Atrial fibrillation and upper limb thromboembolectomy: a national cohort study

Authors


Ljubica Vukelic Andersen, Department of Clinical Pharmacology, Aarhus University Hospital, Wilhelm Meyers Allé 4, DK-8000 Aarhus C, Denmark.
Tel.: +45 8642 1791; fax: +45 8612 8804.
E-mail: ljubica.andersen@farm.au.dk

Abstract

Summary. Background: The risk factors associated with, and the incidence of systemic embolism in patients with atrial fibrillation (AF) are poorly understood. Objectives: We studied the association between AF and upper limb thromboembolectomy involving brachial, ulnar or radial artery in a national cohort study that included all individuals aged 40–99 years with incident AF. Methods: Data were retrieved from the Danish National Vascular Registry, the National Registry of Patients, the Danish Civil Registration System and Statistics Denmark. Results: In total, 131 476 patients (68 042 men and 63 434 women) with AF without previous thromboembolectomy in the upper limb were registered. In the study cohort, 130 men underwent upper limb thromboembolectomy over 220 890 person-years of observation, whilst 275 women underwent thromboembolectomy over 197 777 patient-years. The incidence per 100 000 person-years was 58.9 (95% CI, 49.2–69.8) for men and 139.1 (95% CI, 123.1–156.5) for women. The relative risk of thromboembolectomy among patients with AF compared to the background population was 7.5 (95% CI, 6.3–8.9) for men, and 9.3 (95% CI, 8.3–10.5) for women. Women with AF had a relative thromboembolectomy risk of 1.8 (95% CI, 1.5–2.3) compared to men with AF. Among patients with AF, history of hypertension (HR 2.2–2.9), myocardial infarction (HR 2.9–3.9), heart failure (HR 1.6–1.9) and stroke (HR 2.2–3.8) were significantly associated with increased risk of thromboembolectomy in both men and women. Conclusions: AF substantially increases the risk of upper limb thromboembolectomy. This risk is higher with increasing age, female gender, and associated with hypertension, myocardial infarction, heart failure and stroke.

Introduction

Atrial fibrillation (AF) increases the risk of stroke [1–4]. This risk is not homogeneous and established risk factors for stroke in AF include prior stroke, diabetes, hypertension, systolic heart failure and female sex [5]. The risk factors for and incidence of non-central nervous system (CNS) systemic embolism per se in patients with AF are less known, although it is assumed that the risk factors for non-CNS systemic embolism are similar to stroke risk factors. Indeed, the literature on non-CNS systemic embolism per se in various organ systems such as gastro-intestinal, kidney, heart and limbs largely consists of case reports and single centre case series [6–11].

Major embolism to the arm is usually easy to diagnose. Symptoms of sudden upper limb ischaemia are pulselessness, pain, pallor paresthesia and paralysis [12]. Thromboembolectomy is considered the treatment of choice in patients who present with acutely developed ischaemia of the upper limb of suspected embolic origin [8,12]. Thus, assessment of non-CNS systemic thromboembolism using upper limb thromboembolectomy as a surrogate measure uses a well-defined clinical endpoint.

In this national cohort study, we tested the hypothesis of a strong association of AF with upper limb thromboembolectomy involving the brachial, ulnar or radial artery. We also tested the hypothesis that risk of upper limb thromboembolectomy is associated with age, sex, hypertension, diabetes, myocardial infarction, heart failure and stroke in patients with AF.

Materials and methods

Data were retrieved from the Danish Vascular National Registry, the Danish National Registry of Patients, and from the Danish Civil Registration System and Statistics Denmark [13]. The Danish National Vascular Registry was initiated in 1989. The purpose of the registry is primarily monitoring of the quality of procedures. Information in the Danish National Vascular Registry includes patients’ civil registration numbers, admission hospitals and departments, dates of admission and discharge, and surgical procedures, using the Danish Classification of Surgical Procedures (until December 31, 1995) or the Nordic Classification of Surgical Procedures (from January 1, 1996) coding system. The surgical procedure codes were assigned by the operating surgeon immediately after the operation.

The Danish National Registry of Patients was established in 1977 and records 99.4% of all non-psychiatric hospital admissions in Denmark. Since 1995 it has also included data on outpatient and emergency clinic patients. Reporting to the Danish National Registry of Patients is mandatory, and it contains data on patients including their civil registration number, hospital and department, dates of admission and discharge, surgical procedures performed, using the Danish Classification of Surgical Procedures or the Nordic Classification of Surgical Procedures coding system and one or several discharge diagnoses using ICD-8 or ICD-10. The discharging physician assigned all discharge diagnoses.

All residents in Denmark are assigned a unique 10-digit code, given to each individual having, or having had, a registered address in Denmark since April 1968. Data are stored in the Danish Civil Registration System containing electronic records of any change in vital and emigration status in the Danish population.

We included all individuals aged 40–99 years, with first-time AF (outpatients or admitted patients) occurring between January 1, 1990 and December 31, 2002. During the same period, first-time upper limb thromboembolectomy was identified by searching for the surgical procedure codes, thromboembolectomy in brachial and thromboembolectomy in ulnar or radial artery, coded according to the Danish Classification of Surgical Procedures (codes: 86823 and 86824) or the Nordic Classification of Surgical Procedures (codes PBE20 and PBE30). The surgical procedures were analysed jointly, since emboli can often be found in all three arteries at the same time [12]. Follow-up for upper limb thromboembolectomy ended on December 31, 2002. Thromboembolectomy was considered to have been confirmed when one of the authors (LVA, LF) verified a description of the surgical procedure in a retrievable medical record. In those records, information about AF, including electrocardiograms was also reviewed. Data on upper limb thromboembolectomy from the medical records were used to validate registries prior to this study and the results have been published elsewhere [14]. Patients with thromboembolectomy prior to the diagnosis of AF were excluded from the analysis (64 men and 106 women). Data on outcomes such as stroke and mortality have been published elsewhere [15–18]. Summary data about the background population were uploaded from Statistics Denmark. For AF, we used the International Classification of Diseases, 8th Revision (ICD-8) or the International Classification of Diseases, 10th Revision (ICD-10) codes 427.93, 427.94 and I48, respectively.

Diagnoses of hypertension, diabetes, myocardial infarction, heart failure and stroke before or at the time of diagnosis of incident AF were also retrieved from the Danish National Registry of Patients using ICD-8 and ICD-10 codes (hypertension: 400–404, 410.09, 411.09, 412.09, 413.09, 414.09, 435.09, 437.00, 437.01, 437.08, 437.09, 438.09, I10-I15; diabetes: 249, 250, E10-E14; myocardial infarction: 410, I21 and I22; heart failure: 425.99, 427.09, 427.10, 427.11, 427.19, 427.99, 428.99, I50; and stroke: 432–434, 436, I63, I64). Vital and emigration status were established by querying the Danish Civil Registration System. The Health Care System in Denmark was, during the study period, tax-financed, non-profit and free of charge for all inhabitants. The Danish National Board of Health and the Danish Data Protection Agency approved the study.

Statistical analysis

We used the unique civil registration number to link records. Descriptive statistics are presented as numbers and percentages or means with standard deviation. Two-sided confidence intervals (95% CI) around the estimates were calculated. We computed relative risk of thromboembolectomy as the proportion of observed to expected number of patients. Excess risk was computed as the difference between observed and expected number of patients. To analyse impact of previous diagnoses of myocardial infarction, hypertension, and diabetes, heart failure and stroke, we used a Cox proportional hazards model. In separate models for each sex, we adjusted for age and comorbidity concerning: hypertension, diabetes, myocardial infarction, heart failure and stroke. Statistical analyses were performed using statistical packages BMPD 8.1 (Statistical Solutions Ltd., Cork, Ireland) and STATA 10, (Stata Corporation, College Station, TX, USA).

Results

We retrieved 131 476 patients with incident AF and without previous thromboembolectomy in the upper limb, of whom 68 042 were men (51.8%) and 63 434 were women (48.3%) (Table 1). In the study cohort, 130 men underwent upper limb thromboembolectomy over 220 890 person-years of observation, whilst 275 women underwent thromboembolectomy over 197 777 patient-years. The incidence per 100 000 person-years was 58.9 (95% CI, 49.2–69.8 for men and 139.1 (95% CI, 123.1–156.5) for women. Table 2 illustrates that the incidence of thromboembolectomy increases by increasing age.

Table 1.   Characteristics of 131 476 subjects with an incident hospital diagnosis of AF in Denmark, from 1990 till 2002
CharacteristicsMenWomen
  1. *Diagnoses before or at time of incident diagnosis of AF.

Number of subjects68 042 (51.8%)63 434 (48.3%)
Mean age70.6 (SD 11.5)76.4 (SD 10.6)
Age at AF diagnosis (%)
 40–493488 (5.1)1280 (2.0)
 50–598926 (13.1)3688 (5.8)
 60–6915 848 (23.3)9259 (14.6)
 70–7923 475 (34.5)21 542 (34.0)
 80–8914 512 (21.3)22 866 (36.0)
 90–991793 (2.6)4799 (7.6)
*Comorbidity (%):
 Hypertension5878 (8.6)5848 (9.2)
 Diabetes4315 (6.3)4061 (6.4)
 Myocardial infarction3688 (5.4)2673 (4.2)
 Heart failure12 581 (18.5)12 020 (19.0)
 Stroke3593 (5.3)4236 (6.7)
Table 2.   Thromboembolectomy in patients with AF in Denmark
Age-groupPersons at riskPerson years at riskObserved numberIncidence per 100 000 person-years (95% CI)Expected numberRelative risk (95% CI)Estimated excess riskExcess risk (%)
Men
 All ages68 042220 89013058.9 (49.2–69.8)17.337.5 (6.3–8.9)112.67100
 40–49348817 544211.4 (1.4–41.2)0.1513.3 (1.6–48.2)1.852
 50–59892638 3871026.1 (12.5–47.9)0.7313.7 (6.6–25.2)9.278
 60–6915 84861 6002947.1 (31.5–67.6)3.029.6 (6.4–13.8)25.9823
 70–7923 47570 7634665.0 (47.6–86.7)6.726.8 (5.8–9.1)39.2835
 80–8914 51230 27039128.8 (91.6–176.1)5.976.5 (4.6–8.9)33.0329
 90–99117323264172.0 (46.9–439.7)0.735.5 (1.5–14.0)3.273
Women
 All ages63 434197 777275139.1 (123.1–156.5)29.449.3 (8.3–10.5)245.56100
 40–491280564200.0 (> 0.0–65.5)0.050.0 (> 0.0–73.8)−0.050
 50–59368815 899531.5 (10.2–73.4)0.3016.7 (5.4–38.9)4.702
 60–69925938 0903694.5 (66.2–130.8)1.8119.9 (13.9–27.5)34.1914
 70–7921 54274 88290120.2 (96.7–147.7)9.579.4 (7.6–11.6)80.4333
 80–8922 86656 154117208.4 (172.4–249.7)15.157.7 (6.4–9.3)101.8541
 90–994799710927379.8 (250.4–552.1)2.5610.5 (7.0–15.3)24.4410

The relative risk of needing a thromboembolectomy among patients with AF compared to the background population was 7.5 (95% CI, 6.3–8.9) for men, and 9.3 (95% CI, 8.3–10.5) for women (Table 2). Excess thromboembolectomy increased with age in both men and women. Depending on age group and sex, the relative risk was up to 19.9-fold higher in patients with AF compared to the general population. This relative risk was highest in the three decades from age 40 to 69 for men, and for women in the decades from age 50 to 69 (Table 2).

Figure 1 shows a higher cumulative incidence of thromboembolectomy in women with AF compared to men with AF. After 10 years the absolute cumulative incidence of thromboembolectomy was 0.4% (95% CI, 0.34–0.58) in men, and 1% (95% CI, 0.8–1.1) in women. When adjusted for age, hypertension, diabetes, myocardial infarction, heart failure and stroke, women had a 1.8 (95% CI, 1.5–2.3) fold higher risk of thromboembolectomy than men.

Figure 1.

 Cumulated incidence of thromboembolectomy in patients with incident AF at age 40–99 years in Denmark, from 1990 to 2002.

Table 3 presents the impact of hypertension, diabetes, myocardial infarction, heart failure and stroke on risk of thromboembolectomy. When adjusted for age and other comorbidity, history of hypertension (HR 2.2–2.9), myocardial infarction (HR 2.9–3.9), heart failure (HR 1.6–1.9) and stroke (HR 2.2–3.8) was associated with increased adjusted risk of thromboembolectomy in both men and women (Table 3). For those diagnosed with diabetes, the risk for thromboembolectomy was non-significantly increased by 1.4 and 1.2, for men and women, respectively (Table 3).

Table 3.   Impact of comorbidity on subsequent risk of upper limb thromboembolectomy in patients with AF in Denmark
 MenWomen
  1. *Each estimate is adjusted for age and the other risk factors.

Comorbidity*Hazard ratio (95% CI)PHazard ratio (95% CI)P
Hypertension2.2 (1.4–3.4)0.0012.9 (2.2–3.9)< 0.001
Diabetes1.4 (0.8–2.6)0.221.2 (0.8–1.8)0.47
Myocardial infarction3.9 (2.5–6.0)< 0.0012.9 (2.0–4.2)< 0.001
Heart failure1.9 (1.3–2.9)0.0011.6 (1.2–2.1)< 0.001
Stroke3.8 (2.4–6.0)< 0.0012.2 (1.6–3.2)< 0.001

Discussion

This is the first population-based assessment of risk of upper limb thromboembolectomy associated with AF. In this present analysis, we found that AF substantially increases the risk of upper limb thromboembolectomy. This risk is particularly high in females, and associated with age, hypertension, myocardial infarction, heart failure and stroke.

Since the early 1990s it has been evident that AF is not only a risk factor for stroke, but also that there is causal relationship between AF and stroke [3]. At the same time, the effect of warfarin in reducing risk of stroke in patients with AF became well established from both clinical trials and cohort data [19–21]. Similar research on risk factors for non-CNS systemic embolism is sparse, even though several studies have reported a high prevalence of AF in those patients [7,22,23]. In the present analysis, the increase in relative risk of thromboembolectomy was higher among those with AF (up to 20-fold depending upon age-group and sex) compared to the general population. The incidence of upper limb embolism among patients with AF was highly age-dependent, ranging from 0 to 380 per 100 000 person years. This is consistent with an age-related increase in stroke amongst patients with AF, where the absolute benefit from oral anticoagulation increases with increasing age whilst the value of antiplatelet therapy declines from age 65 and upwards [24]. However, it has been reported that the proportion of vitamin K antagonist users was much lower among elderly patients (80–99 years) with AF compared to younger patients with AF [25].

The risk of thromboembolectomy among patients with AF was increased if they had a history of hypertension, myocardial infarction, heart failure or stroke. The association between thromboembolism and hypertension and stroke would be in accordance with the findings from the studies on stroke and AF, whereby hypertension increases the risk 2-fold [5]. The data for myocardial infarction and heart failure as risk factors for stroke in AF are less consistent, although some evidence does suggest the risk is evident [5,26–31].

Our analyses show that women with AF required thromboembolectomy twice as often as men with AF. One explanation for this finding could be the smaller diameter of arteries in women [32]. Even smaller emboli could, in women, be large enough to occlude the brachial artery, whereas in men, smaller emboli may pass to one of the antebrachial arteries and be clinically silent due to communication between the ulnar and radial arteries through the arterial arcade of the hand. Another explanation could be that men may not react to the symptoms as quickly as women, and spontaneous improvement in acute ischaemia of the upper limb is a well-known phenomenon. Also, age and gender-dependent differences in frequency and quality of antithrombotic treatment have been reported [25].

Increased risk of stroke and thromboembolism in female patients with AF has been reported [33–35]. Indeed, the Stroke in AF Working Group systematic review found female sex to increase risk of stroke in AF by 1.6-fold [5]. The present analysis reports for the first time an increased risk of upper limb thrombeomboectomy in female patients with AF. Several possible mechanisms have been proposed, including a higher prevalence of hypertension among women with AF. Also differences in structure and function of the left atrium, endothelial damage/dysfunction, platelet activation, and level of von Willebrand factor have been suggested [36]. In our study, even after adjustment for comorbidity, women still had higher risk of thromboembolectomy than men.

The registries used in the study, the Danish Vascular National Registry and the Danish National Registry of Patients have been validated previously, and the estimated number of thromboembolectomy cases not registered in any of the two registries was low [14,37]. We ensured validity of the present analysis by reviewing medical records for patients with upper limb thromboembolectomy. We found that AF was considerably underreported to the Danish National Registry of Patients. In a single hospital study, the completeness of reporting of the diagnosis AF in the National Registry of Patients was previously found to be 84%, and appropriate for epidemiological research [38]. Only 34.1% of thromboembolectomy patients had a record of AF in the Danish National Registry of Patients, although clear evidence of AF was present in 62.7% of the medical records. The underreporting of AF to the National Registry of Patients in those patients can be caused by the fact that many patients were transferred from one department to another (from medical to surgical to medical ward and vice versa) leading to lack of responsibility in reporting of conditions of co-morbidity, including AF. Another reason for lack of recording of AF could be that our the mostly old patients may have so many conditions of comorbidity that it was not possible within a reasonable amount of time to record all those conditions. This underreporting of AF to the Danish National Registry of Patients could potentially lead to underestimation of the excess risk ascribable to AF.

Limitations

The main limitation of our study is its registry cohort design, although strength is the use of nationwide registries with long and complete follow-up. We lack precise and detailed data on the use of antithrombotic therapy, which could vary over time. Indeed, a meta-analysis of randomised controlled trials in patients with AF demonstrated that warfarin reduced risk of systemic embolism by 50% when compared to antiplatelet agents, and by 71% when compared to placebo [39]. We also did not adjust for comorbidities in analyses of risk among patients with AF relative to risk in the general population. However, the primary objective of the study was to determine excess risk of thromboembolectomy among patients with AF, not among patients with conditions leading to AF.

Of further concern is the validity of registration of co-morbidity: hypertension, myocardial infarction, diabetes, heart failure and stroke. The positive predictive value of a diagnosis of hypertension in the Danish National Registry of Patients was found to be low (40%–60%) [40]. However, a higher positive predictive value (82%) for a diagnosis of myocardial infarction was found in the National Registry of Patients [41]. The positive predictive value of a discharge diagnosis of heart failure was found to be 84% [42]. A high positive predictive value of diabetes (96%) [43], and stroke (88–97%) [44,45] was found in the Danish National Registry of Patients. Even higher positive predictive value of stroke among patients with AF has been reported, possibly because stroke severity is worse in patients with AF [18].

In conclusion, AF substantially increases the risk of upper limb thromboembolectomy. This risk is higher with advancing age, in females, and associated with prior hypertension, myocardial infarction, heart failure and stroke.

Acknowledgements

This study has been partially supported by Viborg County, the Medical Association of Central Denmark Region, the Danish Medical Association Research Fund/Højmosegaard-Legatet, and the Research Foundation, Silkeborg Hospital.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.

Ancillary