Association between Type 2 Diabetes and 30-Day Clinical Outcomes following Hospital Discharge for Hip Fracture in Older Adults
Article information
Abstract
Background
Population aging is increasing the prevalence of type 2 diabetes mellitus (T2DM) and hip fractures, which are major contributors to morbidity, mortality, and healthcare costs among older adults. The coexistence of both conditions further worsens clinical outcomes and quality of life. This study analyzes the association between T2DM and 30-day outcomes after hip fracture.
Methods
Observational, longitudinal study with a 1-month follow-up after hospital discharge, including a total of 1,164 patients aged ≥65 years with hip fracture treated at a public university hospital between 2017 and 2022. The main variable was the presence of T2DM. Sociodemographic, clinical, and surgical variables, along with 30-day outcomes (mortality, readmission, reoperation, functional change, and change of residence), were collected.
Results
Patients with insulin-treated diabetes showed a higher percentage of 30-day hospital readmission (11.4%) compared with patients without diabetes (3.3%) and those with non–insulin-treated diabetes (2.9%). Insulin-treated diabetes was associated with an increased risk of 30-day hospital readmission (hazard ratio=3.36; 95% confidence interval 1.59–7.08). No statistically significant differences were observed in the remaining outcomes analyzed.
Conclusion
Individuals with insulin-treated diabetes exhibit an increased risk of hospital readmission following hip fracture. These findings highlight the need for targeted perioperative management, closer post-discharge monitoring, and diabetes-specific care pathways for this high-risk subgroup. Implementing individualized follow-up and optimizing glycemic control may help reduce early readmissions and improve short-term clinical outcomes in patients with type 2 diabetes, particularly those treated with insulin.
INTRODUCTION
Population aging represents one of the most profound demographic shifts of the 21st century. Demographic projections indicate that by 2050, the global population of individuals aged 60 years and older will approach 2 billion, effectively increasing their proportion from 10% to 21% of the total population.1,2) This demographic transition is accompanied by a parallel rise in the prevalence of chronic diseases closely associated with aging, such as type 2 diabetes mellitus (T2DM) and osteoporosis, the latter being a major underlying cause of hip fractures.3-5)
Both conditions are highly prevalent and are associated with substantial clinical and functional consequences.6,7)
In this context, the anticipated increase in hip fracture incidence is particularly concerning, with projections suggesting that the total number of cases will double between 2018 and 2050. Specifically, increases of approximately 310% in men and 240% in women are expected compared with figures reported in 1990.8)
Hip fractures constitute a major source of morbidity, disability, and mortality among older adults, especially in individuals with multiple comorbidities or complications during hospitalization. Up to 50% of patients fail to regain their pre-fracture level of functioning, and nearly 20% require long-term assistance with basic activities of daily living following a hip fracture.7,9-13)
The coexistence of T2DM and hip fracture poses an even greater clinical challenge, as each condition independently represents a major public health concern due to its negative impact on quality of life, healthcare burden, and costs.8,14,15) It is therefore plausible that the presence of both conditions may result in worse clinical outcomes. Several studies have reported increased mortality and poorer clinical progression in patients with T2DM after hip fracture.10,16-20) However, existing evidence remains heterogeneous, likely reflecting differences in study populations, methodological approaches, and follow-up duration across studies.16,21-25)
Focusing on short-term outcomes may provide a more precise and clinically actionable framework for healthcare planning and decision-making. Nevertheless, important knowledge gaps persist regarding the specific impact of T2DM on 30-day outcomes following hip fracture, particularly with respect to functional decline, institutionalization, hospital readmissions, and surgical reinterventions.
The aim of this study was to investigate the association between T2DM and key short-term outcomes at 30 days after hip fracture, focusing on mortality, hospital readmissions, surgical reinterventions, changes in functional status, and changes in residential status
MATERIALS AND METHODS
Design
A longitudinal observational cohort study was conducted using data from patients admitted for hip fracture at Hospital Universitario Infanta Sofía. Follow-up extended from hospital admission to 30 days after discharge. All participants were included in the Spanish National Hip Fracture Registry (Registro Nacional de Fractura de Cadera [RNFC]),26) a multicenter observational database established in 2017 that collects standardized clinical and care-related information from patients hospitalized with hip fracture in voluntarily participating hospitals of the Spanish National Health System
Population
Patients aged 65 years and older admitted for hip fracture at Infanta Sofía University Hospital between February 2017 and September 2022 were included. All patients were followed for one month after hospital discharge, and their data were recorded in the RNFC.
Data Collection
Data for this study were obtained from two sources: the RNFC and the local registry of Infanta Sofía University Hospital.
Variables
Primary variable
The primary variable of interest in this study was the diagnosis of T2DM at the time of hip fracture. This variable was classified into three groups: patients without diabetes, patients with insulin-treated diabetes, and patients with non-insulin-treated diabetes, including those managed with oral antidiabetic drugs or without recorded treatment.
Patient-related variables
Patient-related variables included age and sex, pre-fracture residence (home or long-term care facility), and overall health status, assessed using the American Society of Anesthesiologists (ASA) physical status classification.27) Health status was dichotomized into patients with low surgical risk (ASA I–II) and patients with high surgical risk (ASA III–VI).
Pre-fracture functional status was assessed using the Functional Ambulation Category (FAC) scale.28) The original six-level FAC scale, which describes progressive degrees of walking independence, was regrouped into three broader clinical categories, following an approach widely used in previous hip fracture studies to facilitate interpretation and clinical applicability.27) The first category included patients with independent ambulation; the second comprised patients requiring assistance to walk; and the third included patients who were unable to walk or required assistance from two people.
Fracture-related variables
Fracture type was recorded and classified as intracapsular fracture or extracapsular fracture, the latter including pertrochanteric and subtrochanteric fractures.29)
Hospital care process variables
Variables related to the hospital care process included type of anesthesia, type of surgery, surgical delay, and length of hospital stay.
Outcome variables
The outcome variables analyzed were vital status, 30-day hospital readmission due to causes related to hip fracture, and 30-day surgical reoperation.
Change of residence
Change of residence after hip fracture was analyzed. A change was considered to have occurred when, after hospital discharge, the patient moved from their usual residence to an institutional setting, such as a nursing home or a hospital-based rehabilitation unit.
Functional change
Functional change was evaluated using the same FAC-derived categories previously described. The FAC score obtained one month after hospital discharge was compared with the pre-fracture FAC score. Based on the difference between both measurements, functional evolution was classified as functional decline, functional improvement, or no functional change. This approach allowed a clear assessment of functional progression during the first month following hip fracture.
Statistical Analysis
Quantitative variables with a normal distribution are summarized using mean and standard deviation, while those without a normal distribution are described using median and interquartile range. Qualitative variables are presented as frequencies expressed as number (%). Normality was assessed using the Kolmogorov–Smirnov test. Associations between two qualitative variables were evaluated using the chi-square test. For categorical variables with more than two categories, standardized residuals were analyzed following significant chi-square results to identify specific cell differences. For comparisons involving qualitative variables with more than two categories and quantitative variables, ANOVA was used under normality assumptions, and the Kruskal–Wallis test otherwise.
Time-to-event analysis was performed for 30-day hospital readmission using Kaplan–Meier survival curves according to diabetes status and treatment groups, with differences assessed using the log-rank test. Cox proportional hazards regression analysis was conducted to estimate hazard ratios (HRs). Statistical significance was set at p<0.05. Analyses were conducted using IBM SPSS (version 29; IBM, Armonk, NY, USA) and JAMOVI software (version 2.4; https://www.jamovi.org).
Ethical and Legal Considerations
The study was approved by the Ethics Committee of Getafe University Hospital (Registration No. 24/91) on September 5, 2024, and by the Research Commission of Infanta Sofía University Hospital.
RESULTS
Of the 1,170 patients initially identified, four were excluded due to pathological fractures. A total of 1,164 patients aged 65 years and older admitted with hip fracture between February 2017 and September 2022 were included in the analysis (Fig. 1). Among them, 889 patients had no diabetes, 95 had insulin-treated diabetes, and 180 had non-insulin-treated diabetes. The overall prevalence of T2DM in the cohort was 23.6% (n=275). Of the total study population, 95 patients (8.2%) had insulin-treated diabetes, whereas 180 patients (15.5%) had non-insulin-treated diabetes.
Population diagram. T2DM, type 2 diabetes mellitus.
Baseline characteristics and hospital care variables according to diabetes status are shown in Table 1. Patients with insulin-treated diabetes were younger than those without diabetes (median 86 vs. 88 years; p<0.01). Severe health status at admission was more frequent among patients with insulin-treated diabetes (n=88, 93.6%) compared with patients without diabetes (n=721, 82.7%) (p=0.012).
Baseline characteristics of the sample and hospital care process outcomes
Differences were also observed in fracture type (p=0.003), with a higher proportion of extracapsular fractures among patients with insulin-treated diabetes (n=71, 74.7%) compared with patients without diabetes (n=504, 56.9%). Regarding surgical management, osteosynthesis was more frequently performed in patients with insulin-treated diabetes (n = 72, 78.3%) than in patients without diabetes (n=541, 62.8%) (p=0.013).
Length of hospital stay was longer in patients with insulin-treated diabetes than in patients without diabetes (median 10.0 vs. 8.9 days; p=0.024).
No statistically significant differences were observed among the groups with respect to sex, pre-fracture residence, baseline mobility, type of anesthesia, or surgical delay.
Functional and residential changes after hip fracture according to diabetes status are presented in Table 2. No statistically significant differences were observed in functional change, with functional worsening occurring in patients without diabetes (n=423, 55.7%), patients with insulin-treated diabetes (n=45, 56.3%), and patients with non-insulin-treated diabetes (n=95, 58.3%) (p=0.836). Similarly, no statistically significant differences were observed in change of residence after discharge, with no change in previous residence reported in patients without diabetes (n=585, 80.1%), patients with insulin-treated diabetes (n=54, 76.1%), and patients with non–insulin-treated diabetes (n=123, 78.8%) (p=0.693).
Functional and residential changes after hip fracture according to diabetes status
Clinical outcomes after hip fracture according to diabetes status are shown in Table 3. Mortality did not differ significantly among groups—n=95 (10.7%), n=13 (13.7%), and n=13 (7.2%), respectively (p=0.209), nor did surgical reoperation (p=0.987).
Clinical outcomes after hip fracture according to diabetes status
As hospital readmission was the only outcome with statistical significance, a time-to-event analysis was performed for this outcome. Kaplan–Meier curves showed differences between groups over the follow-up period, and the log-rank test demonstrated statistically significant differences (Fig. 2).
Kaplan–Meier analysis of clinical outcomes by diabetes management.
To estimate the magnitude of the association in terms of risk, a Cox proportional hazards regression analysis was conducted. In the univariable analysis, insulin-treated diabetes was associated with an increased risk of 30-day hospital readmission (HR=3.36; 95% confidence interval [CI] 1.59–7.08; p=0.001). In patients with non–insulin-treated diabetes, the association was not statistically significant (HR=1.19; 95% CI 0.45–3.15; p=0.724).
DISCUSSION
Our results show that patients with T2DM were younger at the time of fracture, had poorer overall health status, experienced a higher proportion of extracapsular fractures, and had longer hospital stays. Regarding clinical outcomes, no significant differences were observed in mortality, functional decline, reoperation, or changes in residence. However, hospital readmissions were more frequent among insulin-treated diabetic patients.
Taken together, these findings outline a distinct clinical profile in patients with T2DM sustaining a hip fracture. When comparing our results with previous studies in older populations with hip fractures, we observed that the prevalence of diabetes reported in the literature ranges between 15% and 30%,21,31,32) placing our findings within this expected interval.
Previous studies have reported conflicting findings regarding fracture type, with some identifying significant differences between intracapsular and extracapsular fractures,33,34) whereas others did not.35) The higher proportion of extracapsular fractures observed among patients with type 2 diabetes both in our cohort and in studies reporting similar findings, may be explained by a combination of metabolic and biomechanical mechanisms. Chronic hyperglycemia promotes the accumulation of advanced glycation end-products, which adversely affect bone material properties, particularly cortical bone, leading to reduced toughness and increased susceptibility to fracture.36)
Furthermore, because the extracapsular (trochanteric) region is predominantly composed of cortical bone, these alterations may preferentially increase the risk of extracapsular fractures in patients with diabetes. In addition, type 2 diabetes is frequently associated with sarcopenia, impaired balance, peripheral neuropathy, and visual disturbances, all of which increase the likelihood of sideways or lateral-impact falls.37) Such fall patterns are more likely to transmit force to the trochanteric region, further contributing to the predominance of extracapsular fractures in this population. Taken together, these metabolic and biomechanical factors provide a plausible clinical explanation for the fracture pattern observed in our cohort, although causal inferences cannot be drawn from the present observational design.
In our cohort, patients with diabetes, particularly those treated with insulin, presented poorer preoperative health status compared with patients without diabetes, a finding consistent with previous reports in hip fracture populations.16,35,37) This observation supports the interpretation that insulin-treated diabetes identifies a more clinically vulnerable subgroup following hip fracture.
This relationship may be attributed to the chronic and multisystem nature of diabetes, which increases the risk of complications and contributes to overall health deterioration.
In our study, higher 30-day mortality was observed in patients with insulin-treated diabetes (13.7%) compared with patients without diabetes (10.7%) and those with non-insulin-treated diabetes (7.2%). Although these differences did not reach statistical significance, they should be interpreted with caution given the limited statistical power of the insulin-treated subgroup.
A population-based hip fracture cohort study conducted in Taiwan showed significantly higher 1-year mortality in patients with insulin-treated diabetes (25.8%) compared with patients without diabetes (17.6%) and those treated with oral antidiabetic agents (14.9%).16)
By specifically examining postoperative outcomes in patients undergoing hip fracture surgery, this study supports the interpretation that insulin-treated individuals constitute a clinically more vulnerable subgroup within the hip fracture population.
The lack of statistically significant differences in mortality and functional outcomes should be interpreted with caution, as the relatively small size of the insulin-treated subgroup limits the statistical power to detect clinically meaningful effects. This increases the likelihood of a type II error, whereby true differences may remain undetected. Notably, our cohort shows a consistent trend toward higher mortality among patients with insulin-treated diabetes, suggesting that these differences might reach statistical significance in studies with greater statistical power. Given that patients with insulin-treated diabetes often represent a subgroup with greater comorbidity and clinical vulnerability, adequately powered studies are needed to determine whether the trends observed here reflect true underlying associations.
Therefore, our preliminary findings call for further investigation through prospective studies with larger sample sizes and multivariate adjustments to clarify the true impact of insulin therapy on mortality in patients with T2DM.
In our study, patients with insulin-treated diabetes showed a significantly higher 30-day hospital readmission rate compared with non-diabetic patients and those with non-insulin-treated diabetes. All recorded readmissions in our cohort were related to complications of the index hip fracture and its management.
This finding is consistent with previous studies conducted in hip fracture–specific cohorts, which have reported higher early readmission rates among patients with diabetes, particularly those treated with insulin, compared with non-diabetic patients or those treated with oral antidiabetic agents.16)
However, as in our study, most previous hip fracture cohorts report readmission as a fracture-related outcome without providing detailed information on the specific complication leading to readmission. Therefore, in the absence of cause-specific data, the higher readmission rate observed among patients with insulin-treated diabetes should be interpreted as an indicator of increased susceptibility to fracture-related complications during the early post-discharge period, rather than as evidence of a single underlying pathophysiological mechanism.
Taken together, these findings highlight the importance of closer post-discharge monitoring and targeted transitional care strategies aimed at the prevention and early detection of fracture-related complications in this high-risk subgroup.
Regarding functionality after hip fracture, no significant differences were observed; however, the non-diabetic group showed a higher percentage of improvement (12.5%) compared to patients with diabetes, especially those treated with insulin (8.8%). Although not statistically significant, this difference may reflect a lower functional recovery capacity in patients with diabetes, although no statistically significant differences were observed.
In our study, change of residence after hip fracture showed no significant differences between patients without diabetes, insulin-treated diabetes, and with diabetes without insulin (p=0.693). Most patients remained in their previous residence, while a smaller percentage required transfer to a nursing home or rehabilitation unit.
The current literature shows a lack of studies specifically addressing changes in functional status and residential arrangements among non-diabetic patients, diabetic patients treated with insulin, and diabetic patients not treated with insulin following hip fracture. Given that functional recovery and changes in living situation are critical outcomes for patient prognosis, further research investigating these variables is essential to enhance understanding and improve clinical management of this population.
In this context, our findings are particularly relevant as they provide novel insights with direct implications for healthcare management during the hospitalization of patients with hip fractures and diabetes.
A relevant limitation of this study is the absence of several diabetes-related variables, such as HbA1c levels, duration of diabetes, and the presence of diabetes-related complications, which would have allowed a more detailed characterization of metabolic control and disease severity. In addition, information on certain comorbidities that may influence postoperative outcomes was not available, and their inclusion could have provided greater analytical precision.
Another limitation is that, although all 30-day hospital readmissions were related to complications of the index hip fracture and its management, detailed information on the specific type of complication leading to readmission was not recorded in our dataset. This limitation precluded a more in-depth exploration of the mechanisms underlying this outcome.
Nevertheless, classification based on insulin treatment provides a clinically meaningful distinction, and the findings offer valuable insights into this patient subgroup. Future studies incorporating these additional covariates and more granular information on fracture-related complications may help refine risk stratification and further clarify the underlying mechanisms.
Despite these limitations, the large sample size, the extended duration of the registry, and the comprehensive collection of clinical variables support the robustness of this study and its contribution to understanding the clinical course of patients with T2DM following hip fracture.
In conclusion, diabetes was not associated with increased mortality, the need for reintervention, or an elevated risk of functional decline or change in usual residence within 30 days following hip fracture. However, a significant association was observed between insulin-treated diabetes and a higher risk of hospital readmission.
These findings underscore the need for differentiated care and closer monitoring of patients with type 2 diabetes, particularly those receiving insulin therapy, given their increased clinical vulnerability and risk of readmission. The information obtained may aid in improving care planning and developing more personalized management strategies, although confirmation through prospective cohort studies with longer follow-up is warranted.
Notes
CONFLICT OF INTEREST
The researchers claim no conflicts of interest.
FUNDING
None.
AUTHOR CONTRIBUTIONS
Conceptualization, RQM, MNA; Investigation, JGC; Methodology, RQM; Supervision, MNA, JGC; Formal analysis, RQM; Writing_original draft, RQM; Writing_review & editing, MNA, JGC; Visualization, RQM.