Research Article: Effect of Chronic Kidney Diseases on Mortality among Digoxin Users Treated for Non-Valvular Atrial Fibrillation: A Nationwide Register-Based Retrospective Cohort Study

Date Published: July 28, 2016

Publisher: Public Library of Science

Author(s): Maurizio Sessa, Annamaria Mascolo, Mikkel Porsborg Andersen, Giuseppe Rosano, Francesco Rossi, Annalisa Capuano, Christian Torp-Pedersen, Thomas Berger.

http://doi.org/10.1371/journal.pone.0160337

Abstract

This study investigated the impact of chronic kidney disease on all-causes and cardiovascular mortality in patients with atrial fibrillation treated with digoxin.

All patients with non-valvular atrial fibrillation and/or atrial flutter as hospitalization diagnosis from January 1, 1997 to December 31, 2012 were identified in Danish nationwide administrative registries. Cox proportional hazard model was used to compare the adjusted risk of all-causes and cardiovascular mortality among patients with and without chronic kidney disease and among patients with different chronic kidney disease stages within 180 days and 2 years from the first digoxin prescription.

We identified 37,981 patients receiving digoxin; 1884 patients had the diagnosis of chronic kidney disease. Cox regression analysis showed no statistically significant differences in all-causes (Hazard Ratio, HR 0.89; 95% confident interval, CI 0.78–1.03) and cardiovascular mortality (HR 0.88; 95%CI 0.74–1.05) among patients with and without chronic kidney disease within 180 days of follow-up period. No statistically significant differences was found using a 2 years follow-up period neither for all causes mortality (HR 0.90; 95%CI 0.79–1.03), nor for cardiovascular mortality (HR 0.87; 95%CI 0.74–1.02). No statistically significant differences was found comparing patients with and without estimated Glomerular Filtration Rate <30ml/min/1.73m2 and patients with different stages of chronic kidney disease, for all-causes and cardiovascular mortality within 180 days and 2 years from the first digoxin prescription. This study suggest no direct effect of chronic kidney disease and chronic kidney disease stages on all-causes and cardiovascular mortality within both 180 days and 2 years from the first digoxin prescription in patients treatment-naïve with digoxin for non-valvular atrial fibrillation.

Partial Text

A main strategy to treat atrial fibrillation is heart rate control. Heart rate control strategy includes the administration of one or more heart rate controlling agents, including non-dihydropyridine calcium channel antagonists, beta-blockers, and digoxin [1–4]. After the post-Digoxin Investigation Group trial era [5], digoxin is commonly used worldwide as heart rate control agent in the treatment of atrial fibrillation [6]. Studies have examined the effect of digoxin on mortality [7–28], but no single study have directly evaluated the effect of chronic kidney disease on mortality in patients that receive digoxin for atrial fibrillation. European Heart Rhythm Associations position paper for heart rate control therapy in patients with chronic kidney disease [29], endorsed by American and Asian Heart Rhythm Association, recommend the same treatment to patients with or without kidney disease with appropriate adjustment of dose according to glomerular filtration rate [29]. Atrial fibrillation and kidney disease are commonly found in the same patient and prior studies have shown that the presence of one condition increased the likelihood of finding the other [30–33]. Patients with chronic kidney disease have generally been excluded from clinical trials, therefore register based studies are currently the best opportunity to gain further insight [29,33].

To our knowledge, this is the first observational study that directly examined the effect of chronic kidney disease and its stages on all-causes and cardiovascular mortality among patients treated with digoxin as unique antiarrhythmic agent for non-valvular atrial fibrillation. To date, only one study indirectly tried to evaluate the effect of renal function on mortality in a subpopulation of patients with atrial fibrillation, showing a non-statistically significant interaction term of renal function on mortality in the comparison between digoxin and non-digoxin users [13]. The current study, instead, directly compared patients with and without chronic kidney disease, and patients with different stages of chronic kidney disease for all-causes and cardiovascular mortality, among patients treated with digoxin as unique antiarrhythmic agent for non-valvular atrial fibrillation. This study design was chosen to avoid the effect of indication bias in the evaluation of effect of chronic kidney disease on mortality comparing patients treated with different heart rate control agents. We believe this was necessary because the use of a specific heart rate control agent was a choice made by the cardiologist based on the clinical judgement, after an appropriate examination of the patient and it is not a random event. An unknown number of factors may influence the physician prescription of antiarrhythmic drug including clinical or prescriber-related factors (recorded or not). Therefore, potential unknown factors could influence this association comparing patients treated with different antiarrhythmic drugs. This paper, therefore contribute to the current literature both by providing the results of a direct evaluation of the effect of chronic kidney disease on all-causes and cardiovascular mortality, both by reducing the effect of indication bias in this evaluation, among patients treated with digoxin as unique antiarrhythmic agent for non-valvular atrial fibrillation. Moreover, our study provide new information to the literature regarding the direct effect of chronic kidney disease stages in all-causes and cardiovascular mortality among patients treated with digoxin as unique antiarrhythmic agent for non-valvular atrial fibrillation. Recently, increased awareness has emerged regarding the effect of the chronic kidney disease stages on clinical outcome, especially among cardiologists treating arrhythmic disorders [29]. The main reason is related to the complexity of the pathological and physiological interactions between the kidney and the heart that could have clinical implications. However, few information are available on this topic and little is known regarding how different antiarrhythmic drugs could potentially change these interactions, especially for drugs with narrow therapeutic ranges like digoxin [47]. Our results shown that no statistically significant difference exist among patients with eGFR <30 ml/min/1.73m2 and those with eGFR ≥ 30 ml/min/1.73m2 or among different chronic kidney disease stages, in patients treated with digoxin as unique antiarrhythmic agent for non-valvular atrial fibrillation, within 180 days and 2 years from first digoxin prescription. This study suggest no direct effect of chronic kidney disease and also chronic kidney disease stages on all-causes and cardiovascular mortality within 180 days and 2 years from the first prescription of digoxin among patients treated with digoxin for non-valvular atrial fibrillation. The evaluation of the effect of chronic kidney disease and its stages on all-causes and cardiovascular mortality was performed using an intention to treat approach. This approach could not account for a possible change in the treatment or covariate during the follow-up period, especially in the longer follow-up period. A limited amount of patients in our population had biomarkers measurement (e.g. creatinine, potassium, etc.) within 180 days before index date potentially due to the data-source used. The LABKA database cover 1.8 million people for a period of more than 10 years only from the North Denmark region and the Central Denmark region. Other limitations included a low prevalence of patients with chronic kidney disease compared to those expected in end-of-life population of patients with atrial fibrillation, short median observation period of 82 days and low persistence in digoxin treatment. Moreover, another limitation include a lower prevalence of patients in treatment with warfarin (1.9%) to those expected in patients with atrial fibrillation. These limitations should be considered when interpreting the results.   Source: http://doi.org/10.1371/journal.pone.0160337