Author + information
- Received April 26, 1996
- Revision received November 8, 1996
- Accepted December 16, 1996
- Published online April 1, 1997.
- David J Driscoll, MD, FACCA,*,
- Steven J Jacobsen, MD, PhDA,
- Co-burn J Porter, MD, FACCA and
- Peter C Wollan, PhDA
- ↵*Dr. David J. Driscoll, Pediatric Cardiology, Mayo Clinic and Mayo Foundation, 200 First Street SW, Rochester, Minnesota 55905.
Objectives. The objectives of this study were to 1) define the incidence of syncope coming to medical attention among children and adolescents, 2) determine the outcome of syncope in these patients, and 3) determine changes over time in the evaluation and charges for evaluating this problem.
Background. Syncope occurs commonly in children and adolescents. However, the mid- and long-term outcome of children and adolescents who experience syncope is unknown.
Methods. Utilizing the Rochester Epidemiology Project, we determined the incidence, outcome and charges for medical evaluation for patients seeking medical attention for syncope during an early 5-year period (1950 to 1954) and a more recent 5-year period (1987 to 1991).
Results. The incidence of syncope coming to medical attention was 71.9 and 125.8/100,000 population for the early and more recent cohort, respectively. The incidence was higher for female than for male patients. The incidence peaked in 15- to 19-year old patients. Acute illness and noxious stimuli were associated with 24% and 23% of the episodes, respectively. Although long-term survival was not different from that of the general population, one child died suddenly, and another had hereditary prolonged QT interval syndrome. These were two of only six patients who had exertional syncope. Total charges for evaluation of syncope were similar in the two time periods. However, charges for testing procedures were greater for the more recent cohort.
Conclusions. In general, syncope in children and adolescents is a benign event. Syncope occurring during exercise may identify patients with a potentially fatal condition. Detailed evaluation should be considered for patients who have syncope during exercise or who have a family history of syncope, sudden death, myocardial disease or arrhythmias. It may be prudent to obtain an electrocardiogram for all patients who seek medical attention for syncope.
(J Am Coll Cardiol 1997;29:1039–45)
© 1997 by the American College of Cardiology
Syncope is relatively common in adolescence. Although syncope can be a benign event, in some circumstances it can herald a potentially lethal problem (). In addition, syncope of benign origin can lead to injury or death if loss of consciousness occurs during a situation in which the patient is in particular jeopardy. Thus, patients and physicians are anxious to determine the cause of syncope, identify and treat potentially lethal underlying causes of syncope and prevent future syncopal episodes.
There is no population-based study of syncope in children and adolescents on which guidelines for evaluation and management can be based. Most of the existing data come from referral centers where the mix of patients may not reflect the population at large. The purpose of this study was to 1) define the incidence of syncope in children and adolescents that comes to medical attention, 2) determine the outcome of syncope in these patients, and 3) determine changes over time in the evaluation of this problem and in the charges for evaluating it.
1.1 Study setting.
The population of Rochester, Minnesota is served by a largely unified medical care system that has accumulated comprehensive clinical records over an extended period of time. Rochester (1990 population 70,745) lies 90 miles southeast of Minneapolis/St. Paul. In 1990, the population was 96% white; 28% were >45 years old and 11% 65 years or older. The population is largely middle class; ∼82% of adults have graduated from high school. With the exception of a higher proportion of the working population employed in the health care industry, the characteristics of the population of Rochester are similar to those of U.S. whites.
1.2 Rochester Epidemiology Project.
Epidemiologic research in Rochester is possible because the city is relatively isolated from other urban centers and nearly all medical care is delivered to local residents by a small number of providers (). Most of this medical care is provided by the Mayo Clinic, a major referral center with >1,000 full-time physicians representing every medical and surgical specialty and subspecialty. Mayo Clinic, along with the Olmsted Medical Group and its affiliated Olmsted Community Hospital, provide comprehensive care for the region in every clinical discipline including primary care.
The epidemiologic potential of this situation is enhanced by the fact that each provider uses a unit (or dossier) medical record system whereby all data collected on an individual are assembled in one place. The Mayo Clinic unit record, for example, contains the details of every inpatient hospital stay at its two large affiliated hospitals, every outpatient visit, every physician visit to nursing homes or private homes, as well as every laboratory result, pathology report (including autopsy data) and piece of correspondence concerning each patient. The unit records of each provider in the county have been maintained and are available for use. These medical records are easily retrievable because the Mayo Clinic has maintained, since the early 1900s, extensive indexes based on clinical and histologic diagnoses and surgical procedures. The Rochester Epidemiology Project has developed a similar index for the records of other providers of medical care to local residents. The result is the linkage of medical records from essentially all sources of medical care available to and utilized by the Rochester population.
1.3 Identification of cases.
The medical records of potential cases were identified through the diagnostic indexes of the Rochester Epidemiology Project. Potential cases were drawn from persons falling under the diagnostic rubrics of syncope, fainting, vasovagal attack, spells, consciousness disorders, vasomotor instability, vasomotor attack, cardiogenic syncope and heat syncope. These indexes do not capture the reason for visits. Hence, patients who presented with syncope but were found to have a more specific diagnosis may not have been identified. Patients evaluated by all physician providers in the community were included in this study. Once a potential case was identified, the community medical records were reviewed by a trained nurse abstracter to verify that 1) the syncopal episode had occurred and was the initial syncopal episode for which medical attention was sought, 2) the subject was a resident of Rochester, Minnesota at the time of the episode, 3) the subject was between 1 and 22 years of age at the time of the episode, and 4) the syncopal episode occurred during calendar years 1950 through 1954 (“1950s cohort”) or 1987 through 1991 (“1990s cohort”).
For the purposes of this study, syncopewas defined as a transient (<2 h) loss of consciousness (see for definitions of specific types of syncope). It included witnessed as well as unwitnessed events. It included instances in which cardiopulmonary resuscitation was performed by a family member but excluded instances in which cardiopulmonary resuscitation was performed by medical or paramedical personnel. Also excluded were episodes of unconsciousness resulting from an apparent seizure disorder. Although incontinence rarely can be associated with syncope, we excluded cases in which the patient was incontinent of urine or stool in an effort to exclude all primary seizure disorders. Episodes preceded within 5 min by major trauma also were excluded.
Residence was verified by using information from birth certificates, city and county directories or earlier medical records. Each subject’s residency status was determined on the index date and 1 year before the index date to identify anyone from the cohort who may have moved to Rochester to specifically facilitate the diagnosis and treatment of conditions associated with syncope.
Once a subject was identified and Rochester residence confirmed, the community inpatient and outpatient medical records were reviewed to determine the circumstances surrounding the syncopal episode, the results of the initial diagnostic evaluation, as well as basic demographic information. In addition, all subjects were passively followed up through their medical records to determine their last known vital status, occurrence of syncope and subsequent diagnostic studies and outcome that might be related to syncope.
All persons not known to have died by June 1994 were contacted by mail questionnaire. Questionnaire elements elicited information about vital status, recurrence of syncopal episodes, date of last episode, injuries associated with syncopal episodes and additional evaluation of syncope performed outside of Olmsted County. Medical questionnaires were sent to 196 patients known not to have died. Completed questionnaires were received from 111 patients. Telephone contact was made and questionnaire information was obtained for an additional 49 patients. Current vital status was determined for 183 of the 196 patients; 13 patients were lost to follow-up.
Once collected, the abstracted data were entered into a computer and verified by trained data entry personnel. The data then were edited with a variety of on-line range and consistency checks. Quality control and adherence to the protocol approved by the Mayo Clinic Institutional Review Board were ensured by the close supervision of the principal investigator and coinvestigators.
1.4 Data analysis.
Incidence rates were calculated by assuming that the population of Rochester was at risk. Denominators for incidence rates were based on decennial census data, with linear interpolation for intracensal estimates (). Poisson regression was used to test for differences in incidence rates due to gender, period and age group. Differences in the distribution of associated signs and symptoms, related activities and physician diagnosis were tested with the Wilcoxon rank-sum test for ordinal variables and the chi-square statistic for categoric variables. When expected frequencies were <5 for any cell in contingency tables, the Fisher exact test was used. A standardized mortality ratio was estimated by comparing the observed number of deaths among cohort members to an expected number based on State of Minnesota decennial life tables ().
Total charges between periods were compared with a two-sample Wilcoxon rank-sum test. To assess the relative charges for medical evaluation, typical commercial non-Medicare insurance fees for 1994 were used. Three sets of charges were examined: 1) the charge for the initial medical encounter, 2) the charges for testing, and 3) the charges for subsequent medical follow-up or consultation. All analyses were carried out with SAS Institutes statistical software.
A search of the Rochester Epidemiology Diagnostic Index yielded 447 potential cases with the diagnostic rubrics outlined earlier. Of these, 194 (43%) fulfilled all inclusion criteria. Reasons for exclusion included non-Rochester resident (n = 101), outside of age criteria (n = 11), outside of examination date (n = 7), no documented true syncope (n = 116), associated incontinence (n = 2) and prior medical attention for a syncopal episode (n = 16).
The 1950s cohort comprised 43 patients, 40 white and 3 of unknown ethnicity; 31 were female and 12 were male. The 1990s cohort comprised 151 patients, 98 female and 53 male; 131 were white, 3 were southeast Asian and 17 were of unknown ethnicity. The median age of the female patients was 18 and 16 years for the 1950s and the 1990s cohort, respectively (p = 0.65). The median age of the male patients was 10.5 and 13 years for the 1950s and the 1990s cohort, respectively (p = 0.51).
Syncopal episodes occurring before the index episode for which the patient sought medical attention were noted for 25 (58%) of the 43 patients in the 1950s cohort and for 47 (31%) of the 151 patients in the 1990s cohort.
The incidence of syncopal episodes coming to medical attention per 100,000 population was 71.9 and 125.8 for the 1950s and the 1990s group, respectively. The age-adjusted incidence rates were higher for female (89.8 and 166.3) than for male (47.8 and 92.9) patients for the 1950s and the 1990s cohort, respectively. For both male and female patients, the rate peaked in the 15- to 19-year old group, but the peak for female patients was dramatically greater than that for male patients (Fig. 1). The gender, period and age group main effect and gender by age group interaction effect are all highly significant (p = 0.04, 0.003, 0.0001 and 0.002, respectively). The period by age group interaction is marginally significant (p = 0.073).
2.2 Signs and symptoms.
Signs and symptoms associated with the index syncopal episode are displayed in Table 1. Injury resulting from the syncopal episode, or light-headedness, nausea/vomiting, dizziness or visual symptoms associated with the episode, occurred in 18%, 16%, 15%, 14% and 12% of patients, respectively. There were essentially no differences in the proportion of signs and symptoms between the two cohorts, with the exception of light-headedness, which occurred more frequently in the 1990s cohort (20% vs. 5%, p = 0.02). Activities or events associated with the episode or patient location during the episode are listed in Table 2. The presence of an acute illness or the use of prescription drugs was associated with 24% and 20%, respectively, of episodes. Noxious stimuli or psychologic/emotional stimuli were associated with 23% and 12%, respectively, of episodes. Syncope occurred during exercise in 6 (3%) of 194 patients.
Most of the index episodes were evaluated initially in the emergency room or the outpatient clinic. However, in the 1950s cohort, 23% were evaluated by house call. Visits for further medical evaluation after the initial evaluation for the index syncopal episode were performed for 56% of the 1950s cohort and for 47% of the 1990s cohort.
2.3 Diagnostic categories.
The majority of patients were thought by the attending physician to have had either a simple faint or vasodepressor or vasovagal syncope (Table 3).
A variety of diagnostic tests were performed (Table 4). The profile of tests performed differed for the two cohorts. For example only 1 (2%) of 43 of patients from the 1950s cohort but 35 (23%) of 151 patients from the 1990s cohort had electrocardiograms (ECGs) performed. Glucose and electrolyte measurements were performed much more frequently in the later than in the earlier cohort. Fewer diagnostic studies were performed for the 1950s cohort (71 tests in 43 patients) than for the 1990s cohort (479 tests in 151 patients).
Thirty-five patients had ECGs. Among these, the corrected QT intervals ranged from 0.35 to 0.44. Although no patients had a prolonged QT interval, one adolescent who died suddenly (described later) had not had an ECG performed. Another adolescent did not have an ECG performed when evaluated for his index episode of syncope, but after a subsequent episode he was found to have prolonged QT interval syndrome.
Three patients died; all were from the 1950s cohort. In two of these patients, the death had no apparent relation to previous syncopal episodes. One of the two patients committed suicide at age 41 years, and the second died of liver failure secondary to viral hepatitis at age 27 years. Mortality was no different from that of an age-matched population. In a comparable age- and gender-matched population, the expected number of deaths was 3.0.
The third patient died suddenly and unexpectedly at age 13 years. This patient stood up to recite in class, complained of feeling faint, collapsed and died. Four months before his death, he had had syncope (the index episode) while running. It was reported that he was unconscious for 20 min. Seven months before his death he had also had syncope while running. No ECG was ever performed.
Syncopal episodes occurring after the index episode were noted for 3 (7%) of 43 patients in the 1950s cohort and for 14 (9%) of 151 patients in the 1990s cohort.
2.4.3 Associated medical conditions.
As part of the follow-up questionnaire the presence of specific medical conditions was ascertained (Table 5). Migraine headaches were reported in 28 (18%) of 152 of the patients. Subsequent development of diabetes mellitus was more common in the 1950s than in the 1990s cohort. This difference probably reflects the longer follow-up period for the 1950s cohort.
2.5 Charges for medical evaluation.
The total charge for the initial medical encounter, testing and subsequent medical follow-up or consultation was not significantly different for the 1950s cohort (mean $577.77, median $326) and 1990s cohort (mean $641.48, median $303.00). However, the charge per patient for tests performed was significantly (p < 0.05) higher for patients evaluated in the 1990s ($289/patient) than for those evaluated in the 1950s ($77/patient). The charges for follow-up evaluation and consultation were higher for the 1950s cohort ($557/patient) than for the 1990s cohort ($287/patient). This difference was due to a higher rate of emergency room and repeat outpatient evaluation of patients in the 1950s than in the 1990s cohort.
3.1 Syncope in adults.
The epidemiologic features of syncope were reported for 2,336 adult men and 2,873 adult women as part of the Framingham study (). These subjects ranged in age from 30 to 62 years. During the course of 13 biennial examinations, 71 of the men and 101 of the women reported at least one syncopal episode during their lifetime. Isolated syncope (i.e., transient loss of consciousness in the absence of prior or concurrent neurologic, coronary or other cardiovascular disease stigmata) occurred in 79% of the men and 88% of the women with syncope. During 26 years of follow-up, isolated syncope was unassociated with excess stroke, myocardial infarction or mortality from other causes. Kapoor et al. ([6, 7]) prospectively studied 204 adult patients with syncope. A cardiovascular cause of the syncope was identified in 53 patients and a noncardiovascular cause in 54; in 97 patients the cause remained unknown. After 12 months of follow-up the overall mortality rate was 14 ± 2.5%. The mortality rate was 30 ± 6.7% for patients with a cardiovascular cause of syncope and 6.7 ± 2.8% for patients with syncope of unknown cause.
3.2 Syncope in children.
There have been few studies of syncope in children and, to our knowledge, no population-based studies. Pratt and Fleisher () reported the evaluation and outcome of 77 children and adolescents who were evaluated in an emergency room for syncope. These patients ranged in age from 22 months to 21 years (mean 12.7 years). On review of the records, 37 (48%) of the 77 patients were deemed not to have had syncope. This is similar to our finding that 36% of patients coded as having syncope did not have syncope. In the study of Pratt and Fleisher, 50% of the patients had vasovagal syncope, 20% had orthostatic syncope, 7.5% had atypical seizures, 5% had migraine headaches and 5% had minor head trauma. The occurrence of migraine headaches in their study is interesting in light of our finding that 18% of our patients reported migraine headaches. As in the present study only a small number of patients had an abnormal blood count or serum glucose level. Pratt and Fleisher concluded that the routine evaluation of the child with syncope should include a history and physical examination with measurement of heart rate and blood pressure in supine and standing positions, an ECG and serum glucose and hematocrit measurements.
Gordon et al. () reported on 73 children and adolescents evaluated at the Cleveland Clinic for syncope between 1981 and 1986. The patients ranged in age from 2.5 to 20 years (mean 13). Specific problems were identified in seven patients. These problems included myocardial disease, atrioventricular node reentrant tachycardia, muscular dystrophy, sick sinus syndrome and febrile seizures. These 73 patients had a total of 443 diagnostic test and consultations (6/patient). A relatively large number (n = 29) of the patients were admitted to the hospital. The charges for the patients admitted to the hospital were $4,882/patient, whereas the average charge for evaluation for the entire group of patients was $2,973/patient. In the current study the average charge/patient was $627.36 ($577.77/patient in the 1950s cohort and $641.48/patient in the 1990s cohort). The discrepancy of diagnoses encountered and in charges between our study and that of Gordon et al. is most likely due to the fact that their patients reflected a referred group of patients whereas ours represented a local population-based group of patients.
3.3 Study limitations.
We found an incidence of syncope that ranged from 71.9 to 125.8/100,000 population. This represents a lower boundary estimate of syncope in the general population of children and adolescents because only syncopal episodes that came to medical attention were included in the study. From a practitioner’s standpoint, of course, these are the only episodes with which one must deal. Interpretation of the observed changes in incidence is also problematic. It is not possible to determine whether these changes were due to a real change in incidence, changes in care seeking, changes in care seeking behavior or changes in coding practices.
It is possible that we underestimated the occurrence of significant and potentially lethal underlying or associated medical conditions. If a patient presented for evaluation of syncope but was found to have a specific disease entity (e.g., cardiomyopathy), the dismissal code may have been specific for the disease entity and not reflective of the presenting symptom of syncope. This patient may not have been identified by our study method. Depending on the directions and magnitude of these potential biases, the overall outcome for syncope may be better or worse than that reported here.
Although clinicians frequently assign a diagnostic label to patients with syncope, it is difficult to distinguish, for example, among several types of syncope such as vasovagal syncope, simple faint and orthostatic syncope. One may find variations from study to study or for different time periods within a study in the distributions of the descriptive categories for syncope. Consequently, any shifts in the distributions of the descriptive categories in this study should be interpreted with caution.
3.4 Cost of evaluation.
To our knowledge, our study is the only population-based study and is the first to compare the charges for evaluation of syncope in children and adolescents in two different eras. Not surprisingly, the charge for medical tests/patient was significantly higher in the more recent cohort. It is difficult to know if this resulted from the availability of more sophisticated testing procedures, the presence of increased medical-legal pressures or the desire to make a more specific diagnosis. It was surprising to find that the total charge for evaluation was similar for the early and later cohorts. The increased charges for tests performed for the more recent cohort were offset by the greater number of evaluations performed on inpatients in the 1950s cohort.
Appropriate medical and cost-effective evaluation of a condition such as syncope is inherently related to the outcome of the condition. A condition associated with low or no risk of death and minimal morbidity may not require extensive and expensive medical evaluation. If a subset of patients has a different outcome from the rest of the cohort, it is useful to identify that subset so that appropriate medical evaluation can be targeted toward it rather than toward the entire cohort. In the two studies cited earlier there was no mortality associated with syncope in children and adolescents. In the present study with long follow-up, the risk of death was similar to that expected in the general population. However, there was one instance of sudden unexpected death in a young child that potentially could have been prevented. In addition, a second patient ultimately was identified to have a potentially lethal condition (long QT interval syndrome). Thus, although syncope in children and adolescents is a benign event for the vast majority of patients, it can be associated with sudden death in a very small number of patients. It is important to develop cost-effective algorithms to identify these rare persons.
3.5 Exercise and syncope.
It is known that sudden death in children and adolescents can be associated with, among other conditions, hypertrophic cardiomyopathy, anomalous origin of the left coronary artery from the right sinus of Valsalva, myocarditis, long QT interval syndrome, cystic medial necrosis and Wolff-Parkinson-White syndrome. There also is some suggestion that syncope that occurs during exertion may portend a worse outcome than syncope that is unassociated with exercise. Indeed, in the present study only six patients had syncope in association with exercise. Of these, one patient ultimately died and a second had long QT interval syndrome.
3.6 Conclusions and recommendations.
On the basis of the current and previous studies we propose the following strategy. It is very important to obtain a detailed personal and family history for all patients who seek medical attention for syncope. One should seek information regarding a family history of sudden unexpected death, as well as the patient’s personal and family history of myocardial disease, syncope, arrhythmias and other conditions associated with sudden death. All patients should have a physical examination. Patients with exertional syncope or a positive personal or family history should have a more detailed evaluation. The details of that evaluation should be dictated by the specific case. Some might argue that all children and adolescents who faint should have an ECG although this recommendation is not supported by the results of this study. Additional testing will have a low yield and not be cost-effective for patients with negative findings in the personal and family history, a normal physical examination and nonexertional syncope. This approach should be evaluated in a clinical trial.
We thank Margaret E. (Peg) Farrell, RN for diligent review of the medical records and Christine Boos, BS for assistance in analytic programming.
☆ This project was supported in part by Grant AR30582 from the U.S. Public Health Service, National Institutes of Health, Bethesda, Maryland and by the Mayo Foundation for Education and Research, Rochester, Minnesota.
- Received April 26, 1996.
- Revision received November 8, 1996.
- Accepted December 16, 1996.
- The American College of Cardiology
- Driscoll DJ,
- Edwards W
- ↵Bergstralh EJ, Offord KP, Chu C-P, Beard C, O’Fallon WM, Melton LJ. Calculating incidence, prevalence, and mortality rates for Olmsted County, Minnesota: an update. Technical Report Series No. 49, Section of Biostatistics, Mayo Clinic, Rochester, April, 1992.
- ↵Therneau T, Sicks J, Bergstralh E, Offord J. Expected survival based on hazard rates. Technical Report Series No. 52, Section of Biostatistics, Mayo Clinic, Rochester, 1994.
- Savage D,
- Corwin L,
- McGee D,
- Kannel W,
- Wolf P
- Gordon T,
- Moodie D,
- Passalacqua M,
- et al.