Introduction

Hip fracture is a major global public health concern, with an estimated 1.5 million annual cases worldwide1. This number is projected to increase to 2.6 million by 2025 and 4.5 million by 20502. Hip fracture results in decreased functionality, a reduced quality of life, increased dependency, higher physical, mental, and economic burden, and elevated mortality risk. Despite advancements in surgical techniques and patient care, whether mortality rates are improving is unclear. Early mortality rates range from 5–10%3, whereas the mortality rate within one year after surgery can reach 8–36%4,5,6,7. Therefore, more research is urgently required to improve patient prognosis.

The causes of the high mortality rate after hip fracture are debated, with complications such as pulmonary embolism8, infection9, and heart failure9,10 potentially contributing. Factors related to falls and the risk of sustained osteoporotic fractures may also be involved11, along with low bone mineral density12. Additional analysis of mortality and its causes is crucial for identifying risk factors and predicting complications13.

Previous studies indicated cardiovascular disease and pneumonia as common causes of death after hip fracture14,15,16. However, the specific causes of death in the early postfracture period remain unclear. Given the projected increase in hip fractures, evaluating detailed information on specific causes of death within 30 days of fracture can help identify high-risk patients and enable personalized clinical care, optimize outcomes, and reduce mortality risk. This study aimed to analyze the 30-day mortality risk, causes of death and risk factors associated with death after hip fracture.

Patients selection and methods

Data source

This study used data from the National Health Insurance Research Database (NHIRD) from 2000 to 2015. The NHIRD is the health insurance claims database for Taiwan’s National Health Insurance system. It encompasses demographic information, prescription drug usage, and disease records for more than 99% of the Taiwanese population. The National Death Registry contains data on primary and contributing causes of death and the date of death of all citizens. To protect privacy, all claim and registration files in the NHIRD are anonymized and given identification numbers. These anonymized data are available for academic purposes and provide valuable research opportunities. Our study identified diseases of interest using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes. The NHIRD has been validated, with research articles based on it published in reputable scientific journals worldwide17,18,19,20.

Ethics statement

The Institutional Review Board of the Ditmanson Medical Foundation Chia-Yi Christian Hospital, Taiwan (CYCH-IRB No: 2019063) approved this study, waiving the need for informed consent due to the anonymized nature of the data. All procedures adhered to pertinent guidelines and regulations, with full compliance ensured.

Case definition

We used the NHIRD to identify patients 50 years or older upon admission who had been diagnosed with their first hip fracture (ICD-9-CM: 820) between January 1, 2000, and December 31, 2015. Our analysis focused on inpatient cases and considered only the initial hospitalization records. Individuals with missing information on sex or age were excluded (Fig. 1).

Fig. 1
figure 1

Flowchart showing patient selection in this study.

Full size image

Mortality outcomes

We obtained information regarding the vital status, date of death, and cause of death of the study participants from the National Death Registry of Taiwan. We analyzed data on prespecified causes of death, including circulatory system diseases, malignant neoplasms, respiratory system diseases, infectious diseases, digestive system diseases, urogenital diseases, endocrine, nutritional, and metabolic diseases, accidents and unintentional injuries, and other causes. The other causes comprised deaths from specific categories, each accounting for < 3% of total deaths within a 30-day period. These included diseases related to the musculoskeletal system/connective tissue, the skin and subcutaneous tissue, the blood and blood-forming organs and certain disorders involving immune mechanisms, and the nervous system and sense organs; congenital malformations and chromosomal abnormalities; mental and behavioral disorders; intentional self-harm or suicide; and symptoms, signs, and ill-defined causes. The primary cause of death was determined by evaluating the ICD-9-CM codes, displayed in Supplementary Table S1. The data quality of the National Death Registry has undergone rigorous evaluation and is considered both valid and comprehensive, ensuring that the gathered information is reliable21.

Covariates

The covariates included age, sex, the Charlson Comorbidity Index (CCI), urbanization, and income level. The CCI was calculated by assigning weights to the comorbidities22. The identification of comorbidities was based on the ICD-9-CM codes extracted from outpatient and inpatient reimbursement claims in the NHIRD. The comorbidity was defined based on diagnostic codes from at least two outpatient visits or one inpatient visit within 1 year prior to the hip fracture (defined as the index date). The CCI scores were categorized into four groups (scores of 0, 1, 2, and ≥ 3). Urbanization level was classified into three categories (urban, suburban, and rural), whereas economic status was divided into three groups based on the monthly insurable wage in New Taiwan dollars (NTD): <19,100 (low income level), between 19,100 and 41,999 (intermediate income level), and ≥ 42,000 (high income level)23.

Statistical analysis

Continuous variables are reported as mean ± standard deviation and were compared using the t-test. Categorical variables are presented as percentages and were assessed using the Chi-square test for comparisons of categorical data. We examined the crude mortality rate per 100 people for all-cause mortality and the specified causes of death in the study population. Subgroup analyses were conducted based on sex, age, and year of hip fracture occurrence. To evaluate whether the 30-day mortality risk changed between 2000 and 2015, we conducted a joinpoint regression analysis using v.4.8.0.1 of the National Cancer Institute (US)24. This method identifies when changes in trends occur (inflection points). The dependent variable is a percentage, and the year of death is the independent variable. To test whether these inflection points were statistically significant and should be added to the model, four maximum inflection points and 4,499 Monte Carlo Permutation tests were conducted. The changes in the slope of these trends were calculated as the annual percentage change (APC), with a 95% confidence interval for each segment. This analysis identified periods with statistically different log-linear trends in 30-day mortality. In addition, we calculated the APC per period and the average APC over the entire study period. We also calculated standardized mortality ratios (SMRs) within 30 days. First, we collected the observed number of deaths within 30 days for our study population from Taiwanese mortality statistics. Then, we calculated the expected number of deaths by applying age-specific mortality rates from a reference population to the age distribution of our study population. The SMR was calculated as the ratio of observed to expect deaths. An SMR > 1 indicates excess mortality compared with the reference population, whereas an SMR < 1 suggests lower mortality. We conducted additional data analyses, including calculations of confidence intervals and subgroup analyses, to explore the SMR results further. A Cox proportional hazard model was also used to identify factors associated with 30-day mortality risk from hip fracture. All statistical analyses were conducted using SAS v.9.4 software (SAS Institute, Cary, NC, USA), and a significance threshold of 0.05 was set for two-tailed p-values.

Results

Characteristics of the study population

From January 1, 2000, to December 31, 2015, 285,891 patients experienced hip fractures. Among the patients, 8,505 (2.98%) died within 30 days of sustaining a hip fracture. Of these, 4,416 were men (51.92%) and 4,089 were women (48.08%). The patients who died within 30 days were older (79.51 ± 7.25 years) compared to the survivors (75.87 ± 8.99 years) (p < 0.001). They also had a higher CCI than survivors. Furthermore, significant differences in residence and income level were observed between the 30-day survivors and nonsurvivors (Table 1).

Table 1 Demographic characteristics among patients from 2000 to 2015.
Full size table

Trends in hip fracture mortality

The mortality risk within 30 days of hip fracture decreased from 3.76% in 2000 to 2.92% in 2015. The 30-day mortality risk initially trended downward from 2000 to 2006 (APC = − 5.29, p < 0.05). However, it then displayed a gradually increasing trend (APC = 0.96, p = 0.167) (Supplementary Figure S1). Over the 16-year study period, the 30-day mortality risk after hip fracture declined among women from 3.21 in 2000 to 2.27 in 2015, whereas men experienced a decrease from 4.57 in 2000 to 3.98 in 2015. However, males had a consistently higher 30-day mortality risk than females each year (Fig. 2a). We used the annual SMR to compare the 30-day mortality of hip fractures with that of the general population, and we found that it decreased from 1.71 to 1.65 overall and went from 1.79 to 1.62 in women. However, the annual SMR for males increased from 1.75 to 1.83 (Tables 2 and 3). Males had a higher mortality risk than females across all age groups, and their mortality risk increased with age (Fig. 2b).

Fig. 2
figure 2

a Mortality rates of hip fracture for Taiwanese adults aged older than 50, stratified by sex and year (2000–2015); b age-related mortality rates of hip fracture in the overall population, females, and males.

Full size image
Table 2 Crude 30-day mortality and standardized mortality ratio (SMR) in patients with hip fracture by year.
Full size table
Table 3 Crude 30-day mortality and standardized mortality ratio (SMR) in patients with hip fracture by sex and year.
Full size table

Risk factors for hip fracture mortality

The multivariate Cox proportional hazards analysis revealed independent risk factors for hip fracture mortality. These included older age, male sex, higher CCI, residing in suburban, rural, or remote areas, and lower income level (Table 4).

Table 4 Hazard ratios for 30–days mortality after hip fracture based on multivariable cox regression.
Full size table

Significant causes of death after hip fracture

The pie chart depicted in Fig. 3a provides an overview of the percentage of individuals who died within 30 days after sustaining a hip fracture. During this period, circulatory system diseases were the primary cause of death, accounting for 22% of all cases. Accidents and unintentional injuries closely followed at 18%, malignant neoplasms accounted for 10%, and infectious diseases accounted for 9%. Among infectious diseases, pneumonia was the most prevalent, accounting for 64.11% of infectious disease-related deaths (594 out of 794). When examining the data by sex, circulatory system diseases were the leading cause of death among women. Men had a higher incidence of mortality due to circulatory system diseases and accidents/unintentional injuries (Fig. 3b). Analyzing the data by age group, accidents and unintentional injuries (26.81%) were the leading causes of death among individuals younger than 70. Conversely, circulatory system diseases (22.78%) were the predominant cause of mortality among those aged 70 years and older (Fig. 3c). Furthermore, when stratifying by age and sex, similar proportional distributions of the significant causes of death were observed (Supplementary Figure S2).

Fig. 3
figure 3

The proportional distribution of the major causes of death among hip fracture patients. a Overall population; b by sex; c by age group.

Full size image

Discussion

We assessed the factors associated with mortality and specific causes of death within 30 days after hip fracture hospitalization. We observed a decrease in the 30-day mortality risk from 3.76 per 100 people in 2000 to 2.92 per 100 people in 2015. Men consistently had a higher 30-day mortality risk than women across all age groups, and their mortality risk increased with age. We also found that older age, male sex, higher CCI, residing in suburban, rural, or remote areas, and lower income level were risk factors for hip fracture mortality. Furthermore, circulatory system diseases, especially ischemic heart disease and cerebrovascular diseases, were the leading cause of death within 30 days after hip fracture, followed by accidents and unintentional injuries, malignant neoplasms, and infectious diseases, with pneumonia being the most common among the latter.

Our findings are consistent with previous studies, which have indicated that circulatory system diseases like ischemic heart disease and cerebrovascular disease are common causes of death within 30 days after hip fracture25,26. This may be because patients with circulatory system diseases, particularly those with frailty, are more susceptible to falls due to impaired blood circulation, which increases the risk of fragility fractures such as hip fractures27,28,29. Consequently, this elevated fracture risk may also increase the likelihood of mortality after fracture30.

Additionally, previous studies have indicated that patients with a history of stroke have a greater risk of postoperative complications following hip fracture31,32. Patients with stroke typically exhibit a higher prevalence of unfavorable prognostic factors, such as postoperative delirium, nutritional problems, and impaired function. These factors collectively contribute to increased mortality rates. The significant impact of circulatory system diseases on early mortality after hip fracture has crucial clinical implications. Combining this knowledge with an understanding of factors associated with early mortality, high-risk patients could be identified and provided early multidisciplinary interventions.

Pneumonia is a common cause of death and is a preventable factor in patients with hip fractures14,33. It can be classified as either bacterial pneumonia or aspiration pneumonia. Maintaining oral hygiene, quitting smoking, and receiving influenza or pneumococcal vaccinations can prevent bacterial pneumonia34. Aspiration pneumonia is also prevalent in patients with hip fractures, often accompanied by dysphagia and weakness35. Improving swallowing postural training and preventing gastroesophageal reflux in patient education can reduce the occurrence of pneumonia36. Therefore, unless clinically unstable, patients with a history of pneumonia should not have their surgery delayed. Additionally, early initiation of antibiotic treatment and postoperative pulmonary rehabilitation programs are recommended in these patients37.

Previous studies have shown that men have a higher mortality rate than women after hip fracture38,39, which is supported by our findings. However, the exact reasons for this sex difference remain unclear40. There are several possible explanations for this disparity. First, men may be sicker and frailer than women when fractures occur40. Second, males with hip fractures often have poorer bone health (underdiagnosed and undertreated osteoporosis), which contributes to the increased mortality risk41. Third, previous research has shown that men with hip fractures have a higher incidence of complications than women. This difference in outcomes has been linked to the higher prevalence and severity of preexisting chronic diseases among men with hip fractures. These underlying health conditions may increase the risk of postoperative complications in male patients11,42. Furthermore, male hip fracture patients show a more profound impairment in daily activities, functional capacity, and walking speed than women. Therefore, men may experience a greater physiological reserve loss after hip fracture, leading to a higher mortality risk43.

Our findings that residents of non-urban areas have a higher mortality rate following hip fractures are consistent with the results of other studies44,45,46. This increased mortality may be due to several contributing factors. Limited access to healthcare services and medical professionals in rural areas can delay treatment, leading to a higher risk of complications and death. Moreover, lower income levels may restrict access to quality care, while inadequate management of chronic conditions, such as cardiovascular disease, can further increase mortality risk47. The geographic distance to medical centers and the resulting delays in receiving care, as well as reduced social support during recovery, are additional challenges that likely contribute to the elevated mortality rate observed in rural populations.

We identified sex-based differences in the causes of death following hip fracture. Our findings revealed that men were more likely than women to die from respiratory diseases, which is potentially attributed to the higher prevalence of smoking among men, as suggested by a previous study in Taiwan48. This may explain the higher mortality rate observed in men, because they are more susceptible to experiencing exacerbated respiratory issues after hip fracture surgery, including reduced airway secretions, impaired daily activity, and subsequent chest complications. We also found that men had a higher mortality rate due to infectious diseases than women. This disparity may be linked to the age-related decline in immune function, which can increase susceptibility to infections among the elderly49. Several studies have highlighted male sex as a risk factor for significant infections in postfracture populations50,51, with male patients in surgical intensive care units having a higher incidence of severe sepsis or septic shock52, leading to an elevated mortality risk from infections. Moreover, accidents and unintentional injuries also contribute to the higher mortality rate among men. This may be because, in Taiwan, males more often drive and ride motorcycles than women, increasing their risk of vehicle and traffic-related accidents, potentially leading to a higher mortality rate53.

Our data show a 30-day mortality rate of 2.97%, consistent with previous reports from Taiwan54 and comparable to other Asian countries55. Japan (1.2%), Thailand (1.2%), South Korea (1.9%), Singapore (2.2%), China (3.3%), and Hong Kong SAR (3.0%) report rates between 1.2% and 3.3%55, which are notably lower than those in England (9.6%), Scotland (7%), and the US (5.2–9.3%)9,56,57,58. This lower mortality in Asia may be influenced by healthier diets, lower obesity rates, and better healthcare access through national insurance systems, enabling timely treatment. Earlier surgical intervention and comprehensive post-operative care, along with cultural factors like stronger family support and more active elderly lifestyles, also likely contribute to better outcomes. Previous studies have reported a declining trend in the 30-day mortality rate26,58, whereas a study from Sweden showed the opposite trend59. However, our study observed a declining trend followed by a plateau. The exact reasons for this change in the mortality rate remain unclear. The decrease in mortality rates before 2006 may be attributed to the launch in 1995 of the national insurance program, which improved healthcare services in Taiwan. Additionally, advances in surgical techniques and the wide-scale implementation of the case payment system by Taiwan’s health insurance program in 2002 provided better funding and more comprehensive care for patients with hip fracture. These factors may have contributed to the stable mortality rates observed after 2006.

The strengths of this study include its large sample size derived from a national database, a long follow-up period, and its focus on a topic with relatively few existing studies. This underscores the significant opportunities for prevention and intervention. However, some limitations should be noted. Our study relied on national claims data, which did not allow us to evaluate the impact of factors such as body mass index, smoking history, alcohol abuse, nutritional status, surgery-related information, and postoperative health status on the 30-day mortality rate after hip fracture. Additionally, the expiration of our database access rights prevented us from analyzing the specific causes of hip fractures and assessing how different factors, such as fragility fractures versus fractures caused by metastases or traffic accidents, might influence mortality rates.

Conclusion and future directions

Our research indicates that there has been a declining trend in the 30-day mortality rate after hip fracture in Taiwan. We observed a significant correlation between the 30-day mortality rate and male sex, increasing age, the presence of comorbidities, place of residence, and lower income. The primary causes of death within 30 days were circulatory system diseases, accidents and unintentional injuries, malignant tumors, and infectious diseases. Early identification of high-risk patients is crucial for implementing targeted interventions to reduce in-hospital complications. Since patients with hip fracture may have limited physiological reserves, these patients should receive treatment from experienced clinicians. Identifying and addressing their specific needs can optimize outcomes and minimize complications during hospitalization, thereby reducing mortality rates.