Outcome of Patients Who Present to the Emergency Department With an Elevated International Normalized Ratio

To the Editor: With the combination of increasing indications for anticoagulation and the aging of the population, there has been greater use of coumadin. In turn, there has been an increasing incidence of over-anticoagulation, especially in elderly people. Prior studies have reported that higher levels of anticoagulation and age are independent risk factors for increasing incidence of bleeding.1–5 The majority of these studies examined office-based or clinic-based populations. There are few studies of elderly over-anticoagulated patients presenting to the emergency department (ED). Our hypotheses are that age and international normalized ratio (INR) correlate with morbidity and mortality in ED patients. The INR levels of ED patients at their initial presentation and their age were compared with their admission rates, length of hospital stay (LOS), initial hemoglobin levels, hemoglobin levels on the third day after presentation, and mortality. An institutional review board–approved retrospective chart review was performed on a consecutive cohort of patients presenting to a suburban tertiary-care ED (annual census 60,000) with an initial INR of 4.0 or greater. Subjects were identified by query of laboratory results obtained in the ED over a 1-year period (December 2002 to November 2003). Treating physicians were unaware of the study and made all decisions for treatment. The authors reviewed all charts. Data collected included demographics, laboratory values, and hospital course. Because all data were objective in nature, there were no discrepancies in data collection or abstractor performance. Data were analyzed using Student t-test and Spearman rank correlation. P<.05 was considered significant. There were 269 charts identified (135 men and 134 women). There were no differences between the INR of admitted patients and that of patients discharged from the ED (mean INR±standard deviation=6.5±2.3 for 246 admitted, 6.1±2.2 for 23 discharged; P=.80). The mean initial INR was 6.4±2.3 (Table 1). There were no differences in the age of patients who were admitted and discharged (246 admitted, mean age 73.1±13.9; 23 discharged, mean age 72.6±11.7; P=.87). There were no differences in levels of anemia between patients who were admitted and those who were discharged. No subjects died in this study (Table 2). These results differ from those from prior studies performed in outpatient settings. A study of asymptomatic, ambulatory patients reported that, in a cohort of 114 patients with an INR greater than 6.0, 9% had abnormal bleeding, 4% experienced a major hemorrhagic episode, and two subjects died.6 Another study reported two fatalities and 13% of patients with significant bleeding in the cohort of 77 patients with an INR greater than 8.7 A third study reported a doubling of the risk of death with an INR greater than 3.5.8 However, the results of the current study were similar to a more recent study that reported no fatalities and a low rate of major bleeding (<1%) in ambulatory asymptomatic patients with an INR of 4.0 or greater. It was assumed that, because the patients in the current study had an acute illness compounding their over-anticoagulation, they would be sicker and have poorer outcomes than the subjects described by prior authors.6–10 It was surprising that none died. Although INR correlated weakly with anemia, it did not correlate with admission rate, LOS, or mortality. Nor was a correlation found between age and admission rates or anemia. An association was found between age and LOS. This study has multiple limitations. Data were not collected on why patients were admitted or discharged. Discharged patients were not followed up. Information was not collected on underlying illnesses, so it is unclear whether the correlation between age and LOS was due to their underlying illness or their anticaogulation. Other outcome measures that may have clinical significance, including the use of blood products and the need for surgery, were also not initially defined. The use of blood products may have elevated hemoglobin levels and may have confounded results. In this ED study, a correlation between the level of INR and admission rate, length of stay, and death was not detected, although the height of the INR weakly correlated with the level of anemia on Days 1 and 3. Furthermore, there were no statistically significant differences in admission rates, anemia, or mortality between the elderly subjects and their younger counterparts, although age correlated with length of stay. Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the author and has determined that none of the authors have any financial or any other kind of personal conflicts with this letter. Author Contributions: All authors contributed to the concept, design, acquisition of data, and preparation of manuscript. DCL and AJ contributed to the interpretation of the data. Sponsor's Role: There were no commercial sponsors of this project.