Changes in automated external defibrillator use and survival after out-of-hospital cardiac arrest in the Nijmegen area

From our regional prospective resuscitation registry, comprising urban, suburban and rural areas, we selected a study cohort and historical controls. The former consists of non-traumatic OHCA cases (age ≥18 years) resuscitated by the emergency medical services (EMS) between 2013–2016. For reasons of comparability with previous publications, we focussed on patients transported to a single hospital (Radboudumc Nijmegen) [5, 6] and excluded emergency medical service (EMS) witnessed arrests [13]. The latter consists of patients resuscitated between 2008–2011 [23], who fulfilled the same inclusion and exclusion criteria. Given the observational design of the study, it was not necessary to obtain written informed consent according to the Dutch Act on Medical Research involving Human Subjects.

The Dutch EMS system is activated by calling 112. Paramedics will give instructions to the caller to initiate basic life support (BLS), and usually two ambulances are dispatched to the location of the emergency. Ambulances are staffed by a driver and a paramedic, who are professionally trained to perform advanced life support. CPR was performed according to the prevailing guidelines at the time of the OHCA [10‐12].

Demographic, clinical and arrest characteristics were collected using EMS and hospital records. Variables were defined according to the Utstein style definitions [24]. Location was divided into public or non-public [25]. Outcome data were collected using hospital records and information from municipal registries.

Patients were categorised as follows. The study cohort included patients resuscitated between January 2013 and December 2016, while the patients in the historical control cohort were resuscitated between April 2008 and January 2011.

The primary outcome was survival to hospital discharge. The secondary outcome was sustained ROSC, defined as hospital transportation with ROSC. We furthermore reported on any ROSC in the field, ROSC at the emergency department (ED) and 24-hour survival.

Data were presented and compared as described previously [23]. Analyses were performed separately for 2008–2011 and 2013–2016. To evaluate whether AED use and bystander CPR were associated with survival, multivariate binary logistic regression analysis was performed. All factors that were univariately associated with survival (p < 0.10) were included. Results were reported as adjusted odds ratios (aOR) and their 95% confidence intervals. A p-value of <0.05 was considered statistically significant for all analyses (SPSS Statistics v.22, IBM-Corp., Armonk, NY, USA).

The proportion of patients who received bystander CPR was higher in the study cohort than in the historical cohort (78% vs. 63%, p < 0.001). The proportion of patients in whom an AED was attached was also higher (46% vs. 23%, p < 0.001). Furthermore, a higher proportion of patients received a shock by the AED (39% vs. 15%, p < 0.001), less patients received a shock by the EMS (59% vs. 76%, p < 0.001) and the median number of EMS shocks was lower in the study cohort (2 vs. 4, p = 0.004).

In the study cohort, survival to discharge was higher than in the historical controls (47% vs. 33%, p = 0.002), while ROSC rates were similar (65% vs. 67%, p = 0.74). No significant differences were found in the other outcome parameters (Tab. 2). The survival benefit was restricted to patients with a shockable initial rhythm (58% vs. 41%, p = 0.002 in shockable and 13% vs. 9%, p = 0.51 in non-shockable rhythms).

In the study cohort, we found that younger age, male gender, public location, bystander witnessed, shockable initial rhythm, bystander CPR and AED use were associated with survival. Bystander CPR (aOR 2.83, [1.45–5.51], p = 0.002) and AED shocks (aOR 2.36 [1.40–4.00], p = 0.001) were independently associated with survival. In the historical cohort, younger age, female gender, bystander CPR, shockable initial rhythm and AED shocks were associated with survival. Bystander CPR (aOR 2.45 [1.14–5.27], p = 0.02) and AED shocks (aOR 3.18 [1.25–8.09], p = 0.02) were independently associated with survival.

In 2015, the Dutch Heart Foundation initiated the Working Group Cardiac Arrest Netherlands (weCAN). This working group aimed to unite experts in the field (from Amsterdam, Utrecht, Maastricht and Nijmegen) and report on outcomes and recent developments in OHCA. The main finding was that overall survival increased from 9 to 23% in recent years [3]. Notably, this concerns survival of all OHCA patients, whereas we now focus on patients who were transported to the hospital, similar to other recent publications on this subject [5, 6]. We found that survival rates in recent years increased from 33 to 47%. In Maastricht, a survival rate of 46% was found, compared with a survival rate of 43% in Leiden [5, 6]. Overall, baseline characteristics are comparable among these three cohorts. However, in Leiden there were lower proportions of shockable initial rhythms (61% vs. 74% in our 2013–2016 cohort) and bystander CPR (51% vs. 78% in our cohort). As these are both strong predictors of survival, this might explain our slightly higher survival rates.

Since its development, the AED has been extensively investigated with regard to safety, costs and its effect on survival [13‐16, 26]. In recent years, AEDs have become increasingly available to first responders and bystanders, resulting in improved outcome [17, 27]. In our region, a civilian-based foundation has developed several initiatives to improve resuscitation education and establish a network of AEDs throughout the municipality. This foundation acquires, registers and maintains AEDs and is funded by private sponsoring. Members of the foundation also promote and facilitate registration of BLS providers and AEDs in a nation-wide, text message based civilian alert system. In addition to these continuing and expanding efforts throughout our community, regional government has contributed as well, by moving indoor AEDs to outside locations. This increases the availability of these AEDs from exclusively during business hours to 24/7. These AEDs were also registered in text message alert systems.

In the present study, we found an increase in bystander CPR and AED use. We also observed that survival in our region has significantly increased over the last years. In both cohorts, AED use was independently associated with survival. The fact that the improvement in survival was restricted to patients with a shockable rhythm is in line with results from a previous meta-analysis [26]. These observations support the concept that increased AED use contributed to our improved outcomes [13‐15]. At present, our database does not systematically capture the use of the text message systems, whether the AEDs used were registered AEDs and whether BLS was performed by trained volunteers. This limits conclusions with regard to the contribution of each of the individual factors.

ROSC rates were high in 2008–2011 and have remained high, yet more patients now survive until discharge. Faster restoration of cerebral perfusion may account for this, as patients treated with an AED have a shorter CPR duration and, thus, shorter periods of cerebral hypoxia, a major determinant of survival after OHCA [28]. Even though we did not have the exact times to ROSC, several findings point towards shorter CPR duration. Firstly, we found decreased use of epinephrine and amiodarone as well as a lower number of EMS shocks in the study cohort. Secondly, in the patients with an AED attached (46% study cohort vs. 23% historical control cohort), the proportion receiving a shock was markedly higher (Tab. 1): 85% in the former versus 66% in the latter. This may be related to earlier attachment of the AED, as the chances of finding a shockable rhythm increase with shorter arrest durations [29]. Thirdly, a larger number of patients with a shockable initial rhythm did not receive a shock by the EMS in the 2013–2016 cohort. Notably, all these patients received an AED shock, implying earlier termination of the shockable rhythm. We hypothesise that these early benefits of the AED are sustained throughout the post-resuscitation period and ultimately result in increased survival. However, despite these plausible clues, we cannot firmly objectify earlier AED attachment in lack of data on AED time intervals. Our future registration will also focus on these intervals, to further objectify benefits achieved with lay rescuers.

Our findings underscore the importance of increased rates of bystander CPR and AED use, and support all initiatives to increase awareness for early BLS and use of an AED in case of a cardiac arrest. Although part of the observed increase in bystander CPR will be related to increased awareness through national media campaigns (‘6-minute zones’, Dutch Heart Foundation, https://​www.​hartstichting.​nl/​reanimatie/​6-minutenzone), our local initiative might have contributed as well. Importantly, our findings have a descriptive nature, and should not be interpreted as the result of a local intervention.

Previous publications have shown the value of a text message based alert system in the Netherlands. This system alerts registered volunteers to go to the site of OHCA, directly or via the nearest AED [18‐21]. In recent years, the number of volunteers registered in these systems in the Netherlands increased from 41,805 in 2012 to 95,212 in 2015. The number of registered AEDs also increased, from 5,275 to 7,945. Appreciating the potential benefit of civilian-based initiatives in terms of higher availability of AEDs and more trained volunteers, we strongly support all innovations in this context and encourage healthcare providers to subscribe (www.​hartslagnu.​nl and www.​hartveiligwonen.​nl). Furthermore, we feel that our initiative to increase 24/7 accessibility of AEDs could be a promising intervention to increase early AED use.

This is an observational cohort study, and although the historical controls are part of the same prospective registry, not all differences that develop over time will have been captured which may have affected our results. In particular, even though the ratio public/non-public and median EMS response times did not differ between cohorts, we cannot exclude small differences with regard to the location of the arrest. As for underlying aetiology (presumed cardiac, traumatic, other) no differences were observed between cohorts. Importantly, the vast majority of other baseline characteristics did not differ, except for bystander CPR and AED use. Rates of in-field termination and hospital transportation have been stable over the years as well [23]. Thus, although unmeasured variables may have influenced outcome, this will only have partly influenced the magnitude of the observed survival benefit, and it is unlikely that this has affected the observed direction of the effect on survival. For reasons of comparability [5, 6], our data are restricted to OHCA patients transported to the hospital, which should be noted when comparing survival rates with studies that include all OHCA victims [30]. Due to logistical issues, no data were available for 2012.