Warfarin-related nephropathy: Prevalence, risk factors and prognosis

Abstract

Warfarin has been the most commonly prescribed oral anticoagulantsince its approval in 1954 [1]. Warfarin is widely used in the preventionand treatment of deep venous thrombosis and pulmonary embolism,thromboembolismpreventioninthepatientswithatrial fibrillation,pros-thetic heart valves and indwelling central venous catheters. It interruptsthe synthesis of coagulation factors (II, VII, IX, and X) by inhibiting thevitamin K epoxide reductase (VKOR) and causes disruption of the extrin-sicclottingcascade.Warfarinisstillun deruseinclinicalpracticeespeciallyin China despite the strong evidence of this drug for thromboembolismprevention [2].Thismaybeduetoitsnarrowtherapeuticrange,potentialinteractions with a variety of commonly used drug and food [3], and itspropensity to cause hemorrhage. Also, different warfarin-relatedgenotypes were also related to over-anticoagulation and hemorrhagiccomplications [4].Warfarin-relatednephropathy(WRN)isarecentlyrecognizedcom-plication during warfarin anticoagulation therapy, in which excessivewarfarinization [international normalized ratio (INR) N3.0] leads toacute kidney injury (AKI) by causing glomerular hemorrhage andrenal tubular obstruction by red blood cell (RBC) casts [5,6]. Brodskyet al. [7] was the first to describe this entity through kidney biopsy ina subset of patients with warfarin overdose, hematuria, and AKI, andeachbiopsyspecimendemonstratedtheevidenceofacutetubularinju-ry, glomerular hemorrhage and renal tubular obstruction by RBC casts.They termed clinical WRN as an episode of unexplained acute renalinjury defined as a serum creatinine increase greater than 0.3 mg/dlwithin one week of an INR measurement greater than 3.0 in a patientbeing treated with warfarin without clinical evidence of hemorrhage[5,6]. They also established an animal model of WRN, and showed theexcessive anticoagulation increased serum creatinine levels andhematuriainthe5/6nephrectomyratsbutnotincontrols,whichresultedin the formation of obstructing tubular RBC casts [8].However, WRN could be the result of a problem more widespreadthan complete tubular obstruction by RBCs. Other potential mecha-nismsmayparticipateintheoccurrenceofWRN(Fig.1).First,oxidativestress damage to tubules by RBC, even though the RBC did not obstructthetubule,couldleadtoWRN[9].Thepatientswithchronickidneydis-ease have reduced plasma antioxidant enzyme activities, such as gluta-thioneperoxidaseandcatalase,whichmaycontributetoahigherriskofWRN. Other important mechanisms, including atheroembolism [10],interstitial nephritis [11], apoptosis of glomerular endothelial cells [8]and direct effects of warfarin on the glomerulus [12] may also contrib-ute to the development of WRN. Warfarin has been shown to affectglomerular mesangial cells by interfering with the activation of theproduct of growth arrest-specificgene6[11]. This could affect glomer-ular hemodynamics or aggravate the underlying glomerular disease.Brodsky et al. [5] further investigated the prevalence, risk factors,and consequence of WRN in a large cohort of patients who initiatedwarfarin therapy during a 5-year period at the Ohio State UniversityMedical Center, WRN occurred in 20.5% of the entire cohort, 33.0% ofthe CKD cohort, and 16.5% of the non-CKD cohort, which indicatedthat patients with CKD are at much greater risk of WRN than thosewithout CKD. It suggests that CKD patients may be prone to over-anticoagulation as spend less time within the target range, requiredmorefrequentadjustmentsandhadhigherbleedingrisk.Otherriskfac-torsfor WRNincludedage, diabetesmellitus,hypertension,andcardio-vascular disease. Therefore, WRN may be a common complication ofwarfarin therapy in high-risk patients, especially in the CKD patients.Recently, An et al. [13] showedWRNdeveloped in19.3% of thepatientshavingexcessivewarfarinization.Alowerbaselineserumalbumin,highserum AST at post INR elevation, and heart failure were independentrisk factors for WRN. In contrast, atrial fibrillation significantlydecreased the risk of WRN. However, neither the presence of CKD norbaseline estimated glomerular filtration rate (eGFR) was an indepen-dent risk factor for WRN.The occurrence of WRN also increased long-term mortality in pa-tients with and without CKD. Brodsky et al. [5] reported that one-yearmortality was significantly higher in the patients with WRN compared