Voluntary character of a choice of statistical methods weakens the reliability of the results of heart rate variability assessment and based on them clinical conclusions. The aim of the study was to estimate reliability of the chosen statistical methods of analysing changes in spectral components during tilt testing (TT).
62 patients (34 women), the average age 38.8 +/- 12.8 years, with at least 2 syncopes during the last 6 months that had not been diagnosed by other examinations, were enrolled. The control group consisted of 44 age- and sex-matched volunteers (21 women), the average age 37.4 +/- 12.6 years, who had no history of syncope. All subjects underwent tilt testing according to Westminster protocol. If the result of a passive phase was negative, 0.25 mg nitroglycerine was applied sublingually and the examination was continued for the next 20 minutes. The type of syncopal reaction was defined according to VASIS classification (the vasovagal syncope international study). The evaluation of heart rate variability was based on the frequency domain methods. The low frequency (LF), the high frequency (HF) component and the balance (the LF/HF ratio) were estimated. The response of the autonomic nervous system (AUN) to tilting was evaluated by tests comparing mean values of respective components (the quantitative analysis) and by the modulation index (deltaXFm) reflecting, in percentage terms, changes in analyzed variable values. Taking into account the obtained results there was also a qualitative analysis performed. This analysis referred to the modulation index, which reflected increase or decrease in spectral component value.
Tilt testing explained vasovagal origin of syncope in 35 (60%) patients with syncope. Patients were divided into 3 groups according to anamnesis data and results of tilt testing. Group I consisted of 42 healthy volunteers who had negative result of TT Group II--35 patients with history of syncopes and positive result of TT, and Group III--23 patients with history of syncopes and negative result of TT. In all study groups the quantitative analysis revealed the decrease in the LF and HF component values, whereas the evaluation of autonomic nervous system reactivity based on deltaXFm showed the increase in LF power, concomitant with strongly marked, nearly uniform decrease in HF power in Group I and II. The results obtained from the qualitative analysis did not reveal statistically significant differences between study groups in distribution of defined types of autonomic nervous system reactions.
1. The quantitative analysis has limited credibility in clinical interpretation of changes in autonomic nervous system activity (especially in the low frequency component--LF) occurring during tilt testing. 2. The best estimation of autonomic nervous system activity during tilt testing is assured by the modulation index supplemented by the preliminary selection of data, i.e., rejection of extreme values or those rarely occurring qualitative changes in spectral components.