Persistence of warfarin therapy for residents in long-term care who have atrial fibrillation.

BACKGROUND: Among long-term care (LTC) residents with atrial fibrillation (AF), the use of warfarin to prevent stroke has been shown to be suboptimal. For those who begin warfarin prophylaxis in LTC, persistence on this therapy has not been reported. OBJECTIVE: This study was conducted to estimate persistence on warfarin among LTC residents with AF. METHODS: A retrospective analysis was conducted by using data from an LTC database. Pharmacy dispensing data were used to track warfarin use in residents with a diagnosis of AF who were newly started on warfarin therapy. The main outcome measure was persistence of warfarin over the first year of therapy. Survival analysis included Kaplan-Meier plots and a multivariate Cox proportional hazards model to test the association of resident characteristics and conditions with warfarin discontinuation. RESULTS: A total of 148 residents new to warfarin therapy met all study inclusion criteria. Median age was 84 years; 69% were female. Median time to therapy discontinuation was 197 days (95% CI, 137-249) across all study residents. By 90 days after the initiation of therapy, 37% (95% CI, 28-47) of study residents had discontinued warfarin; by 1 year, 65% (54%-76%) had discontinued warfarin therapy. The multivariate Cox regression analysis found that the following factors were independently associated with discontinuation of warfarin therapy: age 65 to 74 years (hazard ratio [HR] = 3.01 [95% CI, 1.04-8.73]), female sex (HR = 0.45 [95% CI, 0.24-0.87]), Hispanic race/ethnicity (HR = 2.86 [95% CI, 1.30-6.26]), Midwest region (HR = 2.13 [95% CI, 1.02-4.48]), and Alzheimer disease or dementia (HR = 1.97 [95% CI, 1.05-3.68]). CONCLUSIONS: Although clinical practice guidelines exist for the prevention of stroke in AF patients, persistence on warfarin therapy seems suboptimal in many LTC residents with AF.

Warfarin for prevention of thrombosis among long-term care residents with atrial fibrillation: evidence of continuing low use despite consideration of stroke and bleeding risk.

OBJECTIVES: The aims of the study were to evaluate usage rates of warfarin in stroke prophylaxis and the association with assessed stages of stroke and bleeding risk in long-term care (LTC) residents with atrial fibrillation (AFib). METHODS: A cross-sectional analysis of two LTC databases (the National Nursing Home Survey [NNHS] 2004 and an integrated LTC database: AnalytiCare) was conducted. The study involved LTC facilities across the USA (NNHS) and within 19 states (AnalytiCare). It included LTC residents diagnosed with AFib (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] diagnostic code 427.3X). Consensus guideline algorithms were used to classify residents by stroke risk categories: low (none or 1+ weak stroke risk factors), moderate (1 moderate), high (2+ moderate or 1+ high). Residents were also classified by number of risk factors for bleeding (0-1, 2, 3, 4+). Current use of warfarin was assessed. A logistic regression model predicted odds of warfarin use associated with the stroke and bleeding risk categories. RESULTS: The NNHS and AnalytiCare databases had 1,454 and 3,757 residents with AFib, respectively. In all, 34 % and 45 % of residents with AFib in each respective database were receiving warfarin. Only 36 % and 45 % of high-stroke-risk residents were receiving warfarin, respectively. In the logistic regression model for the NNHS data, when compared with those residents having none or 1+ weak stroke risk and 0-1 bleeding risk factors, the odds of receiving warfarin increased with stroke risk (odds ratio [OR] = 1.93, p = 0.118 [1 moderate risk factor]; OR = 3.19, p = 0.005 [2+ moderate risk factors]; and OR = 8.18, p ≤ 0.001 [1+ high risk factors]) and decreased with bleeding risk (OR = 0.83, p = 0.366 [2 risk factors]; OR = 0.47, p ≤ 0.001 [3 risk factors]; OR = 0.17, p ≤ 0.001 [4+ risk factors]). A similar directional but more constrained trend was noted for the AnalytiCare data: only 3 and 4+ bleeding risk factors were significant. CONCLUSIONS: The results from two LTC databases suggest that residents with AFib have a high risk of stroke. Warfarin use increased with greater stroke risk and declined with greater bleeding risk; however, only half of those classified as appropriate warfarin candidates were receiving guideline-recommended anticoagulant prophylaxis.

Usage of medications with high potential to interact with warfarin among atrial fibrillation residents in long-term care facilities.

OBJECTIVES: The rate of potential warfarin drug-drug interactions among long-term care (LTC) residents with atrial fibrillation (AFib) is evaluated. RESEARCH DESIGN AND METHODS: LTC residents from two databases, the National Nursing Home Survey (NNHS) 2004 and the AnalytiCare™ multistate database, with an AFib diagnosis (ICD9 = 427.3x) were studied. Concurrent usage was tabulated for 10 agents/pharmacotherapeutic classes (simple-class screen) and for 96 individual agents (single-agent screen) with strong evidence for warfarin-interaction. RESULTS: Warfarin use was recorded in 502 NNHS and 1,674 AnalytiCare residents with AFib. Using the simple-class screen, ≥ 1 of the 10 classes/agents were prescribed concurrently in 59% (CI: 54.0 - 63.6%) of warfarin users in NNHS and 85% (CI: 82.9 - 86.3%) in AnalytiCare. Using the more comprehensive single-agent screen, usage of any interacting agent was higher in both databases: 87% (CI: 83.9 - 90.3%) in NNHS and 94% (CI: 93.0 - 95.3%) in AnalytiCare. Among 873 interacting agents dispensed to residents while continuously using warfarin (AnalytiCare database), 360 (41%) were changed (added or withdrawn), while 513 (59%) were maintained over two consecutive 45-day periods. CONCLUSIONS: Concurrent use of agents with a high potential for interaction occurs among most LTC residents who are receiving warfarin. Large proportions of interacting agents appear to be added or withdrawn during continuous warfarin therapy and thus have implications for anticoagulation control.

Prevalence of atrial fibrillation in US nursing homes: results from the National Nursing Home Survey, 1985-2004.

OBJECTIVES: To evaluate the prevalence of atrial fibrillation (AFib) in US nursing homes from 1985 to 2004 and to project the prevalence of AFib to 2030. DESIGN: This study is an analysis of cross-sectional data from the US National Nursing Home Survey, years 1985, 1995, 1997, 1999, and 2004. SETTING: Randomly selected long term care facilities in the United States licensed by the state or certified for Medicaid/Medicare reimbursement. PARTICIPANTS: Randomly selected residents within study facilities. MEASUREMENTS: National Nursing Home Survey demographics and current medical conditions data were analyzed. Population estimates were calculated using National Nursing Home Survey sample weights. Absolute observed annual linear growth of the AFib prevalence rate was calculated using linear regression. Predictive margins were estimated using logistic regression models to evaluate effect of changes in resident case-mix over the survey years. Three estimation methods predicted the number residents having AFib in 2030. RESULTS: The sample sizes of surveyed resident groups were as follows: n = 5238 (1985); n = 8056 (1995); n = 8138 (1997); n = 8215 (1999); and n = 13,507 (2004). Prevalence rates of AFib by year were 2.8% (95% confidence interval [CI]: 2.3-3.4%; 1985), 5.1% (95% CI: 4.6-5.6%; 1995), 5.8% (95% CI: 5.3-6.3%; 1997), 6.9% (95% CI: 6.3-7.4%; 1999), and 10.9% (95% CI: 10.2-11.5%; 2004). Population estimates of nursing home residents with AFib (in thousands) were 42.2 (95% CI: 34.1-50.3; 1985), 78.7 (95% CI: 70.8-86.7; 1995), 93.6 (95% CI: 84.9-102.3; 1997), 111.8 (95% CI: 102.1-121.5; 1999), and 162.1 (95% CI: 152.4-171.7; 2004). Absolute annual linear growth in the prevalence rate of AFib was +0.38% observed (P = .022), +0.39% using unadjusted predictive margins (P = .007), and +0.37% using adjusted predictive margins (P = .007). Projected estimates showed that 272,000 (95% CI: 197,000-347,000), 300,000, or 325,000 residents would have AFib in the year 2030. CONCLUSION: The prevalence of AFib in US nursing home residents increased from 1985 to 2004 and is projected to grow substantially over the next 20 years, potentially resulting in an increased nursing home staff burden owing to increased stroke risk evaluations.