While most inappropriate therapies observed in patients with implanted intracardiac defibrillators are secondary to supraventricular rhythm disorders or lead dysfunction, T wave oversensing is the most common cause of inappropriate therapies for subcutaneous devices. The occurrence of inappropriate therapies was more frequent for early defibrillator platforms. subcutaneous but has decreased significantly in newer devices. This decrease has been facilitated by the systematic implementation of a pre-implantation screening procedure, routine programming of a conditional zone, and the use of new algorithms during the detection and certification phases. To classify a cycle, there are three successive phases: a detection phase with input signal filtering, a certification phase to remove extracardiac signals, and a decision phase.

• The device uses an adaptive detection threshold based on the amplitude of the two previous certified cycles. Therefore, oversensing of P waves and/or T waves is favored by the low amplitude of the QRS complexes. When the R wave is of low amplitude, the device quickly reaches high levels of sensitivity, favoring oversensing of the T wave, particularly during exercise (possible decrease in the amplitude of the R wave and slight increase in the amplitude of the T wave). A low-amplitude R wave is mostly observed in patients with arrhythmogenic right ventricular dysplasia, Brugada syndrome, cardiac sarcoidosis, or dilated cardiomyopathy.
• The sensitivity level increases in parallel with the increase in heart rate to optimize detection capabilities during tachycardia. Depending on the frequency, the duration of refractory periods and the adjustment of the sensitivity level depend on the amplitude of previous signals and the programming. For the same lower limit of 200 bpm, programming two zones with a conditional shock zone of 200 to 230 bpm makes the device less sensitive and therefore less prone to oversensing than programming a single shock zone at 200 bpm.
• Once the signal is detected, the certification phase classifies the cycle as certified or as noise/oversensing. Signals with too high a frequency (myopotentials, interference, probe fracture) are classified as noise and are excluded from the heart rate calculation. The signals then go through four additional certification steps to diagnose double counting of the R wave and/or oversensing of the T wave. 

The various algorithms involved in diagnosing oversensing of the T wave, P wave, or R wave (double counting) are based on measuring the minimum and maximum amplitude of the detected complexes, the measurement of the delays between detections and the width of the signals, and on the comparison of the morphology of the detected signal with the morphology of the previous complex and the morphology of the reference complex.

• The first algorithm compares the morphology of the detected signal with the reference morphology; oversensing is suspected when there is an alternation between two signals, one correlated with the reference and one uncorrelated: repetition of sequences correlated signal-uncorrelated signal-correlated signal, suggesting oversensing of a supernumerary cardiac signal.
• The second algorithm searches for double counting of QRS complexes by looking for a repetitive pattern of alternating morphology in short and fixed coupling detections.
• The third algorithm searches for alternation between the duration of two cycles (alternation between two intervals) and is therefore not based on an analysis of the signal morphology.
• The fourth algorithm analyzes the morphology of three successive detected complexes and searches for the presence of a sequence of morphology 1 (QRS complex) – morphology 2 (different from the first; T wave) – morphology 1 (QRS complex). To improve the algorithm’s performance, the width of the complexes and intervals are also included in the analysis; the operation of this algorithm therefore does not depend on the quality of the reference morphology recording. When one of the algorithms diagnoses an oversensing, the signal is marked as uncertified with a • on the trace without specifying which algorithm led to the diagnosis.

These different algorithms can be misled in cases of variability in signal amplitude or morphology, intermittent oversensing, or double oversensing. In addition, some of the oversensing diagnostic algorithms have only been integrated into the most recent platforms and are therefore absent from older devices. Oversensing of the T wave can be complicated by the occurrence of multiple inappropriate shocks with a risk of inducing VF, as the shock may be delivered during a vulnerable period.