LAUSR

Patients with end-stage renal disease (ESRD) often have indications for long-term anticoagulation such as atrial fibrillation, thrombosis of arteriovenous fistulas and grafts, metallic heart valves, and hypercoagulable states. Traditionally, this has been accomplished with warfarin, but questionable effects on outcomes coupled with the potential for serious side effects, the need for continuous therapeutic monitoring, and interactions with diet and other medications has, in the past, called into question the use of warfarin in ESRD patients. More recent data have heightened this concern, particularly its lack of efficacy, the frequency of hemorrhagic complications in the case of atrial fibrillation, and effects on vascular calcification. While the potential for warfarin to exacerbate vascular calcification has been recognized, the magnitude and consequences of this are often underappreciated. Arterial calcification is a significant problem in ESRD even in the absence of warfarin. This calcification is primarily in the medial layer, as opposed to the intimal calcification that occurs in atherosclerosis, and leads to arterial stiffening and impaired pulsatile blood flow distally. Not surprisingly, medial arterial calcification is associated with left ventricular hypertrophy and peripheral arterial disease, particularly in ESRD. Warfarin reduces the activity of matrix gla protein (MGP), a vitamin K-dependent, gamma-carboxylated calcification inhibitor produced by vascular smooth muscle. This results in calcification of vascular smooth muscle (medial arterial calcification) in rodents, with similar findings in mice with deletion of MGP. Warfarin is also strongly associated with calcific uremic arteriolopathy (formerly calciphylaxis), a serious, life-threatening disorder related to medial calcification and thrombosis of small arteries in subcutaneous tissue. Using calcification of breast arteries, which is easily detected on routine mammograms, as a marker of medial arterial calcification in humans, we found a 50% increase in prevalence in women treated with warfarin that jumped to a three-fold increase in those treated for >5 years. A similar increase was seen in lower extremity radiographs in both men and women. That this is due specifically to warfarin was shown in our subsequent longitudinal studies, in which the rate of progression of breast arterial calcification on yearly mammograms was 4-fold greater in women receiving warfarin but not other anticoagulants. This acceleration was observed only during warfarin treatment and not before or after. Warfarin produced a similar proportional increase in progression in ESRD patients, but because of the underlying propensity towards arterial calcification in ESRD, the rate of progression in ESRD patients treated with warfarin was almost 20-fold greater than in women without ESRD not treated with warfarin. Thus, in this population that is already at risk for vascular calcification, the effect of warfarin is quantitatively very large and likely to lead to substantial morbidity. The greater effect of warfarin in ESRD is consistent with recent studies in which genetic inactivation of MGP did not result in vascular calcification unless the mice were fed a high phosphate diet, presumably mimicking the hyperphosphatemia that is common in ESRD. The severity and rapidity of warfarin's effect is illustrated by a patient recently seen by the author, a 47-year-old male on hemodialysis for 10 years with noncompliant phosphorus control who underwent a mechanical aortic valve replacement with initiation of warfarin. Nine months later, he presented with severe pain and tenderness in all four distal extremities with examination showing cold and cyanotic hands and feet consistent with ischemia, with rapid progression to gangrene. Radiographs of the hands and feet showed diffuse, severe medial arterial calcification that was not present on radiographs performed 10 months earlier (Figure 1). It is clear that warfarin promotes vascular calcification in humans and, when combined with the existing risk in ESRD, this calcification can be extensive and clinically significant. The evidence strongly suggests that warfarin should be used sparingly, if at all, in ESRD patients. One could argue that it could be used in patients with excellent control of their mineral metabolism but our data show that warfarin increases arterial calcification in patients with normal renal function and presumably normal mineral metabolism. With data now showing that direct-acting oral anticoagulants can be safe and effective in ESRD, including in patients with calcific uremic arteriolopathy, there seems little need to use warfarin. Unfortunately, as in the patient presented above, these newer agents are not approved for anticoagulation in patients with mechanical heart valves. Received: 2 September 2022 Revised: 9 November 2022 Accepted: 27 November 2022