Results
Among the 11,878 individuals who participated in the survey both by undergoing an ECG and answering to the questionnaire, we calculated a mean age of 21±6 years-old, ranging from 6 to 40 years-old, of whom 30.6% were children (<18 years-old), and 59.6% (n=7078) were female. We found a cumulative incidence of TLOC of 26.5% (n=3153), 18.9% (n=500) in children and 28.8% (n=2653) in young-adults. The majority of participants with TLOC were female (76.2%, n=2403) and almost half (43,9%, n=1379) were engaged in regular physical activity.
Among individuals with a history of TLOC, 2095 (66.4%) had RS and 259 (8.2%) reported episodes compatible with OH. Several other causes for the TLOC episodes were identified: 122 (3.9%) other diseases (ex: anemia, diabetes), 120 (3.8%) drugs/alcohol, 53 (1.7%) head trauma and 40 (1.3%) epileptic seizures (Figure 1 ).
We found a history of US in 466 participants (14.8%), 354 young-adults and 112 children, corresponding to a frequency of 13.3% and 22.4% among all TLOC episodes in both strata, respectively (OR 1.875; 95%CI 1.48-2.38,p<0.001). Among individuals with US, 33.9% were males, versus 22.0% in the non-US group (OR 1.815; 95%CI 1.47-2.24,p<0.001), 34.8% had at least 1 major high-risk feature (OR 2.047; 95%CI 1.66-2.53,p<0.001) and 7.3% had 2 of these features. Further information on the syncope high-risk features, as well as on relevant personal and familiar medical history are provided in Table 1 . Regarding ongoing medication, we found no association between psychiatric drugs and US (4.1% vs 5.0%, OR 0.804; 95%CI 0.49-1.31,p=0.382). However, we observed an association between oral anticonceptive agents and a lower incidence of events (49.3% vs 59.8%, OR 0.656; 95%CI 0.52-0.83,p=0.001).
Syncope associated with fever was detected in 3.9% (n=18) of the individuals with US, comparing to 0.6% (n=17) of the individuals with the remaining causes of TLOC (OR 6.310; 95%CI 3.23-12.34, p<0.001). Among those with US associated with fever, four had an rSr’ pattern in V1-V2 precordial leads with an r’-wave >2mm suggestive of non-Type 1 Brugada Pattern (T1BrP), and one had an ECG suggestive of T1BrP. Drug provocative testing was not performed in individuals with non-T1BrP, therefore the diagnosis was not confirmed. In total, only five individuals presented with a ventricular pre-excitation pattern (Supplementary Material 2 and 3 for the ECGs), four of which described a syncopal episode compatible with reflex vasovagal syncope and one reported syncope in a context of disease. Further resting ECG comparisons between participants with US and other causes of TLOC are depicted in Table 2 . Incomplete right bundle branch block was the most frequent conduction perturbance detected (in 15.2% of the cases), followed by 1stdegree AV block (in 1.0%, with a maximum PQ interval of 292ms). There were eleven individuals with complete right bundle branch block (BBB), one individual with complete AV block and two with 2nddegree Mobitz I. LVH voltage ECG criteria were detected in 11.8% of the individuals with US, comparing to 8.1% of the remaining causes of TLOC (Table 2). Additional features commonly associated with pathological LVH such as T-wave inversion, ST-segment depression and pathological Q-waves were present in only 3.3% of the participants with voltage criteria and any TLOC episode. In total, we found that at least 0.8% (n=26, corresponding to 1,3% of participants with unexplained syncope and 0.7% with remaining causes of TLOC, p=0.231) of the participants who reported any past TLOC event presented abnormalities in the resting ECG that can be considered unrelated to regular training or expected physiologic adaptation to exercise and may require further diagnostic investigation. (9)
During the period of implementation of the SCD-SOS survey, high-risk patients were identified and referred for a Cardiology consultation at the National Services Hospital. The case of T1BrP who had a history of US received an implantable cardioverter-defibrillator, as did a patient who was diagnosed with hypertrophic cardiomyopathy. Although we do not have a complete follow-up of patients with syncope (only the individuals with high-risk features were identified and studied), we are currently raising funding to perform this study and present this analysis in future articles.
On a multivariate analysis, the variables that remained independently associated with US were age<18 years-old, male gender, participation in competitive sports, syncope with at least one major high-risk feature, syncope after exertion, number of previous events and fever context and a history of palpitations requiring medical care. This model which included eight variables had a modest accuracy (C-statistics of 0.707; 95%CI 0.69-0.73) (Table 3 ). Finally, the electrocardiographic parameters that have shown differences in the univariate analysis were not independent predictors of US.