarin Dose Requirements by Kidney Transplant Status

Variable	β	% Dose Change (95% CI)	p Value
Clinical and demographic variables only (Model 1; R2 =21.1%)a
Intercept	1.848
African-American	0.177	19.3 (14.0, 25.0)	<0.0001
Age (per 10 years)	−0.076	−7.3 (−8.7, −5.9)	<0.0001
Female	−0.137	−12.8 (−16.7, −8.8)	<0.0001
Body mass index (kg/m2)	0.009	0.9 (0.6, 1.2)	<0.0001
Congestive heart failure	−0.071	−6.9 (−11.6, −1.9)	0.0077
Kidney functionb	−0.064	−6.2 (−9.4, −2.8)	0.0004
Amiodarone	−0.211	−19.0 (−24.7, −13.0)	<0.0001
Transplant status	−0.202	−18.3 (−27.0, −8.6)	0.0004
Clinical, demographic, and genetic variables (Model 2; R2 = 42.3%) c
Intercept	2.182
African-American	−0.063	−6.1 (−11.2, −0.7)	0.0262
Age (per 10 years)	−0.075	−7.2 (−8.6, −5.9)	<0.0001
Female	−0.133	−12.4 (−16.3, −8.4)	<0.0001
Body mass index (kg/m2)	0.009	0.9 (0.6, 1.2)	<0.0001
Congestive heart failure	−0.088	−8.4 (−13.0, −3.5)	0.0009
Kidney functionb	−0.054	−5.2 (−8.5, −1.9)	0.0027
Amiodarone	−0.246	−21.8 (−27.2, −16.1)	<0.0001
CYP2C9*2d	−0.195	−17.7 (−22.2, −13.0)	<0.0001
CYP2C9*3d	−0.423	−34.5 (−39.3, −29.3)	<0.0001
CYP2C9*5, *6, *11e	−0.161	−14.9 (−28.5, 1.3)	0.0689
VKORC1d	−0.308	−26.5 (−29.2, −23.7)	<0.0001
CYP4F2d	0.038	3.9 (−0.1, 8.1)	0.0564
rs12777823	−0.092	−8.8 (−12.9, −4.5)	0.0001
Transplant status	−0.210	−18.9 (−27.8, −8.9)	0.0004

β denotes parameter estimates.

CI = confidence interval.

^aThe referent patient was a 40-year-old, European-American man without a kidney transplant, with an estimated glomerular filtration rate (eGFR) > 60 ml/min/1.73 m², body mass index (BMI) 25 kg/m², with no congestive heart failure (left ventricular ejection fraction >55%), and not taking amiodarone.

^bKidney function was categorized into 3 groups based on eGFR: ≥ 60, 30–59, and < 30 ml/min per 1.73 m². This was included as an ordinal variable in the model with eGFR ≥ 60 ml/min/1.73 m² serving as the reference group.

^cThe referent patient was a 40-year-old, European-American man without a kidney transplant with an eGFR > 60 ml/min/1.73 m², BMI 25 kg/m², with no congestive heart failure (left ventricular ejection fraction >55%), not taking amiodarone, and not possessing variants in CYP2C9*2, CYP2C9*3, CYP4F2, and VKORC1; and CYP2C9*5, *6, and *11 together; and rs12777823.

^dCYP2C9*2, CYP2C9*3, CYP4F2, and VKORC1 were included as additive; 0 if no variants; 1 if heterozygous and 2 if homozygous for the variant allele.

^eCYP2C9*5, *6, and *11 together, and rs12777823 were categorized as 0 if no variants and 1 if heterozygous or homozygous for the variant allele.

Time to attainment of therapeutic INR (INR ≥ 2) was shorter among KTRs compared to non-KTRs (Table 3). However, KTRs were monitored more frequently compared to non-KTRs (2.74 vs 2.26 visits/patient/month, p < 0.005). As expected based on the higher prevalence of VTE indication for therapy, which requires a shorter duration of treatment compared to AF, the follow-up time accrued was shorter for KTRs (14.0 vs 16.4 months/patient, p = 0.046).

Table 3.

Indices of Anticoagulation Attainment and Maintenance by Kidney Transplant status among Warfarin Users with a Therapeutic INR Range of 2–3

Characteristic	Patients without Kidney Transplants (n=1630)	Kidney Transplant Recipients (n=65)	p Value
Time to first therapeutic INR (days)	19.2 ± 26.2	13.5 ± 20.3	0.033
Follow-up/patient (months)	16.4 ± 10.6	14.0 ± 8.7	0.046
Total no. of visits/patient/month	2.26 ± 1.96	2.74 ± 1.26	< 0.005
Percentage of time below therapeutic range (PTBR)	28.2 ± 26.7	35.4 ± 29.0	0.029
Percentage of time in therapeutic range (PTTR)	48.7 ± 47.4	46.3 ± 41.5	0.34
Percentage of time above therapeutic range (PTAR)	18.0 ± 16.9	18.5 ± 14.8	0.8
Poor anticoagulation control (PTTR < 60)	1023 (62.8)	47 (72.3)	0.11
Good anticoagulation control (PTTR ≥ 60)	607 (37.2)	18 (27.7)

Data are mean ± SD values or no. (%) of patients.

INR = international normalized ratio.

Quality of Anticoagulation Control

A high frequency of poor anticoagulation control (PTTR < 60%) was observed in both KTRs and non-KTRs (72.3% and 62.8%, respectively); however, there was no statistically significant difference in PTTR by transplant status. The difference in PTTR remained unchanged after adjustment for demographic and clinical variables, including eGFR and genetic variants known to influence warfarin response (p = 0.34). KTRs also had a greater PTBR (p = 0.03) compared to non-KTRs.

Risk of Overanticoagulation

Transplant status was associated with risk of overanticoagulation (INR > 4): HR 1.8 (95% CI 1.3–2.3, p = 0.002). However, after adjustment for demographic, clinical, and genetic variables, transplant status was no longer associated with overanticoagulation risk (HR 1.4, 95% CI 0.94 – 2.0, p = 0.11).

Incidence (Absolute Risk) of Hemorrhagic Events

One hundred eighty-six hemorrhagic events occurred during 2143.5 person-years of follow-up (incidence rate [IR] 8.7/100 person-years, 95% CI 7.5 – 10.0). Compared to non-KTRs (8.3/100 person-years, 95% CI 7.1 – 9.7), KTRs (18.3/100 person-years, 95% CI 10.0 – 30.7) had a higher incidence of major hemorrhage (IR 2.1, 95% CI 1.2 – 3.7, p = 0.01). Major hemorrhages by site included gastrointestinal (60%), genitourinary (12.9%), retroperitoneal (4.3%), intracranial bleeds (8.6%), hemoptysis (2.9%), and hematomas (11.4%), and did not significantly differ by transplant status.

Relative Risk of Hemorrhagic Events

Compared to non-KTRs, KTRs had a 2-fold risk of major hemorrhage (HR 2.1, 95% CI 1.2 – 3.8, p=0.008; Figure 1a). However, after accounting for kidney function, clinical, and genetic factors, the difference in risk was not statistically significant (HR 1.44, 95% CI 0.73 – 2.84, p = 0.29; Figure 1b). Older age (p = 0.035), African-American race (p = 0.036), hypertension (p = 0.016), concurrent use of an antiplatelet agent (p = 0.01), poor kidney function (p = 0.006), and PTTR < 60% (p < 0.0001) were associated with an increased risk of hemorrhage.

Figure 1.

Time to major hemorrhage among warfarin users by transplant status. Red and black lines indicate data for kidney transplant recipients and patients without kidney transplants, respectively, without adjustment for covariates (a) and with adjustment for covariates (b). HR = hazard ratio.

Discussion

Patients with CKD have significant cardiovascular disease burden that persists after kidney transplantation²², including comorbid conditions such as AF and VTE, which frequently require initiation of oral anticoagulant therapy^5,23. When oral anticoagulation is required among persons with advanced CKD (eGFR < 30 mL/min/1.73m²), warfarin is the drug most often chosen, as some non–vitamin K antagonist oral anticoagulants are contraindicated in these patients²⁴. Notably, there are limited data on response to anticoagulants among patients with advanced CKD²⁵, and we found no data on warfarin response among transplant recipients. To our knowledge, this is the first study to evaluate warfarin response in KTRs. We found that KTRs receiving warfarin required lower therapeutic doses and may have a higher risk of hemorrhagic complications due to multiple comorbidities present in this population.

Recognition of the influence of clinical and genetic factors on warfarin dose has spurred the development of dosing algorithms^{15–17,26–35}. Among these, the warfarin dosing algorithm³⁰ and the International Warfarin Pharmacogenetics Consortium algorithm³⁵ are widely used. However, these algorithms do not incorporate kidney function or transplant status to guide dose selection. We have previously reported that patients with impaired kidney function require lower warfarin doses^15,34. In the present study, we showed that an additional lowering of warfarin dose is required in KTRs independent of allograft function. Table 2 can be used as a resource to guide dose selection based on patient-specific clinical, demographic, and genetic characteristics.

Several important clinical observations were made concerning anticoagulation control that may help to guide warfarin use in patients with kidney transplants. KTRs reached a therapeutic INR faster than non-KTRs. However, this may be accounted for by more frequent visits/patient. KTRs spent less time within the target therapeutic INR range, although there was not a statistically significant difference after accounting for clinical and genetic factors. This difference may have been attenuated, however, as KTRs also had more frequent INR monitoring than non-KTRs. KTRs were also found to have more time below the therapeutic INR range than non-KTRs. KTRs taking warfarin had increased risk of overanticoagulation and hemorrhage, but these risks were not statistically significant after controlling for relevant clinical and genetic factors. These data suggest that warfarin can be safely prescribed in KTRs, but that close therapeutic monitoring is necessary to maintain the target level of anticoagulation in this complex population.

There are no clear guidelines for the management of disorders requiring anticoagulation, such as AF, in patients with advanced CKD. Platelet dysfunction and alterations in the coagulation cascade can complicate the use of anticoagulants in patients with CKD²⁵. We have previously shown that patients with CKD have slower rates of INR reversal when presenting with overanticoagulation than do patients with normal renal function⁸. This finding has led us to hypothesize that patients with CKD have a slower rate of carboxylation of clotting factors. Given the observation in this study that KTRs require a lower warfarin dose independent of renal function, we can also hypothesize that KTRs have decreased clotting factor carboxylation beyond what would be predicted by renal function as estimated by serum creatinine concentration. Further studies are needed to assess these hypotheses.

This study has several strengths in that it evaluated warfarin dosing and safety in a high-risk population using a large well-phenotyped prospective cohort accounting for the influence of clinical, demographic, and genetic factors. The prospective follow-up documented anticoagulation control and assessment of major hemorrhage objectively. However, we recognize limitations of the study, including lack of availability of data on proteinuria to refine assessment of kidney function. Our study was not designed to elucidate the mechanisms by which transplant status may influence warfarin response, and the lower warfarin dose observed in KTRs compared to non-KTRs may be influenced by transplant medications such as tacrolimus and prednisone, which the majority of KTRs in our cohort were taking. Tacrolimus-based immunosuppression is the current standard of care at many transplant centers in the United States; however, KTRs receiving alternate immunosuppressive regimens may not require the same warfarin dose reduction as was observed here. Sulfamethoxazole-trimethoprim is also commonly prescribed to KTRs posttransplantation for pneumocystis prophylaxis. As it inhibits CYP2C9, the principal enzyme that metabolizes warfarin, co-therapy is associated with lower warfarin dose requirements. However, we could not assess the effect of this medication on warfarin dose as our cohort enrolled patients at inception of warfarin therapy, which occurred at varying time points posttransplantation. Transplant status was not associated with an increased risk of major hemorrhage after accounting for kidney (or allograft) function and clinical and genetic factors. These findings need to be confirmed in larger independent cohorts.

Conclusion

Although patients with CKD have an increased risk of hemorrhage while taking warfarin⁸, studies support the net clinical benefit of its use in reducing the risk of stroke and mortality³⁶. Our observations suggest that warfarin may need to be initiated at a lower dose and monitored more closely in kidney transplant recipients than in the general medical population. Although KTRs were observed to have an increased rate of major hemorrhage, it is important to recognize that this risk was not statistically significant after adjustment for relevant covariates. There are limited data on warfarin’s effectiveness in preventing recurrent systemic thromboembolism and on clinical outcomes in this subgroup. However, clinicians also need to understand the thromboembolic risk associated with withholding of warfarin therapy in KTRs. Perhaps ongoing and future research efforts, evaluating thromboembolic and hemorrhagic events, can confirm these findings and assess the net clinical benefit of warfarin use in these unique and medically challenging patients.