Discussion
In 2014, 8.0% of all Medicare Fee-For-Service beneficiaries had a diagnosis of AF (28). In the 65 and older population, AF prevalence was higher among males (10.5%) than females (8.4%) and was highest among Non- Hispanic Whites (10.2%), followed by American Indians/Alaska Natives (5.9%), Hispanics (5.0%), Blacks/African Americans (5.1%), and Asian Pacific Islanders (4.9%). The cause of these differences in prevalence is unknown but recent study suggests that there may be some genetic protection against AF in minority populations (29).
Several additional reports have evaluated the prevalence of AF in CKD patients but few have studied if CKD in AF patients independently increases MACE when adjusted for previously defined traditional risk factors for MACE. Our study is one of the first to assess the effects of CKD on MACE in an African American and Latino NVAF population and provides evidence that a decreased GFR was an independent risk factor for MI, Death and MI, stroke and death among these patients. Our data reinforces other studies that have identified renal dysfunction as an added predictor of MACE, although our data is unique in that it is predominantly from a non-Caucasian population.
Different than other studies in non- African Americans and Latinos, our data showed that a decreased GFR was not an independent risk factor for stroke in our patient population. In a recent large prospective cohort study of AF patients, Go et al found that a lower level of estimated glomerular filtration rate was associated with a graded, increased risk of ischemic stroke and other systemic embolism, independently of known risk factors in AF (30). A study performed in Japan evaluated if CKD constitutes a risk for stroke and it concluded that decreased kidney function increases the risk of first symptomatic stroke events in a general Japanese population (31). Piccini et al reported that adding impaired renal function to standard stroke risk scoring scale was a potent predictor for stroke and systemic embolism in a predominantly white population (22).
Multiple studies have demonstrated that stroke rates are higher in AF patients with renal dysfunction. Vazquez et al demonstrated a 9.8-fold increased risk of ischemic stroke among patients undergoing dialysis who had AF compared those who maintained sinus rhythm during dialysis (32). In the U.S. Renal Data System study, patients with end-stage renal failure and AF had a 1.8-fold higher rate of ischemic strokes, whereas hemorrhagic stroke rates were comparable to end-stage renal failure patients in sinus rhythm (33). Conversely, in the Rotterdam study (34), decreased GFR did not significantly increase the risk of ischemic stroke, but was a strong predictor of hemorrhagic stroke. Genovesi et al demonstrated that AF was associated with greater total and cardiovascular mortality among hemodialysis patients and was more notable for cardiovascular than non-cardiovascular mortality (8). Nakagawa et al demonstrated that combined eGFR and CHADS2 score could be an independent powerful predictor of cardiovascular events and mortality in patients with NVAF (35). Long-term mortality, cardiac events, and stroke risk were >8 times higher when decreased eGFR was present with higher CHADS2 score (≥2).
The mechanism by which chronic kidney disease (CKD) increases the risk of MACE is not known. CKD increases vascular calcification, inflammation, valve problems, and fluctuation in electrolytes, sympathetic nervous system activation and modulation of renin angiotensin system. Tanaka (36) noted an inverse relationship of eGFR to thrombin-antithrombin (TAT) and fibrin D-dimer levels, both indexes of thrombogenesis. Shlipak demonstrated that renal insufficiency was independently associated with elevations in inflammatory and pro-coagulant biomarkers (37). These findings lend support to the notion that enhanced coagulation activation appears to be related to a reduction in residual renal function in patients with AF (38).
A manual review of EMR was performed for all patients included in the study to avoid false positives and negative. However, our study has several limitations. It was completed at one center and was done primarily on African American and Latino patient population. The Latino/Hispanic population was not sub-stratified further and there have been differences in prevalence in various subgroups (5). This was a retrospective analysis of such patients. The contradictory result for CKD being an independent risk factor for stroke could be due to our unique patient population, or a relatively small sample size of patients with paroxysmal AF mistakenly excluded from study.
A prospective study in minority AF patients with renal dysfunction should be done in the future to provide further evidence regarding ischemic neurological events in this patient population and verify our findings of MACE.