Long-Term Survival in Frail Older Adults Sustaining a Hip Fracture: Does the Perioperative Period Really Matter?

Purpose

Hip fractures in older adults are a major challenge for public health, it is associated with increased morbidity and mortality, especially in frailty older adults. An accurate evaluation of preventable perioperative risk factors is essential to optimize care pathways and clinical outcomes. This study aims to gauge evidence on the predictive accuracy of perioperative variables, including the Nottingham Hip Fracture Score (NHFS), on overall long-term survival (OS) in frail, hospitalized hip-fractured older adults.

Methods

From March 2020 to September 2021, 433 elderly patients with hip fractures received multidisciplinary orthogeriatric care in Policlinic San Martino Hospital (Genova, Italy). Enrolled patients received geriatric assessment (CGA) within 24 hours from hospital admission and their medical conditions were assessed alongside post-operative complications and survival rates. Statistical analyses, including Cox models, evaluated factor influencing overall survival.

Results

The patients’ most prevalent clinical phenotype was frailty, sustained by loss of muscle strength, malnutrition and functional disability. This phenotype reflects the unique demographic of the Liguria region (Italy) with the highest proportion of oldest-old (≥ 85 years) and frail individuals in the nation. Results showed NHFS > 5, vitamin D deficiency and delayed verticalization as key predictive determinants of long-term mortality (up to 21 months). Resilience to perioperative stressors in frail hip-fractured patients is crucial to the success of treatments aimed at improving their physical recovery.

Conclusions

Our research showed that targeting preventable factors during surgery can greatly impact the functional reserve of frail elderly patients, influencing their recovery and long-term outcomes, including mortality rates.

INTRODUCTION

Aging is associated with increased multimorbidity and frailty. Hip fractures are becoming rapidly a growing public health concern worldwide, as the average age of people who experience hip fractures is 80 years old and over ¹. Italy has the highest incidence of proximal femoral fracture, with an annual prevalence of more than 300 cases/100.000 inhabitants for women and more than 150 cases for men ²; European projections estimate that the burden for proximal fracture in the elderly will increase to 6.26 million by 2050 ³.

Mounting evidence underscored that hip fracture in older frail patients is associated with increased immediate short-term mortality (up to 10% ^4,5) and long-term (1-year) mortality (12-37%) ^6,7]. Overall mortality for any cause after experience hip fracture has been observed to be higher even after two decades of follow-up in such vulnerable patients ⁸.

The risk of short-term mortality is mediated by both pre-operative ^9,10 and post-operative ^11,12 factors. Furthermore, several risk scoring systems have been developed to improve clinical understanding. The Nottingham Hip Fracture Score (NHFS) ¹³ is a highly accurate risk-scoring system that predicts 30-day mortality after surgery. It is based on a multi-component assessment that includes age, sex, number of comorbidities (≥ 2), dementia status (Abbreviated Mental Test Score ≤ 6 out of 10), hemoglobin concentration at admission (≤ 10 g/dL), presence of malignant disease, and nursing home placement, showing good predictive accuracy in older populations.

However, advanced age is correlated with multimorbidity, malnutrition, osteoporosis, sarcopenia and reduced physical activity ^1,14 which may impact on overall survival after hip fracture. Frailty is a common geriatric syndrome, it is characterized by an excessive vulnerability to endogenous and exogenous stressors and is strongly associated with higher mortality, functional disability, and reduced quality of life ^15-17. There is little research that has investigated the role of frailty in the determination of long-term outcomes after hip fracture and it is still undetermined whether or not the status of frail patients with fragility fractures may affect their ability to tolerate surgery and the perioperative period, reducing functional reserve and resilience longitudinally.

With this background, the aim of the study is to assess the role of frailty within a set of perioperative variables, including NFHS, in predicting overall long-term survival in hip-fractured old-age hospitalized patients.

METHODS

From March 2020 to September 2021, 433 patients affected by hip fractures and aged 65 years or over were admitted to the orthogeriatric ward of IRCCS Policlinic San Martino Hospital (Genova, Italy) and enrolled in the study.

Enrolled patients were ≥ 65 years old and had sustained hip fractures due to low-energy impact trauma, such as fragility fractures, lateral pertrochanteric hip fractures and medial hip fractures. Exclusion criteria was surgery prohibited by surgical or clinical instability, high-energy impact trauma involvement, and periprosthetic fractures by nature. All patients received in-hospital multidisciplinary orthogeriatric care from a trained team of geriatricians, orthopedics, physiotherapists and professional nurses.

In the absence of Italian standards, the orthogeriatric service of IRCCS Policlinic San Martino Hospital followed established international protocols, particularly those from the National Institute for Health and Care Excellence (NICE) ¹⁸ and the American Academy of Orthopaedic Surgeons (AAOS) ¹⁹, to guide the management of hip fractures in elderly patients. Only in October 2021 did Italy introduce official national guidelines by the Italian Society of Orthopaedics and Traumatology (SIOT) ²⁰, providing a standardized approach for the management of femur fractures in older adults across the country.

All patients received Comprehensive Geriatric Assessment (CGA) within 24 hours from hospital admission, investigating domain such as functional status, comorbidity, cognition and nutrition.

Demographical data, polypharmacy, number of falls in the previous year and a set of biochemical data, including hemoglobin and vitamin D levels, were collected at hospital admission.

The multidimensional and integrated assessment also included: Handgrip test (HG, by using a GIMA 28791 Smedley dynamometer) to assess muscle strength (defined normal > 27 kg for men and > 16 kg for women ²¹), Barthel Index (score 100 for full independence; 91-99 for slight dependence; 61-90 for moderate dependence; 21-60 for severe dependence; 0-20 for total dependence ²²) and Instrumental Activities of Daily Living (IADL) to assess functional status (0 out of 8 IADL preserved: complete dependence; 1-2 preserved: severe dependence; 3-4 preserved: moderate dependence; 5-6 preserved: slight dependence; 7-8 preserved: full independence ²³). Short Portable Mental Status Questionnaire (SPMSQ) was performed to screen for cognitive status (0-2 errors: normal cognitive status; 3-4 errors: mild cognitive impairment; 5-7 errors: moderate cognitive impairment; 8-10 errors: severe cognitive impairment ²⁴), Mini Nutritional Assessment Short-Form (MNA-SF) was used to screen malnutrition (score 12-14 for normal nutritional status; 8-11 for risk of malnutrition; 0-7 for malnutrition ²⁵), and Cumulative Illness Rating Scale-Geriatric (CIRS)- Comorbidity Index was executed to assess multimorbidity (counting the number of categories in which a score of 3 or higher is obtained, excluding the psychiatric/behavioral disorders category ²⁶).

Frailty was defined as the presence of at least four altered domains of the CGA (cut-off > 3) ²⁷.

NHFS was performed before surgery according to clinical records from the emergency room (ER) and/or the orthogeriatric ward. Vitamin D levels were categorized according to 25-OH-cholecalciferol blood levels normative data: ideal range 50-100 ng/mL; insufficiency < 30 ng/mL (severe deficiency if < 10 ng/mL).

Post-operative complications such as anemia requiring blood transfusion, delirium, acute renal failure, urinary tract infections, respiratory complications such as pulmonary embolism, cardiac complications and pressure sores, time to surgery, time to verticalization and allocation after hospital discharge were also collected.

After hospital discharge overall survival (OS) was recorded through the ASL3 Genova (Italy) county electronic database. The date of censorship was February 2022.

STATISTICS

Descriptive data was reported as mean with standard deviation or median with IQR. Continuous variables (Barthel, IADL, NHFS) were included in the Cox model according to literature cut-offs. Overall survival (OS) was calculated using the Kaplan-Meier method. The association between demographic and clinical variables with OS was assessed by using the proportional hazards Cox model. The best cut-off for NHFS was defined using the survival ROC curve with a timepoint set at 6 months. A multivariable Cox model was performed including all significant clinical variables at the univariable analysis with a p-value < 0.10. Results were reported as a hazard ratio (HR) along with the 95% confidence interval (CI). The Harrell’s concordance index was used to measure the discriminative ability of each clinical variable. Stata (v.16; StataCorp) was used for the computation.

RESULTS

CLINICAL PHENOTYPE OF THE PATIENTS

The clinical characteristics of the monitored patients are illustrated in Table I. The patients’ mean age was 85.7 years old (SD 6.7, range 65-102 years). Women accounted for 58.8% (n = 284). Patients’ most prevalent clinical phenotype was frailty (70.8%), followed by malnutrition (64.4%, n = 278 had an MNA score < 12), loss of muscle strength (88.7%, with a median handgrip strength of 12.2 kg), and functional disability (54.3%, n = 235 was dependent in > 2 IADL).

The median SPMSQ score was 3 (IQR: 5; range: 0-10) and among the 433 patients assessed, 239 (55.2%) showed signs of cognitive impairment. Of these patients, 79 patients (19.4%) had mild cognitive impairment, 87 (21.4%) had moderate cognitive impairment, and 73 (17.9%) had severe cognitive impairment. The median CIRS Comorbidity Index was 4 (IQR 3-5), indicating the level of comorbidity in the population.

Before hospital admission, most patients (91.6%, n = 394) lived at home and more than half of the population (59.9%, n = 259) were on home medication involving over five drugs. Patients experienced an average of 0.82 falls in the previous year (median: 0, range: 0-10), and 90 patients (20.8%) had a Barthel Index below 50, indicating a high level of dependency.

The perioperative variables analyzed showed median 25OH-Vitamin D levels of 11.4 ng/mL, with 102 patients (23.6%) showing deficiency (10-30 ng/mL), 72 patients (16.7%) insufficient levels (< 30 ng/mL) and 202 patients (46.8%) severe deficiency (< 10 ng/mL).

The most common hip fractures were the lateral one (51.6%, n = 222), followed by medial fractures (39.8%, n = 171), and other types of fractures (8.6%, n = 37). Two cases were missing fracture type data.

The median waiting time for surgery following emergency room admission was 2 days (range: 0-11 days) and the median time to verticalization was 1 day (range: 0-18 days).

The most common postoperative complications were anemia requiring blood transfusion (81.7%, n = 353), followed by postoperative delirium (29.6%, n = 128), pneumological complications (18.8%, n = 81), acute renal failure and/or urinary tract infections (15.5%, n = 67), and other complications including cardiologic (4.2%, n = 18) and gastrointestinal (3.5%, n = 12) one.

Most of the patients were discharged in nursing homes (38.2%, n = 65) or at home (31.0%, n = 134), while a minor part was transferred to rehabilitation centers (20.1%, n = 87) or other wards (4.6%, n = 20).

SURVIVAL ANALYSIS

Cox regression did not detect significant difference in age between survivors and non-survivors (p-value = 0.52). Fracture type shows no significant impact on survival (p-value = 0.083). The difference in surgery waiting time between survivors and non-survivors has limited and no significant predictive power for survival (p-value = 0.21). However, the achievement of verticalization (postoperative mobilization) was significantly associated with survival (p-value = 0.001), and patients who were mobilized had a lower risk of mortality (c-index = 0.520, moderate predictive value).

No significant difference in handgrip strength was observed between survivors and non-survivors as it has slightly better predictive value than some other factors, but still low overall (c-index = 0.547).

The difference in nutritional status was not statistically significant (p-value = 0.554), while vitamin D deficiency was statistically significant more common among non-survivors (p-value= 0.041).

Non-survivors had significantly higher levels of multimorbidity (p = 0.002), assessed through CIRS Comorbidity Index, and significantly poorer functional autonomy in activities of daily living when compared to survivors (p = 0.008).

A significant difference in NHFS scores was also observed (p-value = 0.004), with non-survivors having higher NHFS scores, reflecting worse outcomes.

An NHFS cut-off of 5 was the optimal value for the prediction of OS, signifying that patients with NHFS ≥ 5 were at higher overall risk (HR = 2.04; p = 0.002) of mortality (Figure 1).

Supplementary material S2 illustrated Kaplan Meier curves, based on the respective stratification for vitamin D levels and time to verticalization.

The multivariable analysis showed that delayed verticalization > 48 hours (HR 2.31 p = 0.002), NHFS > 5 (HR 1.96, p = 0.004) and vitamin D deficiency (HR 2.31, p = 0.048) were associated with reduced OS (Tab. III).

DISCUSSION

The multidimensional clinical management of frail older patients with hip fractures extends beyond traditional surgical approaches. Frailty stratification can help assess long-term mortality after a surgically stressing event, such as hip fracture ^28-31. Our previous findings demonstrated that a set of perioperative variables can influence long-term overall mortality, even in robust/fit individuals. This evidence highlights the importance of proper clinical and surgical management, especially in a population of vulnerable frail older adults, as those enrolled in our study.

The CGA has been implemented for older people having surgery and has been demonstrated to have efficacy in medical, surgical, and outpatient settings with reductions in mortality up to a year of follow-up ³². In this setting, the CGA gave us a complete overview of the patient’s condition, offering frailty stratification and enabling integrated therapeutic planning.

Enrolled patients had a premorbid advanced frailty status sustained by malnutrition, loss of muscle strength, and functional disability. This phenotype reflects the unique demographic of the Liguria region (Italy) with the highest proportion of oldest-old (≥ 85 years) and frail individuals in the nation ³³.

In the past years, several tools have been developed to predict mortality and to help identify orthogeriatric patients at higher risk. Among these tools, the NHFS has demonstrated the most promising results in predicting 30-day mortality in patients with hip fractures ³⁴.

Our findings are in line with previous studies that have recommended a NHFS ≥ 5 as the threshold for distinguishing between low- and high-risk patients ³⁵.

Surgery had been performed within 48 hours from admission to ER and patients were managed by the orthogeriatric team in a timely manner to maximize recovery, rehabilitation (verticalization < 48 hours ³⁶), and prognostic outcomes ^37,38.

Coherent with our findings, Zerah and colleagues showed that the appropriate management of perioperative variables could play a significant role in the reduction of six-month mortality by up to 25% after hip fracture ³⁹.

Our findings indicated that the most common postoperative complication in our population was anemia requiring blood transfusion (81.7%). According to the literature, effective blood management involves preoperative measures, such as timely iron supplementation, to reduce the need for blood transfusions and complications during the hospital stay. This approach is relevant for all individuals, whether pre-frail or robust, but is especially critical for our older population with hip fractures, which is predominantly composed of frail patients ⁴⁰.

The prognostic impact of a vitamin D deficiency on hip fractures is poorly established. Some evidence has shown a correlation between hyperparathyroidism, vitamin D deficiency and increased 1-year mortality in hip-fractured patients ^41,42, while other studies did not replicate the same findings ^43,44. These clinical discrepancies may be the result of incorrect adjustments regarding comorbidity, frailty and functional status ⁴⁵. In our study, the deficiency of vitamin D in frail patients was associated with long-term mortality after hip fracture surgery, which may suggest the pleiotropic effects of vitamin D on osteosarcopenia and frailty progression ⁴⁶.

Early mobilization within 48 hours from surgery ⁴⁷ is a protective factor for overall survival. Even after adjustment for multimorbidity ^48,49, our findings showed that a delayed time of verticalization commencement was a prevalent barrier to restoring physical function. A series of factors may have contributed to this delay: low pre-morbid impaired mobility disability, older age, frailty status and loss of muscle mass and strength, which then prevented physiotherapists from fulfilling their rehabilitative goals.

Ensuring resilience to perioperative stressors in hip-fractured patients may be the key to delivering appropriate intervention and overcoming existing barriers to improving physical function and outcomes, especially in vulnerable frail patients. Frail older patients have intrinsic decreased functional reserve that can be altered by a multitude of perioperative stressors; future aetiologic studies are required to understand the causal mechanisms by which perioperative optimization of different variables may implement effective pathways of care, restraining excess mortality.

Strengths of this study are the prospective nature of the design and the long-term mortality assessment, including 21 months of observation with a suitable patient sample. The suitability of the sample size, confirmed by preliminary statistical evaluations conducted prior to the start of the study, is further supported by literature in which NHFS was as an effective tool for assessing death risk in elderly hip fracture patients ^50-52.

The methodologically robust frailty stratification allowed to adjust for complex multimorbidity, whereas the inclusion of a wide set of perioperative variables allowed the analysis of a wide spectrum of preventable surgical-related variables.

Limitations of the study are the single-center enrolment and the potential underestimation of confounding variables such as intraoperative blood loss and related blood management strategies, the invasiveness of surgical technique and duration, the type of anesthesia and the rate of intraoperative complications. Blood iron levels and types of pre-surgery anemia were not obtained to differentiate perioperative blood loss or iron deficiency anemia. The lack of standardization in outcome measurement and reporting practices in frail old age patients who were affected by hip fractures limited the ability to synthesize findings across studies.

CONCLUSIONS

Our research contributed to the understanding that targeted intervention of perioperative preventable variables, including the Nottingham Hip Fracture Score (NHFS), may have a great impact on the functional reserve of frail old age patients, shaping clinical trajectories and driving long-term outcomes such as excess mortality.

This preliminary study may serve as a platform for further clinical studies aimed at developing innovative pathways for the care of very old frail patients by fulfilling the growing and complex needs of an aging population and by creating new sustainable healthcare models.

Conflict of interest statement

The authors declare no conflict of interest.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author contributions

AG, MP, AC, FM, CG: designed the study;AG: acquired the data; AS, SO, LT: performed the statistical analysis; AG: wrote the manuscript and reviewed the literature; EP: edited the manuscript; FM, AN: scientifically and critically revised the manuscript.

Ethical consideration

The study was approved by the local ethical committee.

The research was conducted ethically, with all study procedures being performed in accordance with the requirements of the World Medical Association’s Declaration of Helsinki.

Written informed consent was obtained from each participant/patient for study participation and data publication.

SUPPLEMENTARY MATERIAL

S2. Kaplan-Meier curves plotted to show the survival of the population after hospital discharge.

Figure S1.1. Kaplan-Meier curve showing OS after hospitalization.

Figure S1.1. Kaplan-Meier curve showing OS after hospitalization.

Figure S1.2. Kaplan-Meier curve showing OS between patients with ideal range of vitamin D vs deficiency levels.

Figure S1.2. Kaplan-Meier curve showing OS between patients with ideal range of vitamin D vs deficiency levels.

Figure S1.3. Kaplan-Meier curve showing OS between patients who have been verticalized vs those that have not.

Figure S1.3. Kaplan-Meier curve showing OS between patients who have been verticalized vs those that have not.

Figures and tables

Figure 1. Kaplan-Meier curve showing OS between patients with NHFS ≤ 5 (low-risk) vs patients with NHFS > 5 (high-risk).

Figure 1. Kaplan-Meier curve showing OS between patients with NHFS ≤ 5 (low-risk) vs patients with NHFS > 5 (high-risk).

Table I. Clinical phenotype of the patients

Age (yr)
Mean	85.7 years (SD: 6.7)
Range	65-102
Gender (%)
Female	254 (58.8%)
Living at home	394 (91.6%)
Falls in the previous year (n°)
Mean	0.82 (SD: 1.47)
Median	0 (range: 0-10)
Drugs (n° [%])
≥ 5	259 (59.9%)
< 5	173 (40.1%)
SPMSQ median n° of errors (IQR) Median	3 (IQR: 5)
Frailty (n° impaired domains of CGA [%])
≥ 3	306 (70.8%)
Handgrip (kg)
Median	12.2 (IQR: 9.7-15.7)
Impaired	383 (88.7%)
Barthel index
Median	80 (IQR: 50-95)
< 50	90 (20.8%).
Preserved IADL
Median	3 (IQR: 5)
Functional dependence	235 (54.3%).
MNA
Median	10 (IQR: 8-12)
< 12	278 (64.4%)
CIRS (multimorbidity) Median	4 (IQR: 3-5).
Perioperative variables
25OH vitamin D levels (ng/mL)
Median	11.4 (IQR: 4.9-22)
Severe deficiency < 10	202 (46.8%)
Deficiency 10-30	102 (23.6%)
Insufficiency < 30	72 (16.7%)
Ideal range 50-100	56 (13%)
Fracture type
Lateral	222 (51.6%)
Medial	171 (39.8%)
Other	37 (8.6%)
Missing	2
Surgical waiting time since ER admission (days)
Median	2 (IQR: 1-3; range: 0-11)
Verticalization time (days)
Median	1 (IQR: 1-2; range: 0-18)
Post operative complications
Delirium	128 (29.6%)
Anaemia requiring blood transfusion	353 (81.7%)
Acute renal failure and/or urinary tract infections	67 (15.5%)
Pneumological (such as pulmonary embolism)	84 (19.4%)
Cardiologic	81 (18.8%)
Gastrointestinal	18 (4.2%)
Other	12 (3.5%)
Hospital discharge
Home	134 (31%)
Nursing home	165 (38.2%)
Rehabilitation centre	87 (20.1%)
Other ward	20 (4.6%)
SPMSQ: Short Portable Mental Status Questionnaire; IADL: Instrumental Activities of Daily Living; MNA: Mini-Nutritional Assessment; CIRS: Cumulative Illness Rating Scale: CGA: Comprehensive Geriatric Assessment.

Table I. Clinical phenotype of the patients

Age (yr)
Mean	85.7 years (SD: 6.7)
Range	65-102
Gender (%)
Female	254 (58.8%)
Living at home	394 (91.6%)
Falls in the previous year (n°)
Mean	0.82 (SD: 1.47)
Median	0 (range: 0-10)
Drugs (n° [%])
≥ 5	259 (59.9%)
< 5	173 (40.1%)
SPMSQ median n° of errors (IQR) Median	3 (IQR: 5)
Frailty (n° impaired domains of CGA [%])
≥ 3	306 (70.8%)
Handgrip (kg)
Median	12.2 (IQR: 9.7-15.7)
Impaired	383 (88.7%)
Barthel index
Median	80 (IQR: 50-95)
< 50	90 (20.8%).
Preserved IADL
Median	3 (IQR: 5)
Functional dependence	235 (54.3%).
MNA
Median	10 (IQR: 8-12)
< 12	278 (64.4%)
CIRS (multimorbidity) Median	4 (IQR: 3-5).
Perioperative variables
25OH vitamin D levels (ng/mL)
Median	11.4 (IQR: 4.9-22)
Severe deficiency < 10	202 (46.8%)
Deficiency 10-30	102 (23.6%)
Insufficiency < 30	72 (16.7%)
Ideal range 50-100	56 (13%)
Fracture type
Lateral	222 (51.6%)
Medial	171 (39.8%)
Other	37 (8.6%)
Missing	2
Surgical waiting time since ER admission (days)
Median	2 (IQR: 1-3; range: 0-11)
Verticalization time (days)
Median	1 (IQR: 1-2; range: 0-18)
Post operative complications
Delirium	128 (29.6%)
Anaemia requiring blood transfusion	353 (81.7%)
Acute renal failure and/or urinary tract infections	67 (15.5%)
Pneumological (such as pulmonary embolism)	84 (19.4%)
Cardiologic	81 (18.8%)
Gastrointestinal	18 (4.2%)
Other	12 (3.5%)
Hospital discharge
Home	134 (31%)
Nursing home	165 (38.2%)
Rehabilitation centre	87 (20.1%)
Other ward	20 (4.6%)
SPMSQ: Short Portable Mental Status Questionnaire; IADL: Instrumental Activities of Daily Living; MNA: Mini-Nutritional Assessment; CIRS: Cumulative Illness Rating Scale: CGA: Comprehensive Geriatric Assessment.

Table II. Univariate analysis of the examined clinical and perioperative variables between long-term survivors and non survivors.

	Survivors (n = 340)	Non-survivors (n = 92)	HR (95% CI); p-value	c-index
	85.6 (6.9)	86.0 (6.5)	1.01 (0.98-1.04); p = 0.52	0.534
	201 (59.1)	53 (57.6)	0.96 (0.63-1.46); p = 0.86	0.500
Fracture type				0.535
Lateral	174 (51.4)	48 (52.7)	1.00 (ref)
Medial	131 (38.6)	40(44)	1.10 (0.72-1.68); p = 0.65
Other	34(10)	3 (3.3)	0.36 (0.11-1.14); p = 0.083
Surgery waiting time (days)	1.91 (1.27)	2.11 (1.93)	1.08 (0.96-1.23); p = 0.21
Verticalization
Not done	30 (8.8)	17 (18.5)	1.00 (ref)	0.520
Done	310 (91.2)	75 (81.5)	0.42 (0.25-0.71); p = 0.001
Handgrip	13.1 (5.5) 38 (11.2)	12.3 (5.0)	0.97 (0.93-1.01); p = 0.15	0.547
Normal	302 (88.8)	11(12)	1.00 (ref)
Pathological		81(88)	0.94 (0.50-1.77); p = 0.86
Barthel	74.7(24)	67.8 (23.3)	0.90 (0.83-0.97); p = 0.008	0.595
80-100	190 (55.9)	39 (42.4)	1.00 (ref)
60-79	59 (17.4)	17 (18.5)	1.41 (0.80-2.49); p = 0.24
< 60	91 (26.8)	36 (39.1)	1.85 (1.18-2.92); p = 0.008
MNA	10.1 (2.9)	9.6 (3.0)	0.94 (0.87-1.00); p = 0.056	0.554
NHFS			1.29 (1.09-1.53); p = 0.004	0.603
≤ 4	5.7 (1.2)	6.1 (1.1)	1.00 (ref)
> 4	49 (14.4)	7 (7.6)	1.91 (0.88-4.13); p = 0.10
≤ 5	291 (85.6)	85 (92.4)	1.00 (ref)
> 5			2.04 (1.29-3.24); p = 0.002	0.592
CIRS (multimorbidity)	3.9 (1.7)	4.4 (1.7)	1.15 (1.03-1.29); p = 0.016	0.573
25OH vitamin D
Ideal range	50 (14.7)	6 (6.5)	1.00 (ref)
Deficiency (any level)	290 (85.3)	86 (93.5)	2.36 (1.03-5.42); p = 0.041	0.54
IADL: Instrumental Activities of Daily Living; MNA: Mini-Nutritional Assessment; CIRS: Cumulative Illness Rating Scale; NFHS: Nottingham Hip Fracture Score; Ref: reference category; HR: hazard-ratio; for continuous variables. HR was reported for an increase of 1 unit on the corresponding scale except where differently reported.

Table II. Univariate analysis of the examined clinical and perioperative variables between long-term survivors and non survivors.

	Survivors (n = 340)	Non-survivors (n = 92)	HR (95% CI); p-value	c-index
	85.6 (6.9)	86.0 (6.5)	1.01 (0.98-1.04); p = 0.52	0.534
	201 (59.1)	53 (57.6)	0.96 (0.63-1.46); p = 0.86	0.500
Fracture type				0.535
Lateral	174 (51.4)	48 (52.7)	1.00 (ref)
Medial	131 (38.6)	40(44)	1.10 (0.72-1.68); p = 0.65
Other	34(10)	3 (3.3)	0.36 (0.11-1.14); p = 0.083
Surgery waiting time (days)	1.91 (1.27)	2.11 (1.93)	1.08 (0.96-1.23); p = 0.21
Verticalization
Not done	30 (8.8)	17 (18.5)	1.00 (ref)	0.520
Done	310 (91.2)	75 (81.5)	0.42 (0.25-0.71); p = 0.001
Handgrip	13.1 (5.5) 38 (11.2)	12.3 (5.0)	0.97 (0.93-1.01); p = 0.15	0.547
Normal	302 (88.8)	11(12)	1.00 (ref)
Pathological		81(88)	0.94 (0.50-1.77); p = 0.86
Barthel	74.7(24)	67.8 (23.3)	0.90 (0.83-0.97); p = 0.008	0.595
80-100	190 (55.9)	39 (42.4)	1.00 (ref)
60-79	59 (17.4)	17 (18.5)	1.41 (0.80-2.49); p = 0.24
< 60	91 (26.8)	36 (39.1)	1.85 (1.18-2.92); p = 0.008
MNA	10.1 (2.9)	9.6 (3.0)	0.94 (0.87-1.00); p = 0.056	0.554
NHFS			1.29 (1.09-1.53); p = 0.004	0.603
≤ 4	5.7 (1.2)	6.1 (1.1)	1.00 (ref)
> 4	49 (14.4)	7 (7.6)	1.91 (0.88-4.13); p = 0.10
≤ 5	291 (85.6)	85 (92.4)	1.00 (ref)
> 5			2.04 (1.29-3.24); p = 0.002	0.592
CIRS (multimorbidity)	3.9 (1.7)	4.4 (1.7)	1.15 (1.03-1.29); p = 0.016	0.573
25OH vitamin D
Ideal range	50 (14.7)	6 (6.5)	1.00 (ref)
Deficiency (any level)	290 (85.3)	86 (93.5)	2.36 (1.03-5.42); p = 0.041	0.54
IADL: Instrumental Activities of Daily Living; MNA: Mini-Nutritional Assessment; CIRS: Cumulative Illness Rating Scale; NFHS: Nottingham Hip Fracture Score; Ref: reference category; HR: hazard-ratio; for continuous variables. HR was reported for an increase of 1 unit on the corresponding scale except where differently reported.

Table III. Multivariate analysis for OS, including clinical and perioperative variables.

	HR (95% CI); p-value	c-index
Verticalization		0.631
On time	1.00 (ref)
Delayed	2.31 (1.36-3.92); p = 0.002
NHFS
≤ 5	1.00 (ref)
> 5	1.96 (1.24-3.10); p = 0.004
25OH vitamin D
Ideal range	1.00 (ref)
Deficiency (any level)	2.31 (1.01-5.29); p = 0.048
Ref: reference category; HR: hazard-ratio; for continuous variables, HR was reported for an increase of 1 unit on the corresponding scale except where differently reported.

Table III. Multivariate analysis for OS, including clinical and perioperative variables.

	HR (95% CI); p-value	c-index
Verticalization		0.631
On time	1.00 (ref)
Delayed	2.31 (1.36-3.92); p = 0.002
NHFS
≤ 5	1.00 (ref)
> 5	1.96 (1.24-3.10); p = 0.004
25OH vitamin D
Ideal range	1.00 (ref)
Deficiency (any level)	2.31 (1.01-5.29); p = 0.048
Ref: reference category; HR: hazard-ratio; for continuous variables, HR was reported for an increase of 1 unit on the corresponding scale except where differently reported.