Chronological and Biological Age in Predicting Outcomes of Older Patients in Neurosurgery

Abstract

Although chronological age is an important factor in indications and predicting outcomes of neurosurgery, it is essential to consider biological age, particularly in older patients, due to individual differences such as frailty. The simplified 5-factor modified frailty index has recently been introduced. This study investigated its role in predicting the outcomes of meningioma and unruptured cerebral aneurysm surgery by analyzing data from the Diagnosis Procedure Combination database in Japan from 2010 to 2014. Although the 5-factor modified frailty index scores could predict the risk of in-hospital worsening outcomes, mortality, and complications in meningioma surgery, it was more useful in non-elderly patients aged <65 years rather than in elderly patients aged ≥75 years. Additionally, in patients aged <74 years, in-hospital complications of unruptured cerebral aneurysms were more associated with the 5-factor modified frailty index than with chronological age.

Alternatively, in patients with aneurysmal subarachnoid hemorrhage, previous reports have suggested a non-linear correlation between age and the outcome, but no reports have explored this relationship. Therefore, we visualized a clear non-linear correlation between age and poor outcomes, which can aid clinical decision-making and better inform and guide patients with aneurysmal subarachnoid hemorrhage and their families. We could validate this visualization using a separate cohort based on the discrimination property and calibration plot. Progress has been made in predicting outcomes for older patients undergoing neurosurgery in Japan; however, in the future, more individualized and specific predictions will be required.

Introduction

The aging of populations in developed countries presents an unparalleled challenge to future societies. According to the Organization for Economic Cooperation and Development, Japan has the highest average life expectancy globally.¹⁾ Therefore, Japan serves as an appropriate country to conduct research on the effects of aging. As life expectancy increases, neuroimaging techniques such as brain magnetic resonance imaging have led to the detection of asymptomatic lesions such as meningiomas and unruptured cerebral aneurysms (UCAs) in older adults worldwide, necessitating attention to these matters.

The heterogeneity of older adults in terms of physical, mental, and social characteristics complicates the process of using chronological age alone to determine the optimal decision-making process in the treatment of older patients.^2-4) Although chronological age is a traditional determinant of the decision-making process, it is increasingly recognized that relying solely on chronological age may be inadequate. Given the highly variable nature of the aging process at the individual level, it is essential to assess both chronological and biological ages. Therefore, this review article examines the factors that influence the outcome in neurosurgery using the simplified 5-factor modified frailty index (mFI-5),^5-7) which is a straightforward measure of biological age to examine factors that influence outcomes in neurosurgery. Furthermore, the relationship between chronological age and outcomes in aneurysmal subarachnoid hemorrhage (aSAH) is generally a non-linear correlation that needs to be accurately portrayed. However, this relationship is not adequately visualized. Therefore, this study also examined the visualization of the non-linearity between age and outcome.

The simplified mFI-5

A recurring theme in the assessment of frailty is the heightened vulnerability to stressors resulting from diminished physiological reserves.^8-10) Consequently, individuals with frailty are at an elevated risk of experiencing adverse clinical outcomes in response to these stressors.¹¹⁾ Various conceptual models have been proposed to elucidate the pathophysiological mechanisms underlying the onset and progression of frailty, with the 2 most widely cited being the “phenotypic” model described by Fried et al.¹²⁾ and the “deficiency” model by Rockwood et al..¹⁰⁾ In the “phenotypic” model, frailty is defined by the presence of specific clinical features, including reduced lean body mass, diminished muscle strength, decreased endurance, impaired balance, slowed gait speed, and low physical activity. Assessment within this framework is based on the presence or absence of these predefined characteristics. Conversely, the “deficiency” model (e.g., Canadian Study of Health and Aging) conceptualizes frailty as a cumulative state of vulnerability, assessed by quantifying the total number of physiological impairments and clinical deficits. This model incorporates a broad spectrum of deficits, ranging from impairments in activities of daily living to neuropsychiatric symptoms such as mood disorders.

Some research suggests that factors contributing to frailty may emerge before individuals reach old age,¹³⁾ indicating that frailty represents a maladaptive model of the aging process. A perioperative outcome of mortality was associated with frailty in 8 studies.^8,14-20) For both men and women, there is a significant non-linear association between age and frailty index (FI) scores.²¹⁾ Overall, FI scores are found to be a more potent predictor of mortality than age.²¹⁾ In this study, it was observed that mFI-5 scores increased with age, similar to the findings in previous reports examining FI.²²⁾ It is important to recognize that aging is a heterogeneous process, and chronological age does not necessarily reflect an individual's health status. The relative risk of death in young adults is more associated with frailty than with age.¹³⁾

To assess frailty, the 11-factor modified FI (mFI-11) was derived from the Canadian Study of Health and Aging Frailty Index.²³⁾ The index was developed by matching 11 comorbidity and deficiency variables from the mFI-11, which are well-validated health indicators that have been applied to general medicine and surgery data sets.^5-7,24) Recently, the mFI-5 was introduced and validated in various surgical fields. This index includes the following 5 factors: prevalence of functional dependence, history of diabetes, history of chronic obstructive pulmonary disease, congestive heart failure, and hypertension.^25-28) However, this study aimed to determine whether the mFI-5 score is associated with the same level of risk in all age groups in neurosurgery. Specifically, its usefulness was examined in predicting complications and outcomes of meningioma and UCA surgery in different age groups.

The mFI-5 score and chronological age in meningioma surgery

The outcomes of meningioma and UCA surgery were investigated using data from the Diagnosis Procedure Combination database in Japan for the period 2010 to 2014. For multivariate logistic regression analyses, independent variables were selected based on the existing literature, and no variable selection method was applied. In our previous study of meningiomas,²⁹⁾ although chronological age was not associated with in-hospital mortality, an mFI-5 score of 1 or 2 or higher was associated with in-hospital mortality (Fig. 1). When the data were examined according to the age group, chronological age was not associated with complications in the <65-year-old group but was associated with the 65- to 74-year-old group and with individuals aged ≥75 years. Conversely, the mFI-5 score of ≥2 was associated with the <65 years (2.5: 1.8-3.6) and 65-74 years (1.5: 1.1-2.1) year-old groups but not with the ≥75-year-old group, indicating a stronger association in younger patients (Fig. 2). Furthermore, when worsening Barthel index (BI) was considered as a worsening functional outcome, chronological age was associated with only the 65-74- and ≥75-year-old groups. The mFI-5 score of ≥2 was associated only with the <65-year-old group (Fig. 3).

Fig. 1.

Forest plots of the risk factors for in-hospital mortality. *p < 0.05 (cited from ref. 29).

E: elderly; HV: hospital volume; mFI-5: 5-factor modified frailty index; NE: non-elderly; PE: pre-elderly

Fig. 2.

Forest plots of the risk factors adjusted for other variables for any complications in the non-elderly, pre-elderly, and elderly. *p < 0.05 (cited from ref. 29).

HV: hospital volume; mFI-5: 5-factor modified frailty index

Fig. 3.

Forest plots of the risk factors adjusted for other variables for the worsening BI scores in the non-elderly, pre-elderly, and elderly. *p < 0.05 (cited from ref. 29).

BI: Barthel index; HV: hospital volume; mFI-5: 5-factor modified frailty index

The mFI-5 score is a useful predictor of postoperative complications and worse functional outcomes, particularly in patients with meningiomas aged <65 years. Therefore, caution should be exercised when using the mFI-5 score to make decisions in meningioma surgery based on the age of patients. Moreover, preoperative physical fitness could be recommended to improve frailty, especially in patients aged <65 years, who are more likely to regain fitness than older patients.¹³⁾ A recent systematic review of patients aged ≥75 years undergoing meningioma surgery found that the rates of in-hospital mortality, worsening postoperative performance status, neurological impairment, and general complications varied widely across reports and were not necessarily associated with older age.³⁰⁾

When assessing the worsening of BI scores from admission to discharge, the rate of BI score worsening increased with increasing age from 7.1% to 21.1% (mean: 11.0%), which is comparable to the rate of increase reported in previous studies (8.3%-14.8%).^31-33) In our previous study,²⁹⁾ pneumonia and worsening BI scores were found to be significantly correlated with older age. The most common complications after meningioma surgery were new focal neuropathy and pneumonia.^31,33) Notably, these 2 complications are now considered common and unavoidable in older patients with meningioma.³¹⁾

Although older age can lead to postoperative functional decline and complications at discharge, the mFI-5 score could not predict complications and functional deterioration in patients aged <65 years, but it could in patients aged >65 years. Therefore, caution should be exercised when using the mFI-5 score to make decisions based on the age of patients.

mFI5 score and chronological age in UCA surgery

The rate of UCA rupture among Japanese is 2.8 times higher than that of Westerners.³⁴⁾ Old age is also reported to be a risk factor for UCA rupture.³⁵⁾ Therefore, many elderly individuals aged >65 years in Japan undergo surgical treatment for UCA. Both chronological and biological age have been implicated in postoperative mortality and in-hospital complications in patients with UCA, indicating that careful follow-up is an option in older patients (>65 years) (class IIa; level of evidence B); however, the usefulness of examining these factors by age group and the exact indications for surgery in older UCA patients are unknown.^36-39) The risk of treatment in patients with UCA is related to age and the presence of complications.

In our previous study of UCA surgery,⁴⁰⁾ the elderly group was associated with in-hospital mortality, and an mFI-5 score of ≥2 was not associated (Table 1). Probably because, unlike meningioma surgery, the overall in-hospital mortality was only 0.3% in UCA surgery. When examining in-hospital complications across all age groups, chronological age was not associated with complications in non- and pre-elderly groups; nonetheless, mFI-5 scores of 1 and ≥2 were significantly associated with complications (Table 2). The mFI-5 score was more useful than age as an associated factor for in-hospital complications after UCA treatment for patients aged <74 years. Considering both age and mFI-5 score may effectively optimize treatment interventions for patients with UCA.

Table 1.

Results of Multivariate Logistic Regression Analyses of Worsening BI Scores, In-Hospital Mortality, and In-Hospital Complications in All Cases (Cited From Reference 40)

Objective variables (no.)	Worsening BI scores (594/14,4407†)		In-hospital mortality (40/13,696†)		In-hospital complications (1,648/14,465†)
	OR (95% CI)	p-Value	OR (95% CI)	p-Value	OR (95% CI)	p-Value
Sex (male)	1.10 (0.92-1.33)	0.270	1.95 (1.00-3.78)	0.049*	1.06 (0.94-1.19)	0.375
Treatment (endovascular coiling)	0.41 (0.31-0.54)	<0.001*	2.21 (0.85-5.70)	0.102	0.52 (0.44-0.62)	<0.001*
Age group (y)
Non-elderly, <65 y	Reference		Reference		Reference
Pre-elderly, 65-74 y	1.84 (1.52-2.25)	<0.001*	1.88 (0.90-3.95)	0.094	1.21 (1.08-1.36)	0.001*
Elderly, ≥75 y	4.33 (3.47-5.39)	<0.001*	3.26 (1.38-7.71)	0.007*	1.76 (1.51-2.06)	<0.001*
Location of the aneurysms
ICA	Reference		Reference		Reference
ACoA	1.15 (0.89-1.48)	0.263	1.80 (0.76-4.28)	0.185	0.89 (0.75-1.05)	0.155
MCA	0.91 (0.73-1.14)	0.403	0.41 (0.12-1.36)	0.145	0.74 (0.64-0.86)	<0.001*
ACA	0.78 (0.51-1.19)	0.252	N/A		0.66 (0.50-0.88)	0.004*
BA	2.11 (1.54-2.90)	<0.001*	3.42 (1.37-8.53)	0.008*	1.43 (1.17-1.75)	<0.001*
VA	1.54 (0.98-2.43)	0.061	3.37 (1.05-10.83)	0.041*	1.50 (1.17-1.75)	0.002*
Other	3.76 (1.54-9.18)	0.004*	N/A		1.86 (0.94-3.60)	0.064
mFI-5 score
0 item	Reference		Reference		Reference
1 item	1.13 (0.95-1.36)	0.164	1.64 (0.82-3.28)	0.161	1.31 (1.17-1.47)	<0.001*
≥2 items	1.95 (1.52-2.51)	<0.001*	2.32 (0.86-6.26)	0.095	1.79 (1.49-2.15)	<0.001*
Internal oral medication on admission
Antiplatelet	1.67 (1.31-2.12)	<0.001*	0.55 (0.23-1.31)	0.177	2.91 (2.47-3.40)	<0.001*
Anticoagulant	2.01 (1.37-2.96)	<0.001*	1.55 (0.36-6.72)	0.557	1.84 (1.37-2.47)	<0.001*
Statin	1.24 (0.98-1.58)	0.072	0.64 (0.20-2.14)	0.478	1.22 (1.04-1.44)	0.013*
Hospital volume
1	Reference		Reference		Reference
2	0.89 (0.74-1.07)	0.218	1.34 (0.67-2.69)	0.410	0.88 (0.78-1.00)	0.043*
3	0.59 (0.48-0.73)	<0.001*	0.53 (0.21-1.31)	0.170	0.71 (0.63-0.81)	<0.001*

ACA: anterior cerebral artery; ACoA: anterior communicating artery; BA: basilar artery; BI: Barthel index; CI: confidence interval; ICA: internal carotid artery; MCA: middle cerebral artery; mFI-5: 5-factor modified frailty index; N/A: not applicable; no.: number; OR: odds ratio; VA: vertebral artery

*p<0.05.

†Complete cases without missing data.

Table 2.

Results of Multivariate Logistic Regression Analyses of In-Hospital Complications across All Age Groups (Cited From Reference 40)

Age group (no.)	In-hospital complications
	Non-elderly, <65 years (709/7,380†)		Pre-elderly, 65-74 years (638/5,364†)		Elderly, ≥75 years (301/1721†)
	OR (95% CI)	p-value	OR (95% CI)	p-value	OR (95% CI)	p-value
Sex (male)	1.12 (0.94-1.33)	0.190	1.06 (0.87-1.29)	0.545	0.85 (0.61-1.18)	0.331
Treatment (endovascular coiling)	0.66 (0.49-0.88)	0.005*	0.47 (0.36-0.62)	<0.001*	0.37 (0.26-0.56)	<0.001*
Age	1.00 (0.99-1.01)	0.910	1.00 (0.97-1.03)	0.924	1.08 (1.04-1.13)	<0.001*
Location of the aneurysms
ICA	Reference		Reference		Reference
ACoA	0.92 (0.71-1.18)	0.492	0.75 (0.58-0.98)	0.033*	1.30 (0.87-1.93)	0.196
MCA	0.77 (0.61-0.97)	0.025*	0.66 (0.52-0.83)	<0.001*	0.93 (0.65-1.33)	0.701
ACA	0.78 (0.50-1.21)	0.264	0.56 (0.36-1.93)	0.011*	0.64 (0.32-1.28)	0.205
BA	1.33 (0.97-1.83)	0.075	1.42 (1.04-1.93)	0.028*	1.88 (1.13-3.12)	0.015*
VA	1.45 (1.04-2.02)	0.028*	1.60 (1.01-2.52)	0.045*	1.39 (0.61-3.18)	0.432
Other	2.64 (1.21-5.74)	0.014*	0.99 (0.21-4.59)	0.986	0.96 (0.11-8.64)	0.968
mFI-5 score
0 item	Reference		Reference		Reference
1 item	1.35 (1.14-1.60)	0.001*	1.28 (1.07-1.54)	0.008*	1.32 (0.98-1.77)	0.068
≥2 items	1.66 (1.20-2.28)	0.002*	1.72 (1.30-2.29)	<0.001*	1.99 (1.36-2.93)	<0.001*
Internal oral medication on admission
Antiplatelet	2.52 (1.91-3.32)	<0.001*	3.28 (2.57-4.18)	<0.001*	2.81 (1.97-4.00)	<0.001*
Anticoagulant	1.95 (1.10-3.45)	0.022*	1.69 (1.08-2.63)	0.021*	1.96 (1.12-3.41)	0.018*
Statin	1.35 (1.04-1.74)	0.024*	1.02 (0.79-1.32)	0.852*	1.57 (1.12-2.22)	0.010*
Hospital volume
1	Reference		Reference		Reference
2	0.87 (0.72-1.06)	0.164	0.87 (0.71-1.07)	0.190	0.89 (0.65-1.22)	0.430
3	0.75 (0.62-0.91)	0.004*	0.74 (0.60-0.91)	0.004*	0.58 (0.42-0.82)	0.002*

ACA: anterior cerebral artery; ACoA: anterior communicating artery; BA: basilar artery; CI: confidence interval; ICA: internal carotid artery; MCA: middle cerebral artery; mFI-5: 5-factor modified frailty index; no.: number; OR: odds ratio; VA: vertebral artery

*p < 0.05.

†Complete cases without missing data.

Patient age is a poor prognostic factor in the treatment of UCA.^41,42) This is because Japan is one of the most aged countries globally, and the world's aging population is expected to follow a similar pattern. Consequently, the results of our study may become a global standard.

In neurosurgery, mFI-5 scores have been reported to be associated with postoperative outcomes in patients with pituitary adenomas, meningiomas, gliomas, and UCA.^26,27,43-45) To date, only a few studies have examined patients in different age groups using the national UCA database.^45,46) An analysis of national data by each age group (including older adults) revealed that the mFI-5 score is a potentially relevant factor in postoperative outcomes for patients with meningiomas and UCAs, regardless of age group. Consequently, mFI-5 scores and treatment should be integrated into the decision-making process for patients undergoing surgery for meningiomas and UCAs, regardless of age. Recently, there have been numerous reports on the efficacy of mFI-5 in trauma and spinal surgery.^47-51) As the global population aging is expected to accelerate, other countries are expected to follow Japan's trajectory. Given the increasing proportion of older adults, it will become increasingly important to evaluate long-term outcomes and healthy life expectancy in the elderly. Moreover, the effectiveness of preoperative frailty-targeted interventions is expected to gain wider application in this population.

Visualization and validation of non-linear correlation between age and outcome in patients with subarachnoid hemorrhage

The case fatality rate for aSAH is reported to be approximately 35% and remains high in older individuals.^52-54) Older age and poor initial neurological status have been reported to have an unfavorable impact on the outcome of patients with aSAH treated by surgical clipping or endovascular coiling procedures.^55-57) The clinical reference age is 70 or 75 years, depending on the reports.^58-61) Our analysis of the outcomes of subarachnoid hemorrhage in Japan using the Diagnosis Procedure Combination database found that the good outcome group on the modified Rankin Scale (mRS) of 0-2 was 67.5% for surgical clipping and 66.5% for endovascular coiling. The mortality rates were 6.3% and 10.1% for surgical clipping and endovascular coiling, respectively.⁵⁷⁾ In patients aged ≥65 years, good outcomes were 38.3% and 37.4%, and mortality rates were 10.1% and 14.2% for surgical clipping and endovascular coiling, respectively.⁵⁷⁾ The risk factors for poor outcomes were age and male sex. Multivariate analysis of risk factors for poor outcome by age group <65 years showed that chronological age, men, neurologic findings on admission, diabetes mellitus, anticoagulants, hypertension, high-volume hospital, and antiplatelet agents were inversely associated. For patients aged >65 years, chronological age, being male, surgical clipping, neurological symptoms on admission, diabetes mellitus, and anticoagulants were identified as risk factors for poor outcomes, whereas the location of the aneurysm, high-volume hospitals, hypertension, antiplatelet agents, and statins were inversely associated.⁵⁷⁾

In the Stroke Data Bank in Japan, the change over time in subarachnoid hemorrhage outcomes improved regarding discharge outcomes for women after adjusting for age, and an improvement was observed for both men and women with respect to poor outcomes and death.⁶²⁾

Predictive models for aSAH outcomes have been published, but to our knowledge, no visual tools are available,^63-66) to aid clinical decision-making and inform patients and their families by plotting the non-linear correlation between age and poor outcome. Although the correlation between age and outcome is known to be non-linear in some clinical areas,^67,68) no detailed analysis and no visualization of such non-linear correlations in aSAH have been reported. Therefore, using a nationwide clinical registry in Japan,⁶⁹⁾ we represented the non-linear correlation between age and poor outcome at discharge visually in patients undergoing radical treatment for aSAH while adjusting for other covariates and addressing heterogeneity due to the initial neurological status of patients.⁷⁰⁾ An analysis of risk factors was conducted for poor outcomes using random forests, a technique within the realm of machine learning. Our findings revealed that the World Federation of Neurosurgical Societies (WFNS) grade, the neurological severity at admission, was the highest, at 1, followed by age, at 0.57, and the other factors at 0.2 or less. Conventional logistic multivariate analysis using odds ratios in 10-year increments could not adequately display the results. To address this, generalized additive models (GAMs) were used to depict the results using the modified conditional plot method (Fig. 4) and the partial dependence plot method, which were largely consistent. The graphs clearly showed the non-linearity between age and outcome for WFNS grades I and II surgical clipping and WFNS grades I, II, and III endovascular coiling. The outcome remained relatively constant up to age 50 years, increased almost linearly from age 50-70 years, and increased exponentially for individuals aged ≥70 years. Although WFNS grades IV and V are often classified together as severe cases, the gradient differed with age, both increasing linearly with increasing age from younger patients and plateauing somewhat at age ≥70. This study has not been validated and is currently undergoing validation in different databases.

Fig. 4.

(A, B) Modified conditional plot, based on a modified Rankin Scale score ≥3 at discharge, according to the World Federation of Neurological Societies (WFNS) grade on admission (I-V) and treatment (adjusted for other covariates). (A) Surgical clipping, adjusted; (B) Endovascular coiling, adjusted. Shaded areas represent 95% confidence interval (cited from ref. 70).

In this study, a GAM was employed in a subgroup of patients with different WFNS grades and treatment approaches without considering age classification. To visualize the GAM, 2 methods were employed: modified condition plots and partial dependence plots. The modified condition plots have the advantage of displaying the 95% confidence interval of the predictive model but do not adequately reflect the distribution of covariates in the study population. Conversely, the partial dependence plots method illustrates the correlation between age and poor outcome, considering the distribution of covariates in the study population, but does not show the reliability of prediction. Thus, the 2 methods serve as complementary tools to evaluate clinical risk in the context of correlations with key independent variables.

Graphs such as those presented in this document could effectively convey crucial information, such as regarding the operating room, to those without specialized knowledge. Patients with subarachnoid hemorrhage and their relatives require reliable information about the prognosis of individuals with aSAH to participate effectively in joint decision-making. Otherwise, clinical decisions may be at the discretion of individual clinicians, which varies from person to person.⁷¹⁾ Considering that reliable, objective prediction of subarachnoid hemorrhage outcomes will reduce patient transfers between expensive intensive care units and inexpensive wards and facilitate more effective allocation of limited resources, our results will aid clinical decision-making and guide inexperienced practitioners.⁷²⁾ This study visually depicts a non-linear correlation between age and poor prognosis in patients with aSAH, addresses the influence of other covariates, and demonstrates that this correlation is heterogeneous across disease severity and treatment modalities.

A machine learning-based prediction model was developed and validated to assess the non-linear association between advanced age and clinical outcomes in patients with aSAH.⁷³⁾ Patient data were obtained from the Japanese Stroke Databank (derivation cohort, n = 9,657) and the Predict for Outcome Study of aSAH (validation cohort, n = 5,085), including those treated with surgical clipping or endovascular coiling between 2003 and 2019. GAMs were employed to predict poor outcomes (defined as mRS score ≥3 at discharge), incorporating spline-based age transformations for each WFNS grade. Model performance was evaluated using discrimination metrics and calibration plots in the validation cohort. The derivation and validation cohorts included 3,610 and 3,251 patients, respectively. In the unadjusted models, the areas under the receiver operating characteristic curves (AUCs) were 0.835 and 0.827 for the derivation and validation cohorts, respectively. In the adjusted models, AUCs were 0.844 and 0.836, respectively. An unbiased correlation was confirmed between predicted and observed probabilities of poor outcomes (Fig. 5). GAMs provided an interpretable framework for visualizing the non-linear association between age and clinical outcomes. This model may support quantitative decision-making across a broad spectrum of clinical scenarios.

Fig. 5.

(A, B). Calibration plots of the derivation and validation cohorts in the unadjusted (A) and adjusted (B) model. Plots for examining prediction bias are shown. The green curve indicates calibration in the derivation cohort, while the purple curve indicates calibration in the validation cohort. If the model is free of prediction bias, a calibration curve will lie on the diagonal (y = x). However, if the prediction is biased upward or downward, the curve will deviate from the diagonal accordingly (cited from ref. 73).

This review has several limitations. First, all included studies were retrospective observational in design, and therefore, the influence of confounding factors could not be entirely eliminated. Second, post-discharge outcomes could not be assessed, as the database used only contained in-hospital data. Third, the datasets utilized in the reviewed studies were relatively outdated, potentially limiting the generalizability of the findings to current clinical practice.

Conclusions

Progress has been made in predicting outcomes for older patients undergoing neurosurgery in Japan. Clinical research involving older individuals is particularly relevant to Japan and is an area in which Japan can demonstrate its originality. In the future, more individualized and specific predictions will be required.

Author Contributions

All authors have made substantial contributions to the intellectual content of the paper, have approved the final manuscript, and agree with submission to this journal. Fusao Ikawa is the corresponding author for this study and the principal investigator. He takes responsibility for data management, accuracy of statistical analysis, conducting the research, and drafting the manuscript.

Data Availability Statement

The anonymized data for this study were shared with the corresponding author by any qualified investigator after request. Primary data from the Diagnosis Procedure Combination database and the Japanese Stroke Databank (JSDB) were made available after reasonable request, in accordance with each institutional review board.

Conflicts of Interest Disclosure

All authors have no conflict of interest.

Funding Statement

This work was supported by Grant-in-Aid for Scientific Research (C) 23K08521 from the Japan Society for the Promotion of Science.