By Michael H. Crawford, MD, Editor
A Finnish national database study of newly diagnosed atrial fibrillation from 2007-2018 has shown that the independent association of ischemic stroke risk with female sex initially was high but trended downward to nonsignificant at the end of the study. This has implications for the risk stratification of atrial fibrillation patients regarding oral anticoagulation therapy.
Teppo K, Airaksinen KEJ, Jaakkola J, et al. Ischaemic stroke in women with atrial fibrillation: Temporal trends and clinical implications. Eur Heart J 2024;45:1819-1827.
In the CHA2DS2VASc scoring system, female sex is considered an independent risk factor for ischemic stroke (IS) based on older trials of oral anticoagulants (OAC) and observational data. However, this observation is age-dependent and related to other factors encompassing sociocultural gender. Thus, these investigators from Finland hypothesized that the risk associated with female sex may change over time. Accordingly, they performed a sub-study of the Finish Anti Coagulation in Atrial Fibrillation (FinACAF) study, which is a nationwide retrospective cohort study of all patients diagnosed with atrial fibrillation or flutter (AF) from 2004-2018.
This sub-study evaluated those with a new diagnosis of AF between 2007 and 2018 who remained in Finland. They identified 229,565 such individuals (50% women; mean age 73 years; mean follow-up time four years). The women were older and had more comorbidities, lower income, and higher CHA2DS2VASc scores. The mean age of AF diagnosis increased over time, especially in the men. Also, the prevalence of comorbidities, and consequently the CHA2DS2VASc scores, increased in both men and women.
Despite these changes, one-year mortality decreased from 13% to 9% over the study period in both sexes but was persistently higher in women. About 70% of all patients were treated with OAC. The crude incidence of IS was higher in women throughout the study period (21 vs. 15 events/1,000 patient-years; P < 0.001) but decreased over time in both sexes.
After adjusting for confounders, female sex was no longer associated with higher rates of IS. These findings were similar when considering OAC use vs. none and when only the period before OAC use was considered. During the first year of the study period, female sex was independently associated with a 20% to 30% higher IS rate in the adjusted analysis, but this was attenuated over time such that at the end of the observation period, the difference was not statistically significant.
Also, in the one-year follow-up data, those with the highest initial CHA2DS2VASc scores (> 2 in men and > 3 in women) exhibited a higher risk of IS in women, which also decreased over time and became nonsignificant. These findings were not observed in those with lower scores in the one-year follow-up data.
The authors concluded that since the association with female sex and IS in AF patients attenuated over a 12-year period, female sex could be eliminated from the IS risk calculation.
COMMENTARY
The CHA2DS2VASc score was based on about 1,000 patients with AF from the Euro Heart Survey in 2003-2004 treated at a hospital. Thus, the association of IS with female sex probably was affected by selection biases and confounded by socioeconomic factors. Since then, it has been noted that IS rates in both sexes have declined, probably because of better diagnosis and treatment of risk factors for AF. Also, changes in lifestyle and reductions in inequities in AF management may have contributed.
These trends prompted the FinACAF investigators to hypothesize that the risk in women specifically may have changed. They found that the independent association of female sex with IS in AF patients has decreased between 2007 and 2018. Initially, women were noted to have a 20% to 30% higher risk of IS compared to men, which decreased to nonsignificant by the end of the study. When they examined just those with a one-year follow-up, they noted a similar pattern, but only in the highest risk patients (CHA2DS2VASc score > 1 in men and > 2 in women). In those at moderate to low risk of IS, there was no difference in IS rates between men and women at any time during the study period.
To the authors’ knowledge, this is the first study to look at temporal trends in AF risk by sex, and the results challenge the status quo. Importantly, the results were adjusted for OAC use, and the subgroup who were censured at the start of OAC use also showed reduced risk of IS in women over time. The major strengths of FinACAF are the large population of AF patients diagnosed at all levels of care, not just in the hospital, which reduces selection biases.
There are weaknesses to consider. In retrospective database studies, inaccuracies in the diagnosis of comorbidities and outcome data can occur. Also, there could have been residual confounding from unmeasured factors. The diagnosis of IS was made in the hospital, but in Finland almost all patients with IS are hospitalized at least initially. Other than alcohol use, they had no other lifestyle information. In addition, they had no information on OAC compliance. Finally, all the patients were Finns who remained in the country, so the findings may not apply to other ethnic or racial groups.
The FinACAF study makes a strong argument for reconsidering the use of sex in the CHA2DS2VASc score, which, in the majority of cases, is enough to recommend OAC therapy in most older women.
