Impact of population aging on survival outcome in out-of-hospital cardiac arrest: a retrospective observational cohort study

2.1 Study design and setting

This retrospective observational cohort study analyzed data concerning adult patients (age ≥18 years) who experienced nontraumatic OHCA in Ulsan, South Korea, spanning from January 2017 and December 2022.

Ulsan, situated on South Korea’s east coast, covers an area of 1057.136 km² and had a population of over 1.1 million in 2022. Over the study period, the total population of Ulsan showed a progressive decline (1,157,077 in 2017; 1,150,116 in 2018; 1,143,692 in 2019; 1,135,423 in 2020; 1,120,753 in 2021 and 1,110,516 in 2022), while the proportion of individuals aged over 65 years steadily increased (9.9% in 2017; 10.6% in 2018; 11.4% in 2019; 12.4% in 2020; 13.5% in 2021 and 14.6% in 2022), indicating a shift towards an aged society by 2022 (Fig. 1) [8].

Fig. 1.Changes in population distribution by age group during the study period, in Ulsan, South Korea.

Ulsan’s EMS system follows South Korea’s National EMS systems, providing multiple dispatches, a basic-to intermediate-level service staffed by emergency medical technicians (EMTs). In 2022, Ulsan had 30 fire stations and a central dispatch center. EMS teams, typically comprising three highly trained personnel, are equipped to deliver immediate on-site cardiopulmonary resuscitation (CPR) and ensure continuous CPR during ambulance transport to the emergency department (ED). These personnel are certified registered nurses or possess qualifications of level 1 and 2 EMT, which correspond to EMT basic and intermediate levels in the United States [9]. EMS personnel confirm cardiac arrest based on the absence of signs of circulation, specifically absence of consciousness, breathing including abnormal breathing and pulse, or the presence of asystole or pulseless electrical activity on the electrocardiogram monitor. EMS personnel are authorized not to initiate or to cease CPR under specific criteria that are met, such as the return of spontaneous circulation, confirmation of death (livor mortis, rigor mortis or putrefaction) or the presence of a do-not-resuscitate (DNR) order. The decision not to initiate CPR initially is the resuscitation withhold, and to cease CPR upon identifying unmistakable signs of death after initiating CPR is the resuscitation withdrawal. The formal declaration of death is the responsibility of physicians within the hospital EDs. Ambulances are not staffed with physicians. EMS personnel conduct advanced life support procedures under the supervision of a medical director as required by law. Medical directors, mostly emergency medicine physicians, provide direct medical oversight in the dispatch center through communication via real-time video phone consultations with EMS teams at the scene, as well as indirect oversight in the fire station through feedback with post-activity reviews. These procedures encompass tasks such as airway management, establishing intravenous (IV) access, administering fluids or medications, and making informed decisions regarding resuscitation withholding or withdrawal [10].

2.2 Study population

All patients evaluated by EMS personnel as experiencing OHCA during the study period were considered for inclusion. Exclusion criteria comprised cases where resuscitation was withheld or withdrawn due to signs of death (livor mortis, rigor mortis or putrefaction) or a DNR order, suspected traumatic arrest (including trauma, intoxication or drowning) and individuals under 18 years of age.

2.3 Data source and collection

Data for this study were obtained from both prehospital and hospital stages. Prehospital data were sourced from the Ulsan Fire Agency headquarters, where information on factors influencing survival outcomes was electronically compiled from EMS dispatches and prehospital cardiac arrest patient care reports. Hospital data, encompassing survival outcomes, were collected from the EDs of the 17 receiving hospitals within the region.

Data collection adhered to the Utstein style reporting guidelines for OHCA, ensuring standardized reporting practices [11]. Cardiac arrest was confirmed based on the absence of signs of circulation. Patient variables included age, sex, witnessed arrest status, arrest location and comorbidities, such as hypertension, diabetes mellitus, cardiovascular disease, pulmonary disease, liver disease, renal failure and malignancy. Bystander-related variables included bystander CPR and the utilization of bystander automated external defibrillators (AEDs). EMS-related variables included initial rhythm, advanced airway management, mechanical compression device usage, IV access, epinephrine administration and EMS process times (response time, scene time and transport time). Response Time Interval (RTI) was defined as the time from EMS dispatch to EMS arrival at the scene. Scene Time Interval (STI) was defined as the time from EMS arrival at the scene to EMS departure from the scene. Transport Time Interval (TTI) was defined as the time from EMS departure from the scene to EMS arrival at the ED. Hospital variables included survival to hospital discharge and favorable neurological outcomes. All patients, including those transferred to other hospitals, were monitored until discharge. Neurological outcomes were evaluated using the Cerebral Performance Categories (CPC) scale in discharge records, with CPC scores 1 and 2 considered favorable outcomes [12].

2.4 Outcome measures

The primary outcome was the annual trends of survival to hospital discharge and favorable neurological outcomes rate. The secondary outcome was the association between age and survival outcomes.

2.5 Statistical analysis

Descriptive statistics were used to summarize the characteristics of the study population across each year from 2017 to 2022. Categorical variables were presented as frequencies and percentages, while continuous variables were described using the median and interquartile range (IQR), based on non-normal distributions indicated by the Kolmogorov-Smirnov and Shapiro-Wilk tests.

In trend analysis, a chi-square test was utilized to assess the variability in the proportion across years. Linear regression analysis was employed to investigate temporal trends in these rates, evaluating for consistent patterns of increase or decrease and assessing the statistical significance of observed trends.

Multivariable logistic regression models were constructed to identify the independent associations between predictor variables (age) and the primary outcomes (survival to hospital discharge and favorable neurological outcomes) across all years. Adjusted odds ratios (OR) with 95% confidence intervals (CI) were calculated to quantify the strength and direction of these associations while controlling for potential confounding variables.

Cubic spline models were employed to explore the association between age and survival outcomes. Cubic splines are a specific type of spline curve model that offers a flexible approach to capture complex, non-linear patterns in the data, thus providing a more comprehensive understanding compared to traditional linear models. By considering the entire age range of this study, we thought that the associations may not be discernible through the traditional linear modeling technique. Knots, representing points of flexibility in the cubic spline curve, were strategically placed based on cumulative percentiles of the study population at the 5th, 25th, 50th, 75th and 95th percentiles, ensuring optimal capture of underlying data patterns [13]. Statistical analyses were conducted using SAS software (version 9.4; SAS Institute Inc., Cary, NC, USA); p-value < 0.05 was considered indicative of statistical significance.

3.1 Annual trends of the study population and the survival outcomes

The total EMS calls increased from 44,937 in 2017 to 54,296 in 2022. For the EMS assessed OHCA cases, a linear regression analysis was performed to assess the trend observed over the years. The analysis revealed a slope of 0.0008 and an intercept of −1.5807. The correlation coefficient (r) was calculated as 0.7187, a p-value of 0.1076 with a standard error of the estimate of 0.0004, indicating a moderate positive relationship between the variables. It suggests an increasing trend that did not reach statistical significance. The number of EMS treated OHCA cases shows a similar trend. A linear regression analysis was a slope of 0.0004 and an intercept of −0.8493. The correlation coefficient (r) was calculated as 0.6731, a p-value of 0.1429 with a standard error of the estimate of 0.0002, suggesting an increasing trend but statistically insignificant. However, adult nontraumatic OHCA cases treated and transported by EMS personnel (study population) exhibited a significant increase annually with a slope of 0.0029, an intercept of −5.4574, a correlation coefficient (r) of 0.8849, p-value of 0.0250 and a standard error of the estimate of 0.0010 (Table 1).

The median age of study population showed a gradual rise, ranging from 65.0 years in 2017 to 73.5 years in 2022. The proportion of patients aged over 65 years increased from 51.0% in 2017 to 70.0% in 2021. A linear regression analysis reveals a slope of 0.0357 and an intercept of −69.7693. The correlation coefficient (r) was calculated as 0.9564, a p-value of 0.0016 with a standard error of the estimate of 0.0075. It indicates that the proportion of patients aged over 65 shows a statistically significant increasing trend over time (Table 2, Fig. 3).

Fig. 3.The annual trends of the study population.

Survival to hospital discharge rate declined consistently from 24.5% in 2017 to 17.5% in 2018, 14.9% in 2019 and 12.3% in 2021, with a slight increase to 13.1% in 2022. The trend analysis was a slope of −0.0010 and an intercept of 2.0954, a correlation coefficient (r) of 0.0494, and a p-value of 0.6138 with a standard error of the estimate of 0.0023, suggesting a slightly decreasing trend over time, but statistically not significant. Favorable neurological outcomes decreased from 10.1% in 2017, 5.4% in 2018, 7.1% in 2019, 6.6% in 2020, 3.4% in 2021, with a subsequent rise to 5.8% in 2022. A linear regression analysis was a slope of 0.0040 and an intercept of −7.4546. The correlation coefficient (r) was calculated as 0.7539, p-value of 0.0421 with a standard error of the estimate of 0.0020. It indicates a favorable neurological outcomes rate shows an increasing trend over time (Table 2).

3.2 Factors impacting survival outcomes

Age showed a significant association with survival to hospital discharge (OR = 0.970, 95% CI: 0.961–0.978, p-value < 0.001). Patients experiencing OHCA in non-public arrest locations (OR = 0.640, 95% CI: 0.473–0.867, p-value = 0.004), those with unwitnessed OHCA events (OR = 0.532, 95% CI: 0.403–0.702, p-value < 0.001), and those with non-shockable rhythm (OR = 0.273, 95% CI: 0.203–0.366, p-value < 0.001) exhibited lower odds of survive to hospital discharge compared to those sustaining OHCA in public locations, witnessed OHCA, and shockable rhythm (Table 3).

Increasing age was significantly associated with unfavorable neurological outcomes (OR = 0.957, 95% CI: 0.941–0.973, p-value < 0.001). Unwitnessed OHCA events (OR = 0.442, 95% CI: 0.256–0.762, p-value = 0.003) and non-shockable rhythm (OR = 0.055, 95% CI: 0.031–0.096, p-value < 0.001) were associated with decreased odds of favorable neurological outcomes (Table 4).

3.3 Association between age groups and survival outcomes

To perform the cubic spline models, age categories were placed based on cumulative percentiles of the study population at the 5th, 25th, 50th, 75th and 95th percentiles. Age categories were placed based on cumulative percentiles of the study population at the 5th, 25th, 50th, 75th and 95th percentiles. Accordingly, categories were used: age <42 (147 patients, 5.5%), 42 ≤ age < 59 (522 patients, 19.7%), 59 ≤ age < 72 (680 patients, 25.6%), 72 ≤ age < 81 (690 patients, 26.0%), 81 ≤ age < 91 (458 patients, 17.2%) and age ≥91 (159 patients, 6.0%). Patients aged between 72 and 81 years exhibited significantly decreased survival in both unadjusted and adjusted analyses after controlling for covariates including sex, comorbidities, witnessed status, bystander AED, bystander CPR, initial rhythm, EMS process time, advanced airway, mechanical compression, IV access and epinephrine use (unadjusted OR = 0.336, CI: 0.218–0.518, p-value < 0.001; adjusted OR = 0.422, CI: 0.253–0.703, p-value = 0.001). However, for patients aged ≥81 years, the lack of significant results was attributed to the absence of cases in this age group (Table 5).

The spline curve depicting the survival to hospital discharge reveals a notable decline around the age of 50, with a gradual slowing of this decline observed after the age of 80 years. A similar pattern was observed for the OR curve for favorable neurological outcomes (Fig. 4).

Fig. 4.Cubic spline model of age and its impact on survival outcomes in the study population. (A) survival to hospital discharge; (B) favorable neurological outcome. The gray area represents the 95% confidence interval. OR: odds ratios.