Ethics, participants and sample size

This study was approved by the Ethics Committee of Kuopio University Hospital (Study 91/2012, 26 March 2013,ClinicalTrials.gov Identifier: NCT01824212). The hypothesis was that the quality of bipolar ECG readings recorded within an area the size of a mobile phone is good and that cardiologists and AED software can analyse ECG readings correctly to distinguish between shockable and non-shockable rhythms with an accuracy rate of 100%. The current system recognises ~80% of OHCAs in EMCC. The target for OHCA recognition should be 100%. Using this target in sample size calculations, we compared the accuracy rate of 100% with the current accuracy rate of 80% for OHCA recognition at EMCCs. According to the power calculations using a one-sided binomial test (risk level, 0.05; power, 95%), 22 observations would be required to detect differences between the expected success rate (99%) and the current success rate of 80%.

Consequently, 22 patients scheduled to undergo ICD surgery were recruited. The patients provided their informed consent before participating. All patients were evaluated by cardiologists to be appropriate for induction of VF under sedation after implantation of the ICD. The characteristics of the participants are provided in Table 11.

Table 11. Characteristics of the patients. Numbers shown are data values or mean ±SD calculations.

6.3.2 Study design

Before ICD implantation the height and weight of the patients as well the length of the sternum from the jugular notch to the xiphoid process were measured, and the midpoint of the sternum was marked. After the standard implantation procedure the patient’s spontaneous rhythm was recorded and printed with two AEDs.

The first AED (R-series ALS; Zoll Medical Corporation, Chelmsford, Massachusetts, USA) recorded the rhythm vertically at the midpoint of the sternum with small ECG pads (3 × 3 cm, with a 3 cm distance between pads) cut from standard pads (CPR Stat-Pads; Zoll Medical Corporation, USA) to fit an area typical of a mobile phone (5 × 11 cm). Pads were trimmed to be just outside the wiring zone so that wiring remained intact for the recordings. The trimmed sternal pad served as the proximal electrode and the apical pad as the distal one. The centre of the electrode-free area was set at the midpoint of the sternum (Fig.7).

Recordings with the second AED were performed with standard defibrillation pads placed at standard points on the body. The sternum electrode was placed below the right clavicle and the apical electrode was placed along the mid-axillary line below the left armpit (Fig. 7). The recordings were made at the same time as recordings made with the first AED, and served as a reference.

Patient sedation was induced by an intravenous administration of propofol (Propolipid® 10 mg/ml;

Fresenius Kabi, Graz, Austria) after registration of a normal heart rhythm. VF was induced according to the standard procedure with 2 s of direct current, followed by automatic defibrillation induced by the implanted ICD (St Jude Medical, Saint Paul, Minnesota, USA; or Medtronic, Dublin, Ireland). During the VF period, ECG readings were recorded at the positions described previously. All ECG recordings were printed for later analysis.

Figure 7. ECG recording sites. Recordings with large standard pads were performed at standard points using small pads within an area the size of a mobile phone.

Age

(years) Sex

(male/female) Weight (kg) Height (cm) Length of sternum (cm)

Reason for implantable cardioverter defibrillator, primary/secondary prevention

58 ± 13 14 /8 80±17 171±8 22±3 8/14

6.3.3 Measurements

Recordings collected with small electrodes at the midpoint of the sternum were classified with AED software as shockable or nonshockable. The size of the ECG amplitude of each recording was measured manually from the ECG printouts (mV). The reported amplitude was calculated as the average of five consecutive amplitudes taken from the middle of each recording.

The duration of VF was measured by the ICD as the sum of diagnosis time and ICD charge time. The time taken by the AED for rhythm analysis was measured manually with a chronometer. The chronometer was started when the rhythm analysis function of the AED was switched on and stopped when the AED displayed the recommendation to defibrillate or not. Results are reported as the mean ± SD values.

The quality of all recorded ECG printouts was cross-checked and assessed by cardiologists. Both cardiologists graded each recording using a standardised scheme according to four quality factors (Table 12). Each factor was assessed on a scale of 0-2 such that each recording received a grade from each cardiologist between 0 and 8 (Table 12). The total quality score for each recording was calculated as the sum of the grades of both cardiologists and resulted in scores between 0 and 16. Scores were analysed to determine median and inter-quartile range (IQR) according to the recording position (small pads mid-sternum or standard pads on standard regions) and the recording state (basic rhythm or VF). In addition to assigning scores, cardiologists also identified the rhythm of the ECG print and classified the ECG readings as shockable or nonshockable rhythms.

The Wilcoxon signed-rank test was performed using IBM SPSS Statistics 23 software (IBM, North Castle, New York, USA) to evaluate the differences in the quality scores and the amplitudes in recordings of normal rhythm and VF with standard defibrillation pads placed at standard points or with small pads at the midpoint of the sternum. The significance level of the test was set to 0.05.

Table 12. Assessments of ECG recordings by cardiologists. The assessment of ECG recordings was divided into (A) rhythm analysis and (B) quality analysis according to set criteria. VF, ventricular fibrillation.

A ) Recognition of the rhythm Is the rhythm shockable? Yes/no

Is the rhythm identifiable? Yes/no

What is the rhythm in the ECG-print? Correct/wrong B) Structured assessment of the ECG

print First impression of the readability and the

overall quality of the ECG print 0 (bad)/ 1 (average)/ 2 (good) Is the QRS configuration or equivalent (VF)

identifiable?

0 (badly)/ 1 (average)/ 2 (easily)

The size of QRS configuration or equivalent

(VF)? 0 (too small, hard to read)/ 1 (small

but easy to read)/ 2 (suitable size, easy to read)

Amount of artefact or noise? 0 (noisy, hard to read)/ 1 (some artefact but easy to read)/ 2 (no remarkable artefact)

6.4 RESULTS