Recurrent syncope is a major cause of hospitalizations and may be associated with cardiovascular co-morbidities. Despite this, prognostic factors and the clinical characteristics among patients are not well described. Therefore, we identified and analyzed data on all patients >50 years of age discharged after a first-time episode of syncope in the period 2001 to 2009 through nationwide administrative registries. We identified the clinical characteristics of 5,141 patients ≥85 years of age and 23,454 patients <85 years of age. Multivariate Cox models were used to assess prognostic factors associated with the end point of recurrent syncope according to age. We found that those with syncope and ≥85 years were more often women (65% vs 47%) and generally had a greater prevalence of noncardiovascular co-morbidities, whereas the prevalence of cardiovascular co-morbidities was more heterogeneously distributed across age groups. Overall, significant baseline predictors of recurrent syncope were aortic valve stenosis (hazard ratio [HR] 1.48, 95% confidence interval [CI] 1.31 to 1.68), impaired renal function (HR 1.34, 95% CI 1.15 to 1.58), atrioventricular or left bundle branch block (HR 1.32, 95% CI 1.16 to 1.51), male gender (HR 1.18, 95% CI 1.12 to 1.24), chronic obstructive pulmonary disorder (HR 1.10, 95% CI 1.02 to 1.19), heart failure (HR 1.10, 95% CI 1.02 to 1.21), atrial fibrillation (HR = 1.09, 95% CI 1.01 to 1.19), age per 5-year increment (HR 1.09, 95% CI 1.07 to 1.10), and orthostatic medications per increase (HR 1.06, 95% CI 1.03 to 1.09). Atrial fibrillation and impaired renal function both exhibited less prognostic importance for recurrent syncope in the elderly compared with younger population (p for interactions <0.01). In conclusion, predictive factors of recurrent syncope were closely associated with increased cardiovascular risk profile age and gender. The use of multiple orthostatic medications additively increased the risk of recurrences representing a need for strategies to reduce unnecessary polypharmacy.
Syncope and recurrent syncope are very frequent events and a common cause for hospitalization. Correspondingly, recurrent syncope, along with risk of death, is acknowledged as the 2 important elements to be considered in syncope guidelines on risk stratification for choosing the strategy of diagnostic testing. The impact severity of recurrent syncope ranges from fractures and hospitalizations to injury and inconvenience to restrictions on driving and the use of heavy machinery, which may result in an inability to work and decreased quality of life. Current epidemiologic data on prognostic factors of recurrent syncope are insufficient and based on smaller studies in specific subsets of syncope. General worldwide demographics show an increase in the proportion of elderly patients causing a steady increase in health-care–related costs and increase in polypharmacy. Despite the exponential increasing incidence of syncope in the elderly and very elderly, the characteristics and prognostics of recurrent syncope and the impact of polypharmacy have not been fully investigated in these patients. We hypothesized that, in a real-life nationwide cohort, recurrent syncope was associated with increasing age, cardiovascular co-morbidities (risk factors), and medications known to give orthostatic hypotension.
Methods
All Danish residents are given a personal 10-digit civil registration number at birth or when moving to the country, which enables linkage of nationwide administrative registers on the individual level. Information on all dispensed prescriptions from Danish pharmacies since 1995 is registered according to the Anatomical Therapeutic Chemical (ATC) system in The Register of Medicinal Product Statistics. Information on hospitalization and co-morbidities were obtained from the Danish National Patient Register, in which information on all hospital admissions in Denmark has been stored since 1978. Each discharge from hospitalization is coded with 1 primary diagnosis and if appropriate, ≥1 secondary diagnoses according to the International Classification of Diseases since 1994 the Tenth revision (ICD-10). Demographic information on date of birth, age, gender, and vital status were obtained from the Danish Civil Register.
We identified all Danish residents with age of >50 years with a first-time discharge for syncope from all public hospital departments when classified as the primary discharge diagnosis (ICD-10 code R55.9—“syncope and collapse”) from January 1, 2001 to December 31, 2009. If any recurrent syncope occurred within 7 days after discharge of the initial syncope, the patient was excluded from the analysis to avoid readmissions due to the same syncopal event and/or interhospital transfers. We have previously validated the R55.9 diagnosis and found a positive predictive value of syncope of 95% and that 47% are of unknown origin after a mean follow-up of 2.5 years.
Identification of and information on major co-morbidities related to syncope up to 5 years before inclusion were based on hospital discharge diagnosis codes according to a modified Charlsons comorbidity index. We obtained information through the Danish National Patient Register based on primary or secondary ICD-10 codes as defined in the Supplementary material . Validation of several of the used ICD-10 discharge diagnoses has been done previously.
Information on concomitant drug use up to 180 days before and after inclusion was provided through The Register of Medicinal Product Statistics using ATC codes as defined in the Supplementary material .
Chronic obstructive pulmonary disorder was redefined as a claimed prescription for a bronchial dilating medication for inhalation (ATC code R03) and/or an admission for a chronic obstructive pulmonary disorder (J42-44). Patients were defined as having hypertension if receiving at least 2 types of antihypertensive drugs as done and validated previously. We defined diabetes mellitus as a claimed prescription for a glucose-lowering drug (ATC code A10) and/or admission for diabetes with or without complications (E10-14).
Pharmacotherapy commonly used and commonly known to have orthostatic hypotension or syncope as a side effect based on available medicine leaflets were gathered in groups consisting of 0, 1, 2, and ≥3 of these medications at baseline and 180 days after initial syncope to assess the importance of use of multiple medications on the risk of recurrent syncope and changes in these medications after the first syncope. Pharmacotherapy defined as having orthostatic hypotension or syncope as side effects were β blockers, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, loop diuretics, thiazides, calcium channel blockers, nitrates, class I and III antiarrhythmic medications, antipsychotic medications consisting of first and second generation of antipsychotics, anticholinergic antiparkinson medications, α blockers, and antidepressants (all groups).
The primary outcome was recurrent syncope defined as hospital admission and primary discharge diagnosis of syncope (ICD-10 diagnosis R.55.9) in the period after discharge of the initial first-time syncope with 7 days quarantine.
Changes in defined orthostatic medications were evaluated comparing medications redeemed up to 180 days before initial syncope with orthostatic medications redeemed up to 180 days after initial syncope.
The population was divided into 4 age groups: <65, 65 to 74, 75 to 84, and ≥85 years of age. Statistical trends in the clinical characteristics according to these age groups were calculated with the Cochran-Armitage trend test for the categorical values. The population was also divided into 2 age groups: <85 and ≥85 years of age. Comparisons between these 2 age groups were done by the chi-square test for the categorical variables. Cox proportional hazards regression models were used to identify predictors of recurrent syncope by a stepwise selection of variables in the model with limits set at 0.05. The available covariates for selection in the final model were all the covariates representing the co-morbidities as defined including age, gender, medications known to cause orthostatic hypotension (as a continuous variable), and the year of inclusion. Relative risks are presented as hazard ratios with 95% confidence intervals. We systematically tested interactions of age with the available covariates in the data set that were a priori suspected to have a significant impact on the end point of recurrent syncope. Interactions were being considered if they were significant (p value <0.05) in univariate and multivariate models. Significant interactions were found between age and atrial fibrillation, age and impaired renal function, as well as age and number of medications known to cause orthostatic hypotension or syncope (as a continuous variable). Therefore, the predictive model found by stepwise selection was also stratified by age groups to evaluate the interactions with age. A 2-sided p value <0.05 was considered statistically significant. All analyses were performed using SAS (version 9.2, SAS Institute, Cary, North Carolina).
The study was accepted by the Danish Data Protection Agency (Ref. 2007-58-0015, int. ref: GEH-2010-001). Ethical approval is not compulsory for retrospective register-based studies in Denmark.
Results
During a mean follow-up of 3.4 years (SD 2.5), 6,267 patients (22%) had at least 1 recurrent syncope with a mean time to first recurrent syncope of 2.0 years (SD 2.0), and 50% of these had the recurrence within 1.2 years. Proportionately, recurrent syncope occurred evenly in the 4 age groups; 18%, 22%, 24%, and 23% for ages <65, 65 to 74, 75 to 84, and ≥85 years, respectively. Male gender was less common among the very elderly patients but more frequent in age groups 50 to 64 and 65 to 74 years (p <0.05; Table 1 ). The prevalence of hypertension, coronary heart disease, previous myocardial infarction, diabetes, chronic obstructive pulmonary disorder, cancer, and patients with an implantable cardioverter-defibrillator showed an increase in prevalence with increasing age until the 8th decade of life, where after the prevalence decreased in the very elderly. Other cardiac co-morbidities such as congestive heart failure, atrial fibrillation, and atrioventricular or left bundle branch block increased continuously with age. Comparative to the general trend in cardiac co-morbidities, a decrease in the use of medications such as statins, β blockers, and angiotensin-converting enzyme inhibitors after the 8th decade was observed, whereas the use of diuretics and anxiolytics increased continuously with age. The cumulative number of medications known to cause orthostatic hypotension or syncope increased continuously with age. A total of 40% of the very elderly received ≥3 of these medications ( Table 1 ).
| Variable | Age (yrs), % | p Value for Trend | Age (yrs), % | p Value for Comparison | ||||
|---|---|---|---|---|---|---|---|---|
| <65 | 65–74 | 75–84 | ≥85 | <85 | ≥85 | |||
| Patients | 7,526 | 6,854 | 9,074 | 5,141 | 23,454 | 5,141 | ||
| Men | 59.5 | 55.6 | 46.5 | 35.2 | <0.001 | 53 | 35 | <0.001 |
| Women | 40.5 | 44.4 | 53.5 | 64.8 | <0.001 | 47 | 65 | <0.001 |
| Hypertension | 28 | 42 | 46 | 39 | <0.001 | 39 | 39 | 0.382 |
| Coronary heart disease | 11 | 19 | 22 | 19 | <0.001 | 18 | 19 | 0.090 |
| Previous myocardial infarction | 4 | 6 | 7 | 6 | <0.001 | 6 | 6 | 0.023 |
| Heart failure | 6 | 10 | 11 | 14 | <0.001 | 9 | 14 | <0.001 |
| Cerebrovascular disease | 7 | 11 | 14 | 14 | <0.001 | 11 | 14 | <0.001 |
| Peripheral artery disease | 1 | 2 | 2 | 2 | <0.001 | 2 | 2 | 0.805 |
| Atrioventricular or left bundle branch block | 2 | 3 | 3 | 4 | <0.001 | 3 | 4 | <0.001 |
| Aortic valve stenosis | 1 | 2 | 4 | 4 | <0.001 | 2 | 4 | <0.001 |
| Atrial fibrillation | 5 | 11 | 17 | 20 | <0.001 | 11 | 20 | <0.001 |
| Diabetes | 10 | 13 | 12 | 9 | 0.245 | 12 | 9 | <0.001 |
| Impaired renal function | 1 | 2 | 2 | 2 | <0.001 | 2 | 2 | 0.027 |
| Chronic obstructive pulmonary disorder | 10 | 15 | 15 | 10 | 0.043 | 13 | 10 | <0.001 |
| Cancer | 3 | 5 | 5 | 3 | 0.107 | 5 | 4 | 0.003 |
| Implanted pacemaker | 1 | 2 | 2 | 3 | <0.001 | 2 | 3 | <0.001 |
| Implanted cardioverter-defibrillator | 0 | 1 | 0 | 0 | <0.001 | 1 | 0 | <0.001 |
| Statins | 20 | 31 | 23 | 10 | <0.001 | 24 | 10 | <0.001 |
| β Blockers | 19 | 27 | 30 | 25 | <0.001 | 26 | 25 | 0.296 |
| Angiotensin-converting enzyme inhibitors or angiotensin receptor blockers | 26 | 27 | 39 | 31 | <0.001 | 34 | 31 | <0.001 |
| Loop diuretics | 9 | 18 | 24 | 32 | <0.001 | 17 | 32 | <0.001 |
| Spironolactone | 4 | 6 | 7 | 7 | <0.001 | 6 | 7 | 0.031 |
| Thiazide | 12 | 19 | 23 | 26 | <0.001 | 18 | 26 | <0.001 |
| Calcium channel blockers | 13 | 21 | 23 | 21 | <0.001 | 19 | 21 | <0.001 |
| α Blockers | 1 | 1 | 1 | 1 | <0.001 | 1 | 2 | 0.046 |
| Nitrates | 5 | 10 | 15 | 18 | <0.001 | 10 | 18 | <0.001 |
| Digoxin | 3 | 6 | 10 | 14 | <0.001 | 6 | 14 | <0.001 |
| Class I antiarrhythmic | 0 | 1 | 1 | 0 | 0.375 | 0 | 0 | 0.030 |
| Class III antiarrhythmic | 1 | 2 | 2 | 1 | 0.026 | 1 | 1 | 0.084 |
| Aspirin | 15 | 28 | 34 | 36 | <0.001 | 26 | 36 | <0.001 |
| Vitamin K antagonists | 4 | 8 | 10 | 7 | <0.001 | 8 | 7 | 0.003 |
| Antipsychotics | 6 | 6 | 6 | 6 | 0.488 | 6 | 6 | 0.807 |
| Anxiolytics | 24 | 29 | 33 | 37 | <0.001 | 29 | 37 | <0.001 |
| 0 Medications ∗ | 41 | 25 | 17 | 16 | <0.001 | 27 | 15 | <0.001 |
| 1 Medication ∗ | 21 | 21 | 20 | 22 | 0.248 | 21 | 22 | 0.011 |
| 2 Medications ∗ | 16 | 21 | 23 | 23 | <0.001 | 20 | 23 | <0.001 |
| ≥3 Medications ∗ | 22 | 34 | 40 | 40 | <0.001 | 32 | 40 | <0.001 |
∗ Consisted of those known to cause orthostatic hypotension: β blockers, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, loop diuretics, thiazides, calcium channel blockers, nitrates, α blockers, class I and III antiarrhythmic medications, antipsychotic medications consisting of first and second generation of antipsychotics, anticholinergic antiparkinson medications, and antidepressants (all groups).
Significant independent predictors of recurrent syncope are listed in Table 2 . In both univariate and multivariate analyses, age interacted significantly with atrial fibrillation (p <0.01), impaired renal function (p <0.01), and the number of orthostatic medications used at first syncope (p <0.01). The multivariate age-dependent Cox models are listed in Table 3 . Generally, the prognostic value of these identified clinical characteristics decreased with increasing age, reaching levels of insignificance in the very elderly, except for the number of orthostatic medications. Atrial fibrillation was highly significant in the younger age groups and became an insignificant predictor in the very elderly. The number of orthostatic medications remained a solid prognostic factor associated with recurrent syncope throughout the age stratification models.
