What Is the Difference in the Risk of Suicide Death Between Spine Fracture in Patients Older Than 65 Years and Matched Controls? A Large-database Study from South Korea

Abstract

Background

Pain and kyphotic deformity after spinal fractures can result in a decrease in a patient’s physical function and quality of life. Furthermore, physical illness, such as respiratory compromise, or mental illness, including depression, may be exacerbated by a spinal fracture. Complications caused by spinal fractures and old age are risk factors for suicide, but studies on these patients are rare.

Questions/purposes

(1) What is the incidence rate of death by suicide after a spinal fracture in patients older than 65 years? (2) How much does the risk of death by suicide increase in patients older than 65 years who have spine fractures compared with well-matched controls? (3) How does this risk change as a function of increasing time after injury?

Methods

Spinal fractures in patients older than 65 years and matched controls were selected from the National Health Insurance Service-Senior cohort (NHIS-Senior) of South Korea. The NHIS-Senior consists of 558,147 people selected by 10% simple random sampling method from a total of 5.5 million people 60 and older in 2002; all people were followed through 2015. A total of 31,357 patients with spine fractures and their 62,714 matched controls remained in the study. The mean follow-up time was 4.3 ± 3.0 years (135,229 person-years) in the spine fracture group and 4.6 ± 3.0 years (290,096 person-years) in the matched control group. We matched the groups for demographic factors such as age, gender, Charlson Comorbidity Index score, medication history, medical history, preoperative disability, number of hospital admissions, as well as socioeconomic factors such as household income level, residential district, and type of national health insurance using a 1:2 risk set propensity score matching by a nearest-neighbor matching algorithm with a maximum caliber of 0.1 of the hazard components. The incidence rate of suicide and the 95% confidence interval were calculated based on a generalized linear model with a Poisson distribution. The effect size was presented as a hazard ratio (HR) using Cox’s proportional hazard model with robust variance estimator that accounts for clustering within matched pairs.

Results

The overall risk of death by suicide throughout the surveillance period, expressed as an incidence rate, was 116 per 100,000 person-years in spinal fracture (157 deaths by suicide over 135,229 person-years). Throughout the entire surveillance period, the risk of death by suicide was greater among patients with spinal fractures than it was in the control group (HR 1.8 [95% CI 1.5 to 2.2]; p < 0.01). This difference was greatest in the first 365 days after the fracture (HR 2.5 [95% CI 1.6 to 3.8]; p < 0.01) (45 deaths by suicide, incidence rate: 156 per 100,000 person-years in spinal fracture). The risk of suicide death in patients with spine fracture from 365 days to the last follow-up was also higher than that of matched controls (HR 1.6 [95% CI 1.3 to 2.1]; p < 0.01).

Conclusions

Considering the substantially increased risk of death by suicide in patients with spine fractures who are older than 65 years, surgeons should consider offering psychiatric evaluation and management more frequently, particularly in patients with chronic pain, functional disability, and depressive mood. Future studies should investigate the underlying causes of suicide, such as deteriorating socioeconomic support or depression, and whether early initiation of psychological support after injury can reduce the suicide rate.

Level of Evidence

Level III, prognostic study.

Introduction

Spinal fractures are the most common type of osteoporotic fracture [13]. Spinal fractures occur mainly in the thoracolumbar junction or midthoracic area and cause pain, kyphotic deformity, and increased risk of recurrent spine fractures [4, 22]. Pain and kyphotic deformity decrease a patient’s physical function and quality of life [5]. Furthermore, physical illnesses (such as respiratory compromise) and or mental illness (such as depression) may be exacerbated by a spinal fracture [22]. Previous reports on the relationship between spine fracture and suicide were limited to the type of spine fracture that occurred during suicide attempts [12, 16, 23]. Complications caused by spinal fractures, such as chronic pain and older age, are risk factors for suicide [9, 25]; however, previous reports about spine fracture as a risk factor for suicide death are rare [5, 10].

Suicide in older patients is not as well studied as it is in younger patients, but it represents an important public health problem in many developed countries [9]. In European countries, the suicide rate for people older than 65 years is 29.3 per 100,000 person-years and the suicide attempt rate is 61.4 per 100,000 person-years [28]. There have been two reports on the relationship between spinal fracture and suicide risk. Chang et al. [5] reported an adjusted odds ratio of 1.53 for suicide risk in all patients with spinal fractures older than 40 years. However, this study did not analyze the effect on risk of suicide death by spine fracture in older patients. Erlangsen et al. [10] analyzed the suicide risk for patients with physical illnesses in Demark who were older than 65 years, and they reported a 1.31-fold increase in relative risk of suicide within 3 years after diagnosis of a spinal fracture, but this study did not calculate the specific incidence rate of suicide for patients older than 65 years with spinal fractures, and they did not analyze the change of risk of suicide death over time.

We therefore asked (1) What is the incidence rate of death by suicide after a spinal fracture in patients older than 65 years? (2) How much does the risk of death by suicide increase in patients older than 65 years who have spine fractures compared with well-matched controls? (3) How does this risk change as a function of increasing time after injury?

Patients and Methods

The National Health Insurance Service (NHIS) of South Korea established the National Health Information Database (NHID), which provides a claims archive for research purposes and stores all records of healthcare and long-term care services [6, 15]. The NHIS maintains all personal information, demographics, and medical treatment data for the entire South Korean population [17, 31]. From the NHID, the NHIS constructed and provided researchers with the NHIS-Senior cohort of South Korea, which contains representative administrative data for health policy and biomedical research purposes. The NHIS-Senior cohort comprises 558,147 people selected in 2002 by a 10% simple random sampling method from a total of 5.5 million participants 60 years and older [15]. Under a compulsory social insurance system implemented under the South Korean National Health Insurance Act, all participants were followed until 2015, except for instances of death or emigration [15, 17]. The key variables in the NHIS-Senior cohort included all inpatient and outpatient medical claims data, such as codes for treatment procedure, prescriptions, and diagnoses [17, 31]. A detailed profile of this cohort has been published elsewhere [15]. For our study, spinal fractures in patients older than 65 years and their matched controls were selected from the NHIS-Senior cohort. The spinal fracture patients were designated as the incident spinal fracture cohort. The design and protocol of this study were approved by the institutional review board at our hospital (EMC-IRB No. 2018-12-009). Written informed consent was waived for all patients involved in this study.

This study is based on data from the Korean National Health Insurance Service (research administration number, NHIS-2019-2-177), and the results of the study are not related to the National Health Insurance Service.

Incident Spinal Fracture Cohort

Based on a previous study [14], the inclusion criteria for the incident spinal fracture cohort included the following diagnostic codes as defined by the ICD-10: S22.0 (fracture of the thoracic spine), S22.1 (multiple fractures of the thoracic spine), S32.0 (fracture of the lumbar spine), M48.4 (fatigue fracture of a vertebra) and M48.5 (collapsed vertebra, not classified elsewhere). The patients were either first-time admissions to an acute care hospital (index admission) or patients who had undergone vertebroplasty or kyphoplasty during the follow-up period (2002 to 2015).

Among the patients who met the inclusion criteria, we excluded certain patients to maintain the validity of the study design. Patients with spinal fractures before January 1, 2005 were excluded to ensure a minimum of a 3-year fracture-free period. Patients under the Medical Aid program were excluded due to possible incomplete information. The incidence date (time-zero) of spinal fracture was defined as the date of admission to the acute care hospital or date of the outpatient visit that fulfilled the inclusion criteria.

The number of spine fracture in patients older than 65 years who met the inclusion criteria was 45,776 (from January 1, 2002 to December 31, 2015). Among these participants, we excluded 4382 patients with a spinal fracture during the first 3 years (2002 to 2004) and 4818 enrollees of the Medical Aid program. During risk-set matching, 123 patients with spinal fracture were not matched to control persons, and 4965 fracture patients entered into the study as control persons for another spine-fracture patient. In addition, 131 patients with spine fractures were additionally excluded due to less than a 1:2 matching ratio.

Therefore, a total of 31,357 incident patients with spine fractures and their 62,714 matched controls remained in the study. The mean follow-up time was 4.3 ± 3.0 years (135,229 person-years) in the spine fracture group and 4.6 ± 3.0 years (290,096 person-years) in the matched control set. In each group, the mean age was 71.6 ± 6.1 years and 79% of the patients were women (Table 1). The comparison of cumulative incidence curve between spinal fracture group and matched controls during the entire follow-up period showed a difference (p for stratified log-rank test < 0.01) (Fig. 1).

Table 1.

Baseline characteristics of the sample

Characteristics	Spine fracture (n = 31,357)	Matched controls (n = 62,714)	Standardized difference
Age (years)a	71.6 ± 6.1	71.6 ± 6.1	0
Womenb	79% (24,691)	79% (49,382)	0
Household income level (decile)b			0.02
First	11% (3470)	11% (6992)
Second	5% (1703)	5% (3399)
Third	7% (2320)	7% (4605)
Fourth	7% (2179)	7% (4195)
Fifth	6% (2025)	6% (3925)
Sixth	7% (2260)	7% (4519)
Seventh	9% (2806)	9% (5609)
Eighth	12% (3804)	12% (7599)
Ninth	16% (5085)	17% (10,479)
Tenth	18% (5705)	18% (11,392)
Type of national health insuranceb			0.02
Self-employed	20% (6384)	20% (12,821)
Dependents of self-employed	18% (5761)	18% (11,190)
Employee	2% (646)	2% (1124)
Dependents of employee	59% (18,566)	60% (37,579)
Registered disabilityb	0.6% (174)	0.5% (309)	0.01
Residential districtb			0.04
Seoul	14% (4441)	14% (8585)
Pusan	7% (2121)	7% (4156)
Daegu	4% (1378)	4% (2556)
Incheon	4% (1322)	4% (2456)
Gwangju	2% (762)	2% (1475)
Daejeon	2% (713)	2% (1421)
Gyeonggi-do	17% (5176)	16% (10,288)
Gangwon-do	4% (1245)	4% (2467)
Chungcheongbuk-do	4% (1363)	4% (2706)
Chungcheongnam-do	7% (2101)	7% (4189)
Jeollabuk-do	6% (2022)	7% (4135)
Jeollanam-do	8% (2483)	8% (5139)
Gyeongsangbuk-do	9% (2831)	10% (6025)
Gyeongsangnam-do	10% (3142)	11% (6637)
Jeju-do	1% (257)	1% (479)
Charlson Comorbidity Index score b			0.05
0	46% (14,305)	48% (29,917)
1	29% (9047)	28% (17,805)
2	14% (4431)	13% (8370)
3	6% (2034)	6% (3732)
4	3% (861)	3% (1576)
5 or more	2% (679)	2% (1314)
Number of hospital admissionsb			0.04
0	69% (21,611)	70% (44,199)
1	18% (5652)	17% (10,885)
2 or more	13% (4094)	12% (7630)
Medication historyb
Antihypertensive agents	38% (11,976)	36% (22,428)	0.05
Antidiabetic agents	10% (3221)	9% (5643)	0.04
Lipid lowering agents	8% (2434)	7% (4237)	0.04
Benzodiazepine	23% (7364)	22% (13,920)	0.03
Opioid (prescription days)			0.03
0 to 5 days	20% (6153)	19% (11,992)
> 5 to 12 days	18% (5732)	18% (11,533)
> 12 to 28 days	21% (6486)	21% (13,182)
> 28 days	24% (7682)	24% (14,888)
Medical historyb
Malignant neoplasm	3% (1035)	3% (2020)	0.004
Ischemic heart disease	9% (2950)	8% (5372)	0.03
Stroke	5% (1634)	5% (2918)	0.03
Dementia	1% (240)	1% (417)	0.01
Depression			0.02
Diagnosis and medication	2% (754)	2% (1412)
Diagnosis only	2% (652)	2% (1180)
Medication only	4% (1279)	4% (2393)

^aData are presented as mean ± SD.

^b% (n).

At the time of spine fracture of each patient, two controls were matched on propensity score estimated by Cox proportional hazard model with predictors included in this table.

Fig. 1.

Cumulative incidence curves show the comparison of the of suicide between patients older than 65 years with spine fracture and their risk set-matched controls (p < 0.001 for the entire period; log-rank test). Cumulative incidences were calculated by product limit (Kaplan-Meier) methods. A color image accompanies the online version of this article.

Identification of Patients who Committed Suicide

As South Korean law requires, all death certificates must be reported to Statistics Korea. We collected information on deaths (date and cause) from this agency and individually linked that information to the patients included in our study using unique personal identification numbers [15, 17]. Death by suicide was identified as the causes of death categorized under “Intentional Self-Harm” (ICD-10 codes X60 to X84) [15, 17]. The time of the event was defined as the date of death by suicide.

During the whole follow-up period, a total of 342 suicide deaths were identified (185 patients in matched control, 157 patients in spine fracture).

Risk-Set Matching on Propensity Score

Although the NHIS-Senior cohort was constructed retrospectively and spinal fracture is not a possible intervention of a randomized controlled trial, our study’s design emulated that of a prospective matched-cohort study. After calculating the time-dependent propensity score, we performed risk-set matching [19, 21]. To adjust for confounding effects, we examined the association between spinal fracture and suicide risk using time-dependent propensity score matching [21]. Hazard ratios (HR) adjusted by propensity scores were estimated using Cox’s proportional-hazards regression model with January 1, 2005 as the baseline and spinal fracture as an event. Age, square of age (used when matching the propensity score to assess a nonlinear response), gender, household income level, type of national health insurance, registered disability, residential district, Charlson Comorbidity Index score (CCI), number of hospital admissions, medication history (antihypertensive, antidiabetic, lipid-lowering drugs, benzodiazepine, and opioid), and medical history (malignant neoplasm, ischemic heart disease, stroke, depression, and dementia) were included as independent variables (Table 1). All these variables were identified in the 3 years (2002 to 2004) before the baseline (January 1, 2005). Age and the square of age were included as continuous variables. The categorical variables were gender, household income levels, type of National Health Insurance, registered disability, residential district as a fixed effect, CCI, number of hospital admissions, medication history, and medical history. The number of comorbidities for each participant was assessed by diagnostic codes using the Quan et al. [24] ICD-10 coding algorithm for the CCI score. Patients who had a record of prescriptions (> 90 days) for antihypertensive, antidiabetic, lipid-lowering, benzodiazepine, and/or antidepressant drugs were considered to be taking corresponding medications. Medication histories of opioid use were divided into 5 groups according to number of prescription days (none, 0 to 5 days, > 5 to 12 days, > 12 to 28 days, > 28 days). Medical histories of depression were divided into four group according to diagnosis and medication (none, diagnosis only, medication only, both diagnosis and medication).

Based on the incident date (time zero), spinal fracture patients were matched for same age and gender with people who were at risk of developing a spine fracture. After constructing the first risk-set, this method of risk-set matching was repeated sequentially until most spinal fracture patients were matched [1, 19, 33]. A 1:2 matching based on the propensity score was then sequentially performed for each risk-set using a nearest-neighbor matching algorithm with a maximum caliper of 0.01 for estimating the HR. To ensure the matching was independent of future events, the matched controls could be either those who never developed or who had not yet developed spinal fractures. Therefore, patients who were in the incident spinal fracture cohort could also act as a matched control for other spinal fracture patients whose incident date was before the control patient [36]. Next, to yield nonoverlapping samples from the risk set, the matched persons were removed from the next risk sets. The matching process was repeated with the next risk set and terminated when there were no more spinal fracture patients.

Statistical Analyses

Statistical tests for the association between incident spinal fracture and suicide risk were performed after considering the nature of the matched pair analyses. To assess covariate balance between treatment groups, baseline characteristics were compared by standardized differences where a difference less than 0.1 (10%) was generally considered negligible [2, 27]. The effect of spinal fracture on suicide risk was deemed to be relatively acute in this study. Therefore, several time frames (before 365 days, after 365 days, and whole follow-up days) were used as follow-up periods for the survival analyses. Cumulative incidence curves for suicide were obtained using the Kaplan-Meier method and comparison of the curves was conducted by the stratified log-rank test [3]. The cumulative incidence of suicide and 95% confidence intervals were estimated using the product limit (Kaplan-Meier) estimator method for survival probability. To manage the overdispersion problem, the incidence rate of suicide and 95% CIs were calculated based on a generalized estimating equation model with a Poisson distribution and robust standard error. The incidence rate was expressed as the number of suicides per 100,000 person-years. The effect size was presented as an HR using Cox’s proportional hazard model with a robust variance estimator that accounted for clustering within matched pairs [3, 27]. Proportional hazard assumption was assessed by graphically using a log of negative log of estimated survivor function, time-dependent explanatory variable, Schoenfeld residuals, cumulative sums of martingale residuals, and a supremum test for proportional hazard assumption [32]. The incident date was set as the date of spinal fracture for both patients with spine fractures and their matched controls. Survival time was defined as the number of days from the incidence date to the dates of death or December 31, 2015, whichever occurred first. In addition to above cause-specific model, which deaths other than suicide were treated as censored observation, a Fine and Gray subdistribution hazard model was used for a sensitivity analysis [35]. Statistical analyses were conducted using SAS Enterprise Guide version 7.1 (SAS Institute, Cary, NC, USA). Means were presented with SDs. P values of < 0.05 were considered statistically significant.

Results

The overall incidence rate of death by suicide in patients with spine fracture throughout the surveillance period was 116 per 100,000 person-years (157 deaths by suicide over 135,229 person-years) (Table 2).

Table 2.

Cumulative incidence of suicide according to cumulative time frames

	Cumulative incidence (%) of suicide (95% CI)
Time frame	Hip fracture patients	Matched controls
90 days	0.07 (0.04 to 0.10)	0.02 (0.01 to 0.03)
180 days	0.11 (0.08 to 0.15)	0.03 (0.02 to 0.05)
1 year	0.15 (0.11 to 0.20)	0.06 (0.05 to 0.09)
2 years	0.27 (0.21 to 0.34)	0.14 (0.11 to 0.18)
3 years	0.38 (0.31 to 0.46)	0.22 (0.18 to 0.26)
4 years	0.46 (0.38 to 0.55)	0.27 (0.22 to 0.32)
5 years	0.55 (0.45 to 0.66)	0.33 (0.28 to 0.38)
6 years	0.70 (0.59 to 0.84)	0.41 (0.34 to 0.48)
7 years	0.83 (0.69 to 0.99)	0.43 (0.36 to 0.50)
8 years	0.87 (0.72 to 1.03)	0.49 (0.42 to 0.58)
9 years	0.94 (0.78 to 1.14)	0.55 (0.46 to 0.66)
10 years	1.08 (0.86 to 1.34)	0.58 (0.48to 0.69)

Table 3.

Incidence rates of suicide in incident spine fracture cohort and matched control showing the association between spine fracture and the suicide risk in elderly patients

Exposure	Number of subjects	Number of suicides	Person-years	Incidence rate (95% CI) per 100,000 person-years	Hazard ratio (95% CI)	p value
Whole follow-up period
Matched control	62,714	185	290,096	64 (55 to 74)	1
Spine fracture	31,357	157	135,229	116 (99 to 136)	1.8 (1.5 to 2.2)	< 0.001
0 to 1 year
Matched control	62,714	37	58,743	63 (46 to 87)	1
Spine fracture	31,357	45	28,775	156 (116 to 209)	2.5 (1.6 to 3.8)	< 0.001
> 1 year
Matched control	55,022	148	231,353	64 (54 to 75)	1
Spine fracture	26,651	112	106,453	105 (87 to 127)	1.6 (1.3 to 2.1)	< 0.001

Throughout the entire surveillance period, the risk of death by suicide was greater among patients with spinal fractures than it was in the control group, which was matched for demographic factors and socioeconomic factors (HR 1.8 [95% CI 1.5 to 2.2]; p < 0.001).

The increased risk of death by suicide among patients with spinal fractures (compared with the matched control group) was greatest in the first 365 days after the fracture event (HR 2.5 [95% CI 1.6 to 3.8]; p < 0.001). The incidence rate of death by suicide in patients with spine fracture in the first 365 days after the fracture was 156 per 100,000 person-years (45 deaths by suicide over 28,775 person-years). From 365 days to the last follow-up, 112 suicides were identified during 106,453 person-years (incidence rate: 105 per 100,000 person-years) among 26,651 patients with spine fractures. The risk of suicide death in patients with spine fracture in this period was reduced compared with that in the first 365 days after fracture, but it was still higher than risk of suicide death in matched controls (HR 1.6 [95% CI 1.3 to 2.1]; p < 0.001).

Results of competing risk models were similar to that of the cause-specific model.

Discussion

Complications caused by spinal fractures, such as chronic pain and older age, are risk factors for suicide [9, 25]; however, previous reports about spine fracture as a risk factor for suicide death are rare [5, 10], and those studies either did not look at the risk of suicide specifically in older patients (who may be at highest risk) [5] or not calculate the specific incidence rate of suicide for patients over the age of 65 years with spinal fractures, and they did not analyze the change of risk of suicide death over time [10]. We found that in patients older than 65 years, spinal fracture was a risk factor for suicide and the suicide risk increased substantially compared with that in matched control groups. Second, the incidence rate of suicide within 1 year after a spinal fracture in this group of patients was 156 per 100,000 person-years, and then decreased to 105 per 100,000 person-years from 1 year to last follow-up. Third, the HR for suicide in the spinal fracture group increased by 2.5 times compared with the matched control group within 1 year after the fracture and then decreased to 1.6 times from 1 year to last follow-up.

There are several limitations in our study. First, because of the limited information available to us, residual confounding such as severity of paralysis and previous history of surgery could explain some of the associations observed in this study. However, we controlled for major confounders, including depression, opioid use, dementia, benzodiazepine use, and the CCI. Second, the inclusion criteria for spinal fracture in patients older than 65 years were not representative of all spinal fractures that may occur. Inclusion criteria were defined according to first-time hospital admission for spinal fracture or certain surgical procedures (vertebroplasty or kyphoplasty). Patients with asymptomatic spinal fractures or those with mild fractures who were not admitted to a hospital were excluded. Thus, the suicide risk presented in this study represented patients with severe spinal fractures. Third, as this study used only the NHIS-Senior cohort database of South Korea, the cultural characteristics of Korea may have been reflected in the results. However, this study compared the relative risk of suicide between patients with spinal fractures and matched controls. We believe that complications and characteristics of spine fracture, including repeating nature, chronic pain, kyphotic deformation, and disability, are similar in any society, regardless of race or sociocultural differences. Thus, we believe that the results of this study have generalizability and may be applied to other societies. Nevertheless, we believe that additional studies are needed to confirm or refute our findings; ideally, these would be done in other cultures, or compare patients across different cultures, to try to address these potential confounding variables. Lastly, the spine fracture group in our study consisted of healthier patients considering the distribution of CCI scores. Because of this, there may be limitations in applying our results to patients older than 65 years who have many comorbidities. However, we believe that patients with poor health have a higher death rate by suicide compared with our study results because of the higher risk of suicide due to underlying comorbidity and physical disease.

The overall risk of death by suicide in the group of patients with spinal fractures in this study, at 116 per 100,000 person-years (157 deaths by suicide over 31,357, 135,229 person-years), is high. To the best of our knowledge, there are no studies reporting incidence rate of suicide death in older patients with spine fracture. Thus, the incidence rate of suicide suggested in this study can be compared with other severe or disabling diseases reported to be associated with increased risk of suicide compared with the general population [8, 26, 34, 37]. Stroke can cause cognitive impairment as well as physical dysfunction, and these complications can contribute to anxiety, sleep disorders, and depression. Stroke and associated complications lead to a high death rate by suicide, with a range from 30 to 272.3 per 100,000 person-years according to severity of disability, socioeconomic status, accompanying psychotic disorder, and sociocultural differences in various countries [8]. Patients with cardiac disease have been reported to have a higher suicide rate versus the general population. According to a study by Wu et al. [34], suicide rates were 59.6 and 44.6 per 100,000 person-years after congestive heart failure and acute myocardial infarction, respectively. Suicide rates for cancer patients have also been reported. The suicide death rate for lung cancer patients in the United States was reported as 27.5 per 100,000 person-years [26]. In a Danish study, the suicide rate of cancer patients was reported to be 6.1 to 30.2 per 100,000 person-years [37]. The suicide death rate of cancer survivors in South Korea was 88.7 per 100,000 person-years [7]. The suicide rate suggested in this study for older patients with spinal fractures appears high compared with that of other serious diseases.

Throughout the entire surveillance period, the risk of death by suicide was substantially greater among patients with spinal fractures than it was in the control group. This is consistent with previous studies. Chang et al. [5] analyzed risk of suicide by several fracture types in patients older than 40 years using Taiwan’s National Health Insurance database between 2000 and 2001. They found spinal fractures were associated with a high risk of suicide (adjusted odds ratio 1.53 [95% CI 1.43 to 1.64]; p < 0.001). Erlangsen et al. [10] used the Danish national administrative register to analyze whether physical diseases increase suicide risk in patients older than 65 years of age. They reported an increasing suicide risk within 3 years after spinal fracture diagnosis (adjusted rate ratio 1.82 [95% CI 1.34 to 2.48]; p < 0.001). Suicide in older patients is caused by the interaction of various factors [9], in particular, depression and suicide risk increase in the presence of physical illness, cognitive impairment, and more insecure socioeconomic conditions [28, 29]. Spinal fractures in older patients cause chronic pain, kyphotic deformation, and a reduction in mobility and the ability to care for oneself [5, 30]. Furthermore, a study by Leslie et al. [18] reported that a gradual increase in health care utilization was observed up to 5 years after the occurrence of a vertebral fracture, which may become an economic burden for patients and families. We believe that these complicating factors increase the risk of suicide death in spinal fracture with older patients.

In our study, the increased risk of death by suicide among patients with spinal fractures compared with the matched control group was greatest in the first year after injury. We also found that the risk of suicide death in patients with spine fracture from 1 year to last follow-up was reduced compared with that in the first 365 days after fracture, but it was still higher than risk of suicide death in matched controls. This is inconsistent some previous research. For example, Erlangsen et al. [10] found that the suicide risk increased within 3 years after spinal fracture, but they did not confirm that the suicide risk increased during the whole surveillance period. This is probably due to the high mortality rate observed in older patients, which has not achieved statistical significance between suicide and spinal fracture. However, in our study, we found that even in the competing risk model, spine fracture increased suicide risk over the entire follow-up period. It is also believed that the timing of the increase in suicide risk is observed in different ways depending on injury or disease. Liu et al. [20] reported that the highest risk of suicide occurred in the first 6 months after a diagnosis of acute coronary syndrome and this risk remained for 4 years postdiagnosis. Wu et al. [34] reported that there were differences in risk of death by suicide depending on the type of cardiovascular disease, but the highest suicide death risk was observed within 2 years after diagnosis. According to a cancer registry study in Lithuania, Kaceniene et al. [11] reported that the risk for suicide was highest in the initial 3 months after a cancer diagnosis. Although this risk decreased within a year, it increased again after cancer recurrence. In our study, the suicide risk in patients with spinal fractures decreased 1 year after the fracture, which is comparable with other diseases. However, the suicide risk remained high from 365 days to last follow-up and may reflect the increased potential for repeated spinal fractures.

In conclusion, spinal fractures in older patients are associated with a substantially increased risk of death by suicide compared with matched controls. The suicide risk was greatest in the first year after fracture, but it remained high throughout the whole follow-up period. In the treatment for older patients with spinal fractures, surgeons should consider offering psychiatric evaluation and management more frequently, particularly in patients with chronic pain, functional disability, and depressive mood. Further studies should investigate the underlying causes of suicide, such as deteriorating socioeconomic support or depression, and whether early initiation of psychological support following injury can reduce the suicide rate.