Abstract
Background
Syncope risk stratification is difficult and has not been implemented clinically.
Hypothesis
The CHADS2 score can be applied as a risk stratification tool for predicting mortality after an episode of syncope.
Methods
All patients discharged from emergency departments with a first‐time diagnosis of syncope from 2001 to 2009 where identified from nationwide registers in Denmark and matched on sex and age with a control population. Risk of all‐cause or cardiovascular death was analyzed by multivariable Cox models.
Results
A total of 37 705 patients were included. There were a total of 7761 deaths (21%), of which 52% were cardiovascular vs 27 862 (15%) deaths in the control population. The risk of cardiovascular death was significantly increased with increasing CHADS2 score (CHADS2 score: 1–2, hazard ratio [HR]: 9.11, 95% confidence interval [CI]: 8.25‐10.07; CHADS2 score: 3–4, HR: 17.32, 95% CI: 15.42‐19.47; CHADS2 score: 5–6, HR: 26.66, 95% CI: 21.40‐33.21) relative to CHADS2 score of 0. A CHADS2 score of 0 was associated overall with very low event rates (15.1 deaths per 1000 person‐years) but was associated with increased relative risk in the syncope population compared to controls. Syncope predicted 1‐week, 1‐year, and long‐term mortality across CHADS2 scores compared to controls but did not reach significance in CHADS2 scores of 5 to 6.
Conclusions
Increasing CHADS2 score significantly predicts mortality in patients discharged with a diagnosis of syncope, and a CHADS2 score of 0 was associated with a very low absolute mortality. Compared to controls, syncope was associated with increased short‐ and long‐term mortality, particularly in the lower CHADS2 scores.
Introduction
The CHADS2 score is a validated clinical tool used for prediction of stroke in the presence of atrial fibrillation and is used to guide initiation of anticoagulant therapy. The score evaluates the risk of stroke through the sum of individual risk factors for stroke (congestive heart failure, hypertension, age ≥75 years, diabetes, and previous stroke or transient ischemic attack [doubled]).1, 2 Recently it has been shown that the CHADS2 score may also be predictive for major cardiac events such as myocardial infarction and cardiovascular mortality.3, 4, 5, 6, 7, 8, 9 Syncope patients, particularly with cardiovascular disease, remain at high risk of arrhythmias and sudden death,10, 11, 12 with risk of death markedly higher than the general population.13
The risk stratification for arrhythmias and sudden death after a syncopal event is important, and several risk assessment methods have been used to stratify risk of death after syncope, but none have been implemented widely. As syncope is a common event,14 we are still in need of a simple implementable tool for arrhythmic and sudden death risk stratification that is useful in everyday life. In the setting of the emergency department (ED), an easy rule‐out algorithm is warranted (ie, which syncope patients should not be admitted for further evaluation and who should be selected for admission due to high risk of cardiovascular death).
The association between the CHADS2 score and the prevalence of death in the postsyncope setting is unknown. As the CHADS2 score is widely known and used for prediction of thromboembolic episodes and initiation of treatment with anticoagulants in patients with atrial fibrillation, it would be easy to implement as a risk stratification of cardiovascular death. We hypothesized that the CHADS2 score can be used as a tool for predicting cardiovascular death, all cause‐mortality in patients who had been discharged from the ED after a syncopal episode, and evaluate whether the CHADS2 score may be applicable in risk stratification.
Methods
A personal and unique civil registration (CPR) number is assigned to all residents in Denmark, 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.15 We obtained information on hospitalization and comorbidities from the Danish National Patient Register, where information on all hospital admissions in Denmark has been stored since 1978.16 Demographic information on date of birth, age, sex, and vital status were obtained from the Danish Civil Register, and cause of death was gathered from the National Danish Registry of Causes of Death.
Study Population
We identified all Danish residents with a first‐time discharge for syncope from the ED when classified as the primary discharge diagnosis (International Classification of Diseases, 10th Revision [ICD‐10] code R55.9 [syncope and collapse]) between January 1, 2001 and December 31, 2009. We have previously validated the R55.9 diagnosis, which carries a positive predictive value of 95%.17
We matched the cohort of syncope patients with 5 controls selected from the background Danish population matched on sex and age. The controls were given the same date of inclusion in the study as the case they were matched to.
Comorbidity and Pharmacotherapy
Identification and information on major comorbidities related to syncope up to 5 years prior to inclusion were based on hospital discharge diagnosis codes. We obtained information through the Danish National Patient Register based on the primary or secondary diagnosis for ICD‐10 codes listed in the Appendix. Validation of several ICD‐10 discharge diagnoses has been done previously18, 19, 20, 21 and has been used in several larger studies.22, 23, 24, 25
Combinations of medications and comorbidity were used to describe hypertension and heart failure. Congestive heart failure (CHF) was defined as a combination of a diagnosis with CHF and the use of loop diuretics as done and validated previously.26 Patients were defined as having hypertension if receiving at least 2 types of antihypertensive drugs as done previously.2 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) as done previously.27
Information on concomitant drug use up to 180 days prior to inclusion was provided through The Register of Medicinal Product Statistics using the ATC codes listed in the Appendix.
Calculation of the CHADS2 Score
The CHADS2 score was calculated with 1 point given for each of the following parameters if present at inclusion according to information on comorbidity: CHF, hypertension, age ≥75 years, and diabetes. Two points were given for prior transient ischemic attack or stroke.
Outcome Measures
The primary long‐term outcomes were cardiovascular death and all‐cause mortality. The secondary outcomes were 1‐year cardiovascular death, 1‐year all‐cause mortality, 1‐week cardiovascular death and 1‐week all‐cause mortality.
Other measures were cardiovascular hospitalization defined by any hospital discharge with any cardiovascular ICD‐10 discharge diagnosis (I10 through I89). These outcomes were gathered from the National Danish Registry of Causes of Death, the Danish National Patient Register and the CPR registry.
Statistics
Patients and controls were divided into 4 groups based on CHADS2 score 0, CHADS2 score 1–2, CHADS2 score 3–4 and CHADS2 score ≥5.
Outcomes were displayed by Kaplan‐Meier plots and compared using log‐rank tests. The χ2 test was used to determine P values between CHADS2 groups for comparison in baseline characteristics. Cox proportional regression analyses were used to determine hazard ratios (HR) and their 95% confidence intervals (CI) for the measured end points.
Two separate models were used. The first model used CHADS2 = 0 as the reference group in the syncope population. The adjusted HRs have been adjusted for year of admission, sex and baseline comorbidities in Table 1 excluding those involved in the CHADS2 risk score calculation. The second Cox model was used for comparison between the individual CHADS2 groups in the patients with the CHADS2 groups in the controls. This model was adjusted for sex, age, year of admission, and baseline comorbidities in Table 1 excluding those involved in the CHADS2 risk score calculation. Age was included as a covariate in this model because the matching between cases and controls was broken when patients were divided into the CHADS2 groups. In all Cox models, the model assumptions (proportional hazards, linearity of continuous covariates, and lack of interactions) were found to be valid. A 2‐sided P value <0.05 was considered statistically significant.
Table 1.
Baseline Characteristics of the Syncope Population
| Characteristics | CHADS2 = 0, N = 21 197 (56.2%) | CHADS2 = 1–2, N = 13 589 (36.0%) | CHADS2 = 3–4, N = 2710 (7.2%) | CHADS2 = 5–6, N = 209 (0.6%) |
|---|---|---|---|---|
| Age, median (range), y | 45 (30–59) | 77 (67–83) | 79 (75–85) | 81 (77–85) |
| Male sex (%) | 9885 (47) | 6203 (46) | 1418 (52) | 110 (53) |
| Hypertension (%) | 0 | 6630 (49) | 1757 (65) | 193 (92) |
| Ischemic heart disease (%) | 390 (2) | 1699 (13) | 729 (27) | 100 (48) |
| Cerebral vascular disease (%) | 0 | 549 (4) | 1694 (63) | 209 (100) |
| Previous myocardial infarction (%) | 132 (1) | 667 (5) | 293 (11) | 41 (20) |
| Cardiac conduction disorder (%) | 117 (1) | 315 (3) | 118 (4) | 13 (6) |
| Previous atrial fibrillation (%) | 254 (1) | 1148 (8) | 587 (22) | 84 (40) |
| Heart failure (%) | 0 | 934 (7) | 837 (31) | 154 (74) |
| Diabetes (%) | 0 | 1511 (11) | 847 (31) | 116 (56) |
| COPD (%) | 1335 (6) | 1423 (10) | 370 (14) | 29 (14) |
| Cancer (%) | 327 (2) | 564 (4) | 146 (5) | 9 (4) |
| Renal disease (%) | 39 (0) | 229 (2) | 108 (4) | 12 (6) |
| Concomitant pharmacotherapy | ||||
| Statins (%) | 799 (4) | 2715 (20) | 900 (33) | 82 (39) |
| β‐blockers (%) | 650 (3) | 3555 (26) | 1002 (37) | 106 (51) |
| ACEI/ARB (%) | 498 (2) | 5609 (41) | 1,515 (56) | 165 (79) |
| Loop diuretics (%) | 381 (2) | 2273 (17) | 940 (35) | 137 (66) |
| Spironolactone (%) | 69 (0) | 621 (5) | 344 (13) | 52 (25) |
| Thiazide (%) | 447 (2) | 3360 (25) | 726 (25) | 53 (25) |
| Calcium channel blockers (%) | 379 (2) | 3192 (23) | 800 (30) | 90 (43) |
| Nitrates (%) | 173 (1) | 1367 (10) | 452 (17) | 60 (28) |
| Digoxin (%) | 99 (0) | 884 (7) | 432 (16) | 61 (29) |
| ASA (%) | 686 (3) | 3694 (27) | 1230 (45) | 104 (50) |
| VKA (%) | 220 (1) | 790 (6) | 351 (13) | 40 (19) |
| Antidepressants (%) | 2238 (11) | 2397 (18) | 714 (26) | 64 (31) |
| Antiepileptics (%) | 678 (3) | 580 (4) | 217 (8) | 25 (10) |
| Anxiolytics (%) | 2957 (14) | 3958 (29) | 900 (33) | 70 (33) |
Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; ASA, acetylic salicylic acid; COPD, chronic obstructive pulmonary disorder; VKA, vitamin K antagonists.
Ethics
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
A total of 37 705 patients and 188 525 controls were included in the study. The median follow‐up was 4.2 years (interquartile range [IQR], 1.8–6.5 years). Median age was 65 years (IQR, 48–78 years), and 48.4% were male. The median age distribution in the 4 assigned CHADS2 groups was 45 years (IQR, 30–59 years), 77 years (IQR, 67–83 years), 79 years (IQR, 75–85 years), and 81 years (IQR, 77–85 years), respectively. Baseline characteristics according to CHADS2 score is depicted in Table 1. Increasing CHADS2 score was associated with increasing frequency of the comorbidities.
Primary End Point
A total of 7761 (21%) deaths occurred among syncope patients during the follow‐up, of which 4020 (52%) were cardiovascular, compared to the control population where 27 862 (15%) died (50% cardiovascular). The patients and deaths were distributed unevenly across CHADS2 score, 56.2% (19.5% of deaths) with CHADS2 = 0, 36.0% (61.1% of deaths) with CHADS2 = 1–2, 7.2% (17.6% of deaths) with CHADS2 = 3–4, and 0.6% (1.8% of deaths) with CHADS2 = 5–6.
Increasing CHADS2 score was significantly associated in a proportional manner with increased all‐cause mortality in unadjusted analysis (Figure 1) and adjusted analyses as shown in Table 2. Similar significant association between increasing CHADS2 score and risk of cardiovascular death was likewise found (Table 2). Overall, there was a very low risk of death due to cardiovascular cause in patients with a CHADS2 score of 0, and the absolute event rates given in Table 3 show an event rate for all‐cause deaths of 15.1 and 4.8 for cardiovascular deaths per 1000 person‐years for CHADS2 scores of 0. The event rates showed a dramatic increase from CHADS2 score of 0 to CHADS2 score of 1–2, which continued in all‐cause death and cardiovascular death to CHADS2 score of 3–4 and 5–6.
Figure 1.
Kaplan‐Meier plot. The probability of all‐cause mortality according to CHADS2 risk score in syncope population compared to the control population. Significantly increased risk of all‐cause mortality is associated with increasing CHADS2 risk score and significant difference between syncope and control.
Table 2.
Univariate and Multivariate Cox Proportional Hazard Models Depicting Hazard Ratios of All‐Cause Mortality and Cardiovascular Death in Patients With Syncope According to CHADS2 Score, With CHADS2 Score = 0 as Reference
| CHADS2 Score | Hazard Ratio (Unadjusted), CHADS2 Score = 0 Is Reference | 95% Confidence Interval | Hazard Ratio (Adjusted), CHADS2 Score = 0 Is Reference | 95% Confidence Interval | |
|---|---|---|---|---|---|
| Death from all‐cause long term | 1–2 | 6.37 | 6.02‐6.76a | 5.60 | 5.28‐5.94a |
| 3–4 | 11.09 | 10.30‐11.94a | 8.29 | 7.66‐8.96a | |
| 5–6 | 19.23 | 16.16‐22.97a | 12.97 | 10.81‐15.56a | |
| Death from all‐cause within 1 year | 1–2 | 5.46 | 4.91‐6.07a | 4.53 | 4.06‐5.04a |
| 3–4 | 10.24 | 9.02‐11.61a | 6.90 | 6.03‐7.10a | |
| 5–6 | 17.45 | 13.38‐22.76a | 10.35 | 7.84‐13.68a | |
| Death from all‐cause within 1 week | 1–2 | 6.87 | 4.88‐9.87a | 5.75 | 3.80‐8.31a |
| 3–4 | 13.16 | 8.71‐19.90a | 9.69 | 6.25‐15.02a | |
| 5–6 | 20.11 | 8.95‐45.19a | 14.61 | 6.29‐33.97a | |
| Cardiovascular death long term | 1–2 | 10.57 | 9.56‐11.66a | 9.11 | 8.25‐10.07a |
| 3–4 | 24.85 | 22.26‐27.75a | 17.32 | 15.42‐19.47a | |
| 5–6 | 47.80 | 38.70‐59.03a | 26.66 | 21.40‐33.21a | |
| Cardiovascular death within 1 year | 1–2 | 10.32 | 8.51‐12.53a | 8.59 | 7.06‐10.45a |
| 3–4 | 28.24 | 23.00‐34.68a | 18.05 | 14.53‐22.41a | |
| 5–6 | 51.80 | 37.23‐72.08a | 26.27 | 18.58‐37.13a | |
| Cardiovascular death within 1 week | 1–2 | 10.48 | 5.85‐18.77a | 8.90 | 4.94‐16.05a |
| 3–4 | 29.76 | 16.15‐54.86a | 20.75 | 10.93‐39.38a | |
| 5–6 | 55.67 | 22.21‐139.53a | 33.42 | 12.75‐87.57a |
Multivariate model adjusted for sex, year of inclusion, and comorbidities noted in baseline characteristics except for the variables involved in the CHADS2 score.
Denotes a P value < 0.001.
Table 3.
Long‐term Mortality Rates per 1000 Person‐Years According to CHADS2 Score in Syncope and Control Population
| All‐Cause Mortality– Syncope | Cardiovascular Mortality– Syncope | All‐Cause Mortality–Control | Cardiovascular Mortality–Control | |
|---|---|---|---|---|
| CHADS2 score 0 | 15.1 | 4.8 | 7.5 | 2.3 |
| CHADS2 score 1–2 | 97.9 | 51.0 | 85.5 | 42.8 |
| CHADS2 score 3–4 | 172.8 | 121.9 | 178.1 | 120.8 |
| CHADS2 score 5–6 | 310.9 | 242.3 | 272.2 | 211.1 |
Finally, when comparing the CHADS2 syncope population with the control population, it was evident that syncope per se was associated with an increased risk of all‐cause mortality and cardiovascular death as shown in Table 4. The relative risk was proportionally higher in patients in the lower CHADS2 scores, but the absolute risk of all‐cause mortality and cardiovascular mortality associated with a CHADS2 score of 0 was low.
Table 4.
Multivariate Cox Model Comparing Syncope With Controls According to CHADS2 Score
| CHADS2 Score | Hazard Ratio Syncope–Control | 95% Confidence Interval | P Value | Events Syncope/Control | |
|---|---|---|---|---|---|
| All‐cause mortality long‐term | 0 | 2.14 | 2.02‐2.27 | <0.001 | 1516/4289 |
| 1–2 | 1.29 | 1.25‐1.33 | <0.001 | 4741/19587 | |
| 3–4 | 1.09 | 1.02‐1.16 | 0.009 | 1368/3683 | |
| 5–6 | 1.25 | 1.01‐1.54 | 0.039 | 136/303 | |
| All‐cause mortality 1 year | 0 | 2.78 | 2.48‐3.13 | <0.001 | 445/837 |
| 1–2 | 1.43 | 1.34‐1.51 | <0.001 | 1477/5181 | |
| 3–4 | 1.18 | 1.06‐1.30 | 0.002 | 524/1317 | |
| 5–6 | 1.27 | 0.93‐1.73 | 0.131 | 62/130 | |
| All‐cause mortality 1 week | 0 | 4.17 | 2.62‐6.62 | <0.001 | 36/40 |
| 1–2 | 2.84 | 2.33‐3.48 | <0.001 | 158/259 | |
| 3–4 | 1.83 | 1.32‐2.54 | <0.001 | 60/95 | |
| 5–6 | 1.26 | 0.51‐3.13 | 0.621 | 7/15 | |
| Cardiovascular death long‐term | 0 | 2.17 | 1.95‐2.41 | <0.001 | 477/1338 |
| 1–2 | 1.29 | 1.24‐1.35 | <0.001 | 2472/9798 | |
| 3–4 | 1.10 | 1.02‐1.19 | 0.010 | 965/2498 | |
| 5–6 | 1.19 | 0.94‐1.51 | 0.152 | 106/235 | |
| Cardiovascular death 1 year | 0 | 2.81 | 2.34‐3.52 | <0.001 | 119/235 |
| 1–2 | 1.41 | 1.29‐1.53 | <0.001 | 748/2542 | |
| 3–4 | 1.20 | 1.06‐1.35 | <0.003 | 388/929 | |
| 5–6 | 1.20 | 0.85‐1.69 | 0.306 | 50/104 | |
| Cardiovascular death 1 week | 0 | 8.01 | 3.22‐19.93 | <0.001 | 13/8 |
| 1–2 | 3.76 | 2.81‐5.03 | <0.001 | 87/104 | |
| 3–4 | 2.34 | 2.60‐3.44 | <0.003 | 49/60 | |
| 5–6 | 1.60 | 0.62‐4.15 | 0.332 | 7/12 |
Multivariate model adjusted for age, sex, year of inclusion, and comorbidities (not in the CHADS2 score).
Secondary End Points
One‐year all‐cause and cardiovascular mortality exhibited the same pattern as long‐term mortality as shown in Table 2. When comparing the 1‐year mortality in syncope with controls, there was a significantly increased risk of both all‐cause and cardiovascular mortality in CHADS2 scores of 0–4, whereas there was no significant difference between syncope and controls in the high CHADS2 scores of 5–6 (Table 3).
One‐week mortality was evaluated in the same fashion, exhibiting increased risk of cardiovascular death with increasing CHADS2 scores. Compared to controls we found a significantly increased risk associated with syncope, where the relative risk was most pronounced in the CHADS2 scores of 0 (HR: 8.01, 95% CI: 3.22‐19.93, P < 0.001), whereas the absolute risk of cardiovascular death within 1 week was low, with 13 and 8 events in 21 197 patients and 118 777 controls. Again, we found no significant association on 1‐week cardiovascular mortality in syncope population with CHADS2 scores of 5–6.
Discharge From the ED and Risk Factor Contribution
A total of 2919 (7.7%) patients of relatively high risk (CHADS2 score >2) were discharged directly from the ED, and of these only 4 patients were (re)admitted for cardiovascular causes within 7 days, whereas 56 (2.1%) patients died of cardiovascular causes within the following week. As seen in the baseline characteristics patients with a CHADS2 score >2 are multicomorbidity patients with a high risk of death.
The contribution of individual risk factors associated with all‐cause mortality is shown in Table 5 as hazard ratios derived from Cox proportional hazard analysis. The risk factor associated with the highest risk is age >75 years (HR: 5.73, 95% CI: 5.55‐5.91), whereas hypertension was associated with the lowest risk of death (HR: 1.80, 95% CI: 1.74‐1.85). In patients with CHADS2 score = 2, the combination of age >75 years and heart failure were the factors depicting the highest risk (HR: 4.93, 95% CI: 4.70‐5.17), whereas the combination of hypertension and diabetes where associated with the lowest risk (HR: 2.08, 95% CI: 1.96‐2.21).
Table 5.
Individual CHADS2 Risk Factor Covariates Contribution in Composing a Total CHADS2 Scores of 1 and 2
| Hazard Ratio | 95% Confidence Interval | P Value | |
|---|---|---|---|
| CHADS2 score = 0 | 1.00 | NA | NA |
| CHADS2 score = 1 | |||
| Hypertension | 1.80 | 1.74‐1.85 | <0.001 |
| Heart failure | 3.84 | 3.68‐4.00 | <0.001 |
| Diabetes | 2.13 | 2.04‐2.22 | <0.001 |
| Age >75 years | 5.73 | 5.55‐5.91 | <0.001 |
| CHADS2 score = 2 | |||
| Diabetes + heart failure | 3.97 | 3.66‐4.31 | <0.001 |
| Diabetes + hypertension | 2.08 | 1.96‐2.21 | <0.001 |
| Diabetes + age >75 years | 3.52 | 3.33‐3.72 | <0.001 |
| Heart failure + hypertension | 3.41 | 3.24‐3.59 | <0.001 |
| Heart failure + age >75 years | 4.93 | 4.70‐5.17 | <0.001 |
| Hypertension + age >75 years | 2.72 | 2.63‐2.82 | <0.001 |
| Previous stroke | 2.54 | 2.43‐2.64 | <0.001 |
Abbreviations: NA, not applicable.
Results from Cox proportional hazard analysis using CHADS2 score of 0 as a reference.
Discussion
In this study we provide evidence that the CHADS2 score is highly prognostic as a risk stratification tool for syncope patients measured on all‐cause and cardiovascular mortality. Importantly, it also provides reliable data for the use of the CHADS2 score when estimating the risk of patients with syncope and a CHADS2 score of 0. Similarly, the data suggest that a CHADS2 score as low as 1 is associated with a relatively high mortality rate, which could be useful in designing new risk score models, as none of the previous has been widely implemented.
Furthermore, syncope as a symptom is associated with adverse prognosis on short‐ and long‐term mortality when compared to a control population in the same CHADS2 scores groups. This indicates that syncope is a symptom that predicts cardiovascular mortality despite a low CHADS2 score.
CHADS2 Score of 0
The event rates for mortality were generally low, indicating low absolute risk of death when CHADS2 score is 0, even after an ED visit for syncope. However, when compared to the control population, it addresses that risk factors other than those used in the CHADS2 score need to be taken into account when designing risk stratification models.
Recently we found that syncope, even in healthy individuals, has a higher risk of cardiovascular events, hospitalizations, and death compared to the background population.28 This study supports these findings and suggests that a CHADS2 score of 0 cannot be used as a safe rule of discharge from the ED of these patients.
CHADS2 Score Above 0
CHADS2 scores above 0 and as low as 1 carry higher risks of all‐cause death as well as cardiovascular death, which was markedly increased compared to a score of 0. We find that all‐cause mortality event rates of patients with syncope when CHADS2 score is above 0 are comparable to event rates in populations with atrial fibrillation.1, 2
The event rates and hazard ratios of cardiovascular and all‐cause mortality are, however, probably attributable to many diseases when evaluating a CHADS2 score on these outcomes. Therefore, we compared our syncope population with controls of the same CHADS2 scores. This consistently showed an increased risk associated with syncope for all‐cause mortality and cardiovascular mortality on a long‐ and short‐term basis.
Patients in the highest CHADS2 scores group did not have increased risk compared to the control population after appropriate adjustments. This is attributable to an overall high mortality rate in this subset, and that very few patients were allocated to this group. Furthermore, it indicates that syncope in this subgroup does not alter the clinical course of these very sick patients.
Individual Risk Factor Contribution to the Risk of All‐Cause Mortality
Not surprisingly, we found age above 75 years and heart failure to be the factors most significantly associated with death after syncope, whereas hypertension accounted for a slight increase in risk. The linearity of the increased risk among the individual factors makes the CHADS2 risk score applicable for 1 point per variable as in the original purpose of the risk score calculation.
The CHADS2 risk score system was developed to estimate the risk of stroke in populations with atrial fibrillation, therefore using the risk factors most strongly associated with that end point. It is not necessarily the same risk factors that are applicable to estimating the best risk of cardiovascular events in a syncope population, but nevertheless, we see the risk factors used in CHADS2 as applicable to an overall cardiovascular risk profile. Using the score in a population with syncope is therefore an estimation of overall cardiovascular risk, and we find that the all‐cause mortality and cardiovascular mortality rates are likewise parallel. When proposing new risk score stratifications schemes for syncope, these considerations should be acknowledged.
Risk Stratification Models
Some risk score stratifications have been proposed for the initial evaluation of short‐term outcome but have failed later validations.29, 30, 31, 32, 33 Structural heart disease is a major risk factor for sudden cardiac death and overall mortality in patients with syncope13, 29, 34, 35 and is a key point in the current risk stratification as defined by the European Society of Cardiology guidelines in 2009,36 along with pathological electrocardiogram, severe anemia, and electrolyte disturbances.
This study is the first to show a significant association between CHADS2 score and short‐ and long‐term all‐cause mortality and cardiovascular mortality in a very large cohort of unselected patients with syncope discharged from the ED. It is of particular importance that in our cohort, many patients with a CHADS2 score >2 actually are not being admitted from the ED for further evaluation, and that approximately 2% of these die within the week without any cardiovascular rehospitalization is worrisome.
Limitations
This is a retrospective, register‐based, observational cohort study, which is a major limitation. The registers unfortunately did not include important clinical variables such as electrocardiogram and echocardiogram, weight, and laboratory values, which would all be appropriate to consider in this study. In observational studies, the use of multivariate adjustments limits confounding, but it does not eliminate unmeasured confound that could affect the results. Although not all ICD‐10 diagnoses are validated, and some are more useful than others, the approach used was the same in the syncope cases and the controls, thus limiting the effect of this particular issue.
Conclusion
The CHADS2 score significantly predicts short‐ and long‐term all‐cause and cardiovascular mortality in patients discharged after syncope. Syncope was associated with increased risk of death compared to the background population. A CHADS2 score of 0 was associated with a very low absolute mortality but a higher relative risk than that of the background population, suggesting that other significant risk factors are of importance in a syncope population.
The following ICD‐10 codes were used to identify comorbidity: cerebral vascular disease (I60‐I69), ischemic heart disease (I20‐I25), previous myocardial infarction (I21‐I22), cardiac conduction disorders (I44‐45), atrial fibrillation (I48), heart failure (I50,I42, J81), acute or chronic renal failure (N17‐19, I12‐13, R34), diabetes with or without complications (E10‐14), chronic obstructive pulmonary disease (J42‐44), and malignancies and metastatic cancer (C00‐97). Chronic obstructive pulmonary disorder was redefined as a claimed prescription for a bronchial dilating medication for inhalation (ATC code R04) and/or an admission for a chronic obstructive pulmonary disorder (J42‐44).
The following ATC codes were used to identify pharmacotherapy: statins (C10A), β‐blockers (C07), angiotensin‐converting enzyme inhibitors (C09), loop diuretics (C03C), spironolactone (C03D), thiazides (C03A), calcium channel blockers (C08), digoxin (C01AA05), glucose lowering medication (A10), acetylsalicylic acid (B01AA0), vitamin K antagonists (B01AA0), antiepileptic drugs (N03), antidepressants (N06A), sedatives and anxiolytics (N05B, N05C), and bronchodilators (R03).
Dr. Martin Huth Ruwald has received unrestricted grants from the Danish Heart Association (12‐04‐R90‐A3806‐22701), The Lundbeck Foundation (R108‐A104415), Helsefonden (2012B018), Knud Hoejgaards Fund, Arvid Nilssons Fund, and Snedkermester Sophus Jakobsens Fund. The authors have no other funding, financial relationships, or conflicts of interest to disclose.
References
- 1. Gage BF, Waterman AD, Shannon W, et al. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA. 2001;285:2864–2870. [DOI] [PubMed] [Google Scholar]
- 2. Olesen JB, Lip GY, Hansen ML, et al. Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study. BMJ. 2011;342:d124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Crandall MA, Horne BD, Day JD, et al. Atrial fibrillation significantly increases total mortality and stroke risk beyond that conveyed by the CHADS2 risk factors. Pacing Clin Electrophysiol. 2009;32:981–986. [DOI] [PubMed] [Google Scholar]
- 4. Kim YD, Cha MJ, Kim J, et al. Ischaemic cardiovascular mortality in patients with non‐valvular atrial fibrillation according to CHADS score. Thromb Haemost. 2011;105:712–720. [DOI] [PubMed] [Google Scholar]
- 5. Nakagawa K, Hirai T, Takashima S, et al. Chronic kidney disease and CHADS(2) score independently predict cardiovascular events and mortality in patients with nonvalvular atrial fibrillation. Am J Cardiol. 2011;107:912–916. [DOI] [PubMed] [Google Scholar]
- 6. Khumri TM, Idupulapati M, Rader VJ, et al. Clinical and echocardiographic markers of mortality risk in patients with atrial fibrillation. Am J Cardiol. 2007;99:1733–1736. [DOI] [PubMed] [Google Scholar]
- 7. Ruiz‐Nodar JM, Marin F, Manzano‐Fernandez S, et al. An evaluation of the CHADS stroke risk score in patients with atrial fibrillation who undergo percutaneous coronary revascularization. Chest. 2011;139:1402–1409. [DOI] [PubMed] [Google Scholar]
- 8. Chao TF, Cheng CC, Lin WS, et al. Associations among the CHADS(2) score, atrial substrate properties, and outcome of catheter ablation in patients with paroxysmal atrial fibrillation. Heart Rhythm. 2011;8:1155–1159. [DOI] [PubMed] [Google Scholar]
- 9. Chao TF, Ambrose K, Tsao HM, et al. Relationship between the CHADS(2) score and risk of very late recurrences after catheter ablation of paroxysmal atrial fibrillation. Heart Rhythm. 2012;9:1185–1191. [DOI] [PubMed] [Google Scholar]
- 10. Pouleur AC, Barkoudah E, Uno H, et al. Pathogenesis of sudden unexpected death in a clinical trial of patients with myocardial infarction and left ventricular dysfunction, heart failure, or both. Circulation. 2010;122:597–602. [DOI] [PubMed] [Google Scholar]
- 11. Huikuri HV, Castellanos A, Myerburg RJ. Sudden death due to cardiac arrhythmias. N Engl J Med. 2001;345:1473–1482. [DOI] [PubMed] [Google Scholar]
- 12. Yap YG, Duong T, Bland M, et al. Temporal trends on the risk of arrhythmic vs. non‐arrhythmic deaths in high‐risk patients after myocardial infarction: a combined analysis from multicentre trials. Eur Heart J. 2005;26:1385–1393. [DOI] [PubMed] [Google Scholar]
- 13. Soteriades ES, Evans JC, Larson MG, et al. Incidence and prognosis of syncope. N Engl J Med. 2002;347:878–885. [DOI] [PubMed] [Google Scholar]
- 14. Ruwald MH, Hansen ML, Lamberts M, et al. The relation between age, sex, comorbidity, and pharmacotherapy and the risk of syncope: a Danish nationwide study. Europace. 2012;14:1506–1514. [DOI] [PubMed] [Google Scholar]
- 15. Gaist D, Sorensen HT, Hallas J. The Danish prescription registries. Dan Med Bull. 1997;44:445–448. [PubMed] [Google Scholar]
- 16. Andersen TF, Madsen M, Jorgensen J, et al. The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan Med Bull. 1999;46:263–268. [PubMed] [Google Scholar]
- 17. Ruwald MH, Hansen ML, Lamberts M, et al. Accuracy of the ICD‐10 discharge diagnosis for syncope [published online ahead of print November 4, 2012]. Europace. doi: 10.1093/europace/eus359. [DOI] [PubMed] [Google Scholar]
- 18. Krarup LH, Boysen G, Janjua H, et al. Validity of stroke diagnoses in a National Register of Patients. Neuroepidemiology. 2007;28:150–154. [DOI] [PubMed] [Google Scholar]
- 19. Madsen M, Davidsen M, Rasmussen S, et al. The validity of the diagnosis of acute myocardial infarction in routine statistics: a comparison of mortality and hospital discharge data with the Danish MONICA registry. J Clin Epidemiol. 2003;56:124–130. [DOI] [PubMed] [Google Scholar]
- 20. Thomsen RW, Lange P, Hellquist B, et al. Validity and underrecording of diagnosis of COPD in the Danish National Patient Registry. Respir Med. 2011;105:1063–1068. [DOI] [PubMed] [Google Scholar]
- 21. Kumler T, Gislason GH, Kirk V, et al. Accuracy of a heart failure diagnosis in administrative registers. Eur J Heart Fail. 2008;10:658–660. [DOI] [PubMed] [Google Scholar]
- 22. Olesen JB, Lip GY, Kamper AL, et al. Stroke and bleeding in atrial fibrillation with chronic kidney disease. N Engl J Med. 2012;367:625–635. [DOI] [PubMed] [Google Scholar]
- 23. Sorensen R, Hansen ML, Abildstrom SZ, et al. Risk of bleeding in patients with acute myocardial infarction treated with different combinations of aspirin, clopidogrel, and vitamin K antagonists in Denmark: a retrospective analysis of nationwide registry data. Lancet. 2009;374:1967–1974. [DOI] [PubMed] [Google Scholar]
- 24. Nielsen SF, Nordestgaard BG, Bojesen SE. Statin use and reduced cancer‐related mortality. N Engl J Med. 2012;367:1792–1802. [DOI] [PubMed] [Google Scholar]
- 25. Lidegaard O, Lokkegaard E, Jensen A, et al. Thrombotic stroke and myocardial infarction with hormonal contraception. N Engl J Med. 2012;366:2257–2266. [DOI] [PubMed] [Google Scholar]
- 26. Andersson C, Norgaard ML, Hansen PR, et al. Heart failure severity, as determined by loop diuretic dosages, predicts the risk of developing diabetes after myocardial infarction: a nationwide cohort study. Eur J Heart Fail. 2010;12:1333–1338. [DOI] [PubMed] [Google Scholar]
- 27. Schramm TK, Gislason GH, Kober L, et al. Diabetes patients requiring glucose‐lowering therapy and nondiabetics with a prior myocardial infarction carry the same cardiovascular risk: a population study of 3.3 million people. Circulation. 2008;117:1945–1954. [DOI] [PubMed] [Google Scholar]
- 28. Ruwald MH, Hansen ML, Lamberts M, et al. Prognosis among healthy individuals discharged with a primary diagnosis of syncope. J Am Coll Cardiol. 2013;61:325–332. [DOI] [PubMed] [Google Scholar]
- 29. Quinn J, McDermott D, Stiell I, et al. Prospective validation of the San Francisco Syncope Rule to predict patients with serious outcomes. Ann Emerg Med. 2006;47:448–454. [DOI] [PubMed] [Google Scholar]
- 30. Colivicchi F, Ammirati F, Melina D, et al. Development and prospective validation of a risk stratification system for patients with syncope in the emergency department: the OESIL risk score. Eur Heart J. 2003;24:811–819. [DOI] [PubMed] [Google Scholar]
- 31. Reed MJ, Newby DE, Coull AJ, et al. The ROSE (Risk Stratification of Syncope in the Emergency Department) study. J Am Coll Cardiol. 2010;55:713–721. [DOI] [PubMed] [Google Scholar]
- 32. Reed MJ, Henderson SS, Newby DE, et al. One‐year prognosis after syncope and the failure of the ROSE decision instrument to predict one‐year adverse events. Ann Emerg Med. 2011;58:250–256. [DOI] [PubMed] [Google Scholar]
- 33. Del Rosso A, Ungar A, Maggi R, et al. Clinical predictors of cardiac syncope at initial evaluation in patients referred urgently to a general hospital: the EGSYS score. Heart. 2008;94:1620–1626. [DOI] [PubMed] [Google Scholar]
- 34. Kapoor WN, Hanusa BH. Is syncope a risk factor for poor outcomes? Comparison of patients with and without syncope. Am J Med. 1996;100:646–655. [DOI] [PubMed] [Google Scholar]
- 35. Middlekauff HR, Stevenson WG, Stevenson LW, et al. Syncope in advanced heart failure: high risk of sudden death regardless of origin of syncope. J Am Coll Cardiol. 1993;21:110–116. [DOI] [PubMed] [Google Scholar]
- 36. Moya A, Sutton R, Ammirati F, et al. Guidelines for the diagnosis and management of syncope (version 2009). Eur Heart J. 2009;30:2631–2671. [DOI] [PMC free article] [PubMed] [Google Scholar]