This retrospective observational administrative database analysis used the Korean National Health Insurance Service (NHIS) data provided by the Korean government. Ethical approval was obtained by the Institutional Review Board at the authors’ institution (No. 2022-07-003), and the study follows principles embodied in the Declaration of Helsinki. Data were obtained from the NHIS under a confidentiality agreement for restricted use.

Patients aged >18 years with an International Classification of Diseases, Tenth Revision (ICD-10)-based SAH disease code (I60) and a concomitant surgical prescription code (S4640, S4641, S4642, M6641, M1661, M1662) between January 2005 and December 2020 were included. Patients who received a hemorrhagic stroke diagnostic code (I60, I61, I62) or aneurysm surgical code (S4640, M6641) during the washout period between January 2002 and December 2004 were excluded. Data regarding the prescription of tracheostomy (O1300, O1301) and mortality were collected until December 2021, ensuring at least a 12-month follow-up for all patients. Additionally, the 16-year observation period was analyzed by dividing it into four 4-year intervals (2005-2008, 2009-2012, 2013-2016, and 2017-2020).

Microsurgical clip operation was defined as cases with a first surgical prescription code for a clipping operation (S4640, S4641, S4642). Endovascular treatment was defined as cases with a first surgical code for endovascular coiling (M6641, M1661, M1662). We assumed that the use of stent-assisted coiling would have increased; thus, simple coils and stent-inserted coils were analyzed separately for patients who underwent endovascular treatment depending on intracranial stent prescription. The location of the cerebral aneurysm was classified according to the diagnostic codes (internal carotid siphon and bifurcation, I600; middle cerebral artery, I601; anterior communicating artery, I602; posterior communicating artery, I603; basilar artery, I604; vertebral artery, I605; unspecified, I60, I606, I607, I608, and I609). Mortality was defined as death within 3 months after the admission date. Poor prognosis was defined as death within 3 months or patients who required tracheostomy.

Age was reported as the mean (±standard deviation) and was compared using an independent t-test. Variables other than age were reported as the frequency (percentage) using Fisher’s exact test or the chi-square test. Statistical significance was set at P<0.05. A multistep logistic regression analysis was performed to evaluate the risk factors involved in mortality and poor prognosis, with tracheostomy used as a proxy measure for a poor prognosis, rather than the modified Rankin scale, which was not recorded in the NHIS database. Variables within a value of P<0.10 were re-entered into the multivariate logistic regression model using the backward stepwise method. The odds ratios (ORs) and 95% confidence intervals (CIs) are presented for significant factors (P<0.05). A simple regression analysis was performed to evaluate the yearly trends in mortality and in poor prognosis. All statistical analyses were performed using Software Enterprise (version 7.1; SAS Institute Inc., Cary, NC, USA) and R Studio (version 4.0.5; R Foundation for Statistical Computing, Vienna, Austria).

Overall, 83,587 patients with SAH and surgical prescription for either clip or coil were included in the study. Female predominance (64.5%) was observed. The mean age was 56.9±13.0 years. The 3-month mortality rate was 12.5%. The characteristics of the patients are presented in Table 1. The proportion of patients whose cerebral aneurysm location was not recorded was 35.9%. Among those for whom the location was recorded, the anterior communicating artery was the most common site (25.0%). Endovascular treatment was slightly dominant over the use of microsurgical clips (50.1% vs. 49.9%), with simple coils and stent-assisted coils accounting for 38.5% and 11.6% of the cases, respectively.

The yearly trend in treatment methods, poor prognosis, and mortality rates are described in Table 2 and presented in Fig. 1. The mortality rate fluctuated between 12.2% and 13.6% from 2005 to 2015, and then remained stable at <12% since 2016. The poor prognosis rate fluctuated between 20.0% and 21.6% from 2005 to 2015, and then remained stable at <20% since 2016. Microsurgical clip use continued to decline, starting from 78.8% in 2005 to 24.4% in 2020. Simple coil embolization gradually increased, starting from 21.2% in 2005, and then plateaued at approximately 50% since 2017 onwards. Stent-assisted coil embolization gradually increased, starting from 0.4% in 2006 and exceeded microsurgical clip use in 2020 (24.7% vs. 24.4%).

The mortality rate per year tended to decrease over time, with significance (rate per year: -0.0010%, R²=0.4607, P=0.0038) (Fig. 2). Furthermore, the poor prognosis rate per year also showed a significant decreasing trend over time (rate per year: -0.0016%, R²=0.4957, P=0.0024) (Fig. 3). The results for the observational period, divided into four 4-year intervals, are displayed in Supplementary Fig. 1 and summarized in Supplementary Table 1.

The results of the multistep logistic regression model for mortality are presented in Table 3. Age (hazard ratio [HR]: 1.04, 95% CI: 1.04-1.04, P<0.0001) and sex (female; HR: 0.74, 95% CI: 0.70-0.79), P<0.0001) were significantly associated with mortality. Endovascular treatment was positively related to mortality, as opposed to microsurgical clipping (HR: 1.13, 95% CI: 1.07-1.19, P<0.0001). Regarding aneurysm location, the posterior communicating artery had the lowest risk (HR: 0.66, 95% CI: 0.56-0.76, P<0.0001) for mortality rate. Contrastingly, the vertebral artery had the highest risk (HR: 2.27, 95% CI: 1.88-2.74, P<0.0001).

The results of multistep logistic regression model for poor prognosis are presented in Table 4. Endovascular treatment showed a positive correlation with mortality and a negative correlation with poor prognosis (HR: 0.93, 95% CI: 0.89-0.98, P=0.005). The posterior communicating artery location had the lowest risk (HR: 0.68, 95% CI: 0.60-0.77, P<0.0001) for a poor prognosis. Contrastingly, locations in the basilar (HR: 1.26, 95% CI: 1.07-1.47, P=0.0051) and vertebral arteries (HR: 2.54, 95% CI: 2.14-3.00, P<0.0001) were strongly related to a poor prognosis (tracheostomy).

Characteristics and clinical outcomes according to the aneurysm location are presented in Supplementary Table 2. In comparison to other sites, the anterior communicating artery (males: 49.9%, mean age: 55.9±12.2) and vertebral artery (males: 46.2%, mean age: 53.3±12.5 years) were involved in a higher proportion of men and tended to appear in younger individuals. In contrast, the posterior communicating artery was more often involved in female patients (females: 83.8%) and tended to be more involved in older individuals (mean age: 60.6±13.2 years). The yearly trends in characteristics and clinical outcomes according to aneurysm location are presented in Supplementary Tables 3-7.

Our trend analysis of nationwide data of patients with SAH due to ruptured cerebral aneurysm who received surgical treatment in the Republic of Korea between 2005 and 2020 confirmed a trend toward decreased mortality (rate per year: -0.0010%) and good prognosis (rate per year: -0.0016%). Microsurgical treatment decreased by two-thirds between 2005 (78.8%) and 2020 (24.4%), whereas endovascular treatment increased rapidly, accounting for 75.6% in 2020. Additionally, the stent-assisted coil embolization rate surpassed that of microsurgical treatment in 2020 (24.7% vs. 24.4%). Endovascular treatment was identified as a risk factor that increased mortality and, conversely, reduced the likelihood of a poor prognosis.

The number of patients who underwent aneurysm repair for SAH had substantially increased between 2005 (4,310) and 2020 (5,893) in the Republic of Korea. Additionally, the average age of patients increased by approximately 5.2 years over the 16-year period (2005: 54.6±12.1 years to 2020: 59.8±13.5 years). Despite the increase in the average age of patients, the prognosis of patients improved. The proportion of female patients gradually increased, accounting for 64.5%. The average life expectancy of women in the Republic of Korea is approximately 6 years longer than that of men, suggesting that the SAH rate in older adults has increased.

The most notable finding was the sharp decline in microsurgical clip treatment and the increase in endovascular treatment, which was reversed in 2013. In the Republic of Korea, an intracranial aneurysm coil device has been covered by insurance since 2001 and a stent device since 2006. Simple coil use has remained stable at approximately 50% since 2017, and use of stent-assisted coils has gradually increased, reaching 24.7% in 2020. For patients with SAH caused by middle cerebral artery rupture, the proportion of surgical treatment decreased by about half, from 91.8% in 2005 to 45.9% in 2020. In patients with the location being the posterior communicating artery, the proportion of microsurgical clip use decreased by more than two-thirds (2005: 76.7% to 2020: 20.5%), and for the anterior communicating artery, its use decreased by three-fourths (2005: 82.3% to 2020: 21.0%). Stent-assisted coils accounted for 24.0% of SAH treatments in the anterior communicating artery and 27.0% for those in the posterior communicating artery in 2020. Availability of low-profile stents, better navigation devices, and stable anti-thrombotic agents may have been attributed to the increase in stent-assisted coil use.

Based on our comprehensive analysis of treatment trends, endovascular treatment was essentially established as the first-line treatment method for ruptured cerebral aneurysms, and surgical clipping was considered only when endovascular treatment was impossible, posed a higher risk, or if the amount of intracerebral hemorrhage was large and required surgical removal8).

Aneurysmal SAH is a disease in which a good prognosis cannot be expected due to the high mortality rate before surgical treatment and the technical difficulty of surgical treatment by either clip or coil. The global estimated mortality rate is approximately 19-20%, but differs widely according to nation [9]. A United States study using the Nationwide Inpatient Sample Database from 2006 to 2018 showed a yearly mortality rate of 13.1-13.7% and did not demonstrate a significant trend in mortality reduction [6]. In our study, the 3-month mortality rate had a declining trend (rate per year: -0.0010%, R²=0.4607, P=0.0038) (Fig. 2) from 2005 (13.0%) to 2020 (11.6%). In addition, the mortality rate has remained stable at <12% since 2016. It is expected that the increase in endovascular treatment, which involves a relatively short learning curve duration, may have contributed greatly to the decline in overall mortality.

It should be emphasized that this study did not use data only from large tertiary hospitals; nationwide data was used, including all citizens and all hospitals. Considering that it is dangerous for patients to be transferred long distances, the widespread use of endovascular treatment is expected to be a good option for treating patients without transport to more specialized hospitals. Furthermore, it is possible that advances in drug treatment and intensive care contributed to the overall decline in mortality.

The most interesting aspect of this study is that endovascular treatment increased mortality but conversely reduced the possibility of a poor prognosis. We encountered limitations in interpreting this phenomenon because the study could not confirm patients’ radiological results or neurological symptoms. Compared to microsurgical clips, the most notable limitation of endovascular treatment is rebleeding. A high rebleeding rate after endovascular treatment may be related to the increased risk of mortality [10,11]. A systematic meta-analysis has shown that endovascular treatment results in a higher incidence of rebleeding (OR: 2.18, 95% CI: 1.29-3.70) than surgical ligation [12]. A Japanese study also argued that clip ligation should be considered as a first-line treatment because it has a lower mortality rate than endovascular treatment [13].

In this study, the patients who died or who underwent tracheostomy were defined as having a poor prognosis. Although it is appropriate to estimate prognosis through the modified Rankin Scale, the NHIS data do not record this information; thus, we used tracheostomy as a proxy for poor prognosis. The data from this study indicated that use of a microsurgical clip reduced the mortality rate and resulted in a higher rate of tracheostomy than endovascular treatment. Several studies have reported that, although endovascular treatment has a higher mortality rate, it has better results in terms of various complications and long-term prognosis than surgical ligation, which is consistent with our results [8,12,14-16].

This study had the strength of being a nationwide study that included all patients in the Republic of Korea, that is, not only large medical institutions but also small medical institutions that would otherwise have involved limited sample sizes which are too small for publication of research results. Globally, studies analyzing treatment outcomes at a national level are rare.

This study had some unavoidable limitations. It used insurance claim-based data, which was not derived for research purposes. The definitions of surgical treatment and clinical outcomes were based on prescription data, which have the risks of mis-recordings and omissions. The location of the aneurysm was unspecified in 35.9% of cases. The ICD-10 codes are limited concerning the location of the aneurysm. The most notable limitation was that patient conditions and radiological examinations were not determined. This study is meaningful in terms of analyzing the trends of treatment methods and clinical outcomes over time; however, the results should be interpreted cautiously.