Can N-terminal pro-B-type natriuretic peptide concentration on admission predict the functional independence measure at discharge in patients aged 75 years or older with postoperative proximal femur fracture?

Abstract

[Purpose] To examine whether functional independence measure scores and effectiveness at discharge can be predicted using N-terminal pro-B-type natriuretic peptide concentrations at admission in postoperative patients aged ≥75 years with proximal femur fractures. [Participants and Methods] This study included 35 patients who were admitted for rehabilitation after proximal femur fracture surgery between April 1, 2020 and September 20, 2023 and were discharged by November 30, 2023. The primary outcomes were the functional independence measure scores and effectiveness at discharge. The explanatory variables analyzed using multiple regression included demographic data; N-terminal pro-B-type natriuretic peptide concentration, estimated glomerular filtration rate, and geriatric nutritional risk index at admission; functional ambulation categories before injury; and motor, cognitive, and total functional independence measure scores at admission. [Results] The motor functional independence measure score at admission and N-terminal pro-B-type natriuretic peptide concentration were significant explanatory variables for the motor functional independence measure score at discharge. The N-terminal pro-B-type natriuretic peptide concentration was a significant explanatory variable in total and motor functional independence effectiveness. [Conclusion] This study, which excluded cognitively impaired patients and focused on individuals aged ≥75 years, suggests that N-terminal pro-B-type natriuretic peptide concentration at admission affects the functional independence measure scores and effectiveness at discharge.

INTRODUCTION

The global population is aging at an unprecedented rate, with the proportion of older adults expected to rise from 9.4% in 2020 to 18.7% by 2042¹⁾. This trend is particularly pronounced in Japan, where the aging population has remarkably increased. As of 2022, 29.0% of the Japanese population was aged 65 or older, and 15.5% was aged 75 or older, marking the highest aging rate in the world¹⁾. Bone mass peaks in one’s 20s, after which a balance between bone resorption and formation is maintained for a period. However, from around the age of 40²⁾, bone mass begins to decline, increasing the risk of fractures³⁾. In Japan, the number of proximal femur fractures is projected to reach approximately 300,000 annually by 2030 and 320,000 by 2042, largely due to the aging population⁴⁾.

Aging is also associated with an increased incidence of heart failure. A U.S. study reported that the incidence of new heart failure cases rises with age, affecting 1.3% of men and 0.7% of women in their 60s, and 8.3% of men and 7.8% of women in their 80s⁵⁾. The global aging population has contributed to what is now termed a “heart failure pandemic”, with 26 million people worldwide affected⁶⁾. In Japan, the number of patients with chronic heart failure is estimated to rise significantly due to the aging population and the increasing prevalence of ischemic heart disease. This number is expected to increase from approximately 1.1 million in 2015 to 1.3 million by 2030, continuing to increase until 2035⁷⁾.

Given the increasing aging population in Japan, the number of patients with heart failure undergoing proximal femur fracture surgery will rise. Consequently, establishing evidence-based practices is necessary for managing these complicated cases in the future.

There is some literature describing patients with complicated heart failure after proximal femur fracture surgery. Itagaki et al.⁸⁾ reported that functional independence measure (FIM^TM) effectiveness was lower in those with a history of heart failure compared with those without such a diagnosis. Tamamura et al.⁹⁾ found that FIM^TM gain and motor FIM^TM effectiveness were lower in patients over 65 years of age with elevated B-type natriuretic peptide (BNP) levels. In addition, they reported that this trend was particularly pronounced in patients with BNP concentrations of 100 pg/mL or higher¹⁰⁾. However, these reports did not exclude patients with cognitive impairments, leaving open the possibility that cognitive decline may influence ADL recovery. Additionally, these studies either lacked age criteria or included a wide age range (65 years and older). Given the correlation between age and the severity of heart failure, distinguishing between ADL capacity decline due to aging and that due to heart failure is difficult.

Therefore, this study aimed to investigate whether FIM^TM scores and FIM^TM effectiveness at discharge could be predicted using N-terminal pro-B-type natriuretic peptide (NT-proBNP) concentrations at admission. The goal was to obtain basic data for predicting ADL recovery in elderly postoperative patients, aged 75 years or older, with proximal femur fractures and no cognitive impairment undergoing rehabilitation.

PARTICIPANTS AND METHODS

This study is a retrospective cohort study that examined predictors of FIM^TM at discharge and FIM^TM effectiveness using multiple regression analysis.

The target sample size was estimated in advance using G*Power 3.1.9.7 (Heinrich Heine University, Düsseldorf, Germany) (Effect size f²=0.5, α=0.05, Power=0.8, Number of predictors=8). Consequently, 39 participants were targeted for inclusion in the study. This study included patients who were admitted to Hikarigaoka Hospital for rehabilitation between April 1, 2020 and September 20, 2023 and were discharged by November 30 of the same year. The participants were patients aged 75 years or older with a first-time postoperative proximal femur fracture. Their plasma NT-proBNP concentrations were measured upon admission due to suspected heart failure. Out of 98 patients initially considered, 35 were ultimately included in the study. Patients were excluded if they had passed away in the hospital (n=1), were transferred to another hospital (n=4), had missing data (n=11), or scored 20 or lower (n=47) on the revised Hasegawa’s dementia scale (HDS-R)¹¹⁾.

The final sample consisted of 29 women and 6 men, with a mean age of 87.3 ± 6.4 years. Among these patients, 18 had a femoral neck fracture, and 17 had a trochanteric fracture. Regarding surgical intervention, 19 patients underwent osteosynthesis, while 16 underwent arthroplasty. This retrospective study evaluated and measured results within the context of routine medical care. All study participants provided informed consent, and the study design was approved by the appropriate Ethical Review Committee of Hikarigaoka Hospital (Acceptance No. 14) and the Kinjo University Ethical Review Committee (Notification No. 2024-05).

Referring to previous studies^{8, 12, 13)} on the effect of proximal femur fracture surgery on ADL, we obtained basic information from medical records, including patient age, sex, disease name, length of hospitalization, and the interval between injury onset and hospital admission. Additionally, we gathered information on pre-injury functional ambulation categories (FAC), NT-proBNP concentrations, estimated glomerular filtration rate, cognitive function, and the geriatric nutritional risk index (GNRI).

The ability to perform ADL was assessed using FIM^TM, which was then used to calculate FIM^TM effectiveness. Total FIM^TM effectiveness was calculated as total FIM^TM gain / (126 points −total FIM^TM at admission), motor FIM^TM effectiveness as motor FIM^TM gain / (91 points −exercise FIM^TM at admission), and cognitive FIM^TM effectiveness as cognitive FIM^TM gain / (35 points −cognitive FIM^TM at admission). The denominator represents the potential for improvement, while the numerator indicates the actual improvement, with effectiveness values ranging from 0 to 1¹⁴⁾. FIM^TM measurements were conducted by the assigned physical or occupational therapist.

Cognitive function was assessed using the HDS-R¹¹⁾, with a score of 21 or higher considered normal and a score of 20 or lower indicating suspected dementia. In this study, a score of 20 or lower was considered as cognitive decline.

FAC was developed by Holden et al.¹⁵⁾ as a method to assess walking ability across six levels. This classification does not consider the use of walking aids, such as braces and canes. Pre-fracture gait ability has been reported to affect FIM^TM effectiveness in patients recovering from hip fractures⁸⁾. In this study, the FAC score before injury was obtained from medical records.

NT-proBNP is a circulating hormone biosynthesized and secreted by cardiac myocytes. It plays a role in diuresis, natriuresis, relaxation of vascular smooth muscle, and inhibition of aldosterone secretion¹⁶⁾. It is secreted by the myocardium in response to increased cardiac preload and myocardial stretching¹⁷⁾. NT-proBNP concentration increases in proportion to the severity of both left and right heart failure¹⁸⁾. According to the “Guidelines for the Diagnosis of Acute and Chronic Heart Failure”¹⁹⁾, the measurement of NT-proBNP and BNP levels is strongly recommended (class I) for diagnosing, determining the severity, and prognostic evaluation of heart failure. Additionally, it is considered class IIa for guiding treatment decisions and screening for heart failure.

The GNRI is a tool used to assess the nutritional status of elderly individuals. It is calculated using the formula: 14.89 × Alb level (g/dL) + 41.7 × (current weight/standard weight). The standard weight is determined differently for men and women: for men, it is calculated as height −100 −(height −150)/4, and for women, as height −100 −(height −150)/2.5. If the current weight exceeds the standard weight, the ratio of the current weight to the standard weight is set to 1²⁰⁾. The GNRI is suitable in predicting mortality and evaluating the risk of undernutrition in the elderly²⁰⁾.

During the hospitalization period, the participants underwent rehabilitation sessions lasting 20 to 60 min, once or twice daily, 5 to 7 days a week. The rehabilitation program, designed by the physical or occupational therapist under the guidance of a physiatrist, included exercises to improve joint range of motion, muscle strengthening, gait training (using parallel bars, walkers, canes, or unassisted walking), and activities of daily living. The timing of discharge was decided by consensus of the multidisciplinary staff during a conference after the attending physician gave permission for discharge.

The FIM^TM score at discharge (total, motor, and cognitive items) and FIM^TM effectiveness (total, motor, and cognitive items) were used as dependent variables. Each assessment item (NT-proBNP concentration at admission, age, period from injury to hospitalization, length of hospital stay, estimated glomerular filtration rate at admission, GNRI at admission, FAC before injury, motor FIM^TM at admission, cognitive FIM^TM at admission, and total FIM^TM at admission) was used as an explanatory variable. Multiple regression analysis was conducted using the stepwise method. Prior to the analysis, multicollinearity was assessed to ensure that the correlation coefficients between explanatory variables did not exceed 0.9.

NT-proBNP concentrations were classified into four groups based on the cut-off values for the diagnosis of heart failure in the “Guidelines for the Diagnosis of Acute and Chronic Heart Failure”¹⁹⁾ issued by the Japanese Society of Cardiology and the Japan Heart Failure Society: 1) less than 125 pg/mL, 2) 125 pg/mL to less than 400 pg/mL, 3) 400 pg/mL to less than 900 pg/mL, and 4) 900 pg/mL and above. These categories were used as dummy variables. SPSS (version 29) (IBM Corp., Armonk, NY, USA) was used for statistical analyses, and the significance level was set at 5%.

RESULTS

The values of each evaluation item are shown in Table 1. Tables 2, 3, 4 shows the results of a multiple regression analysis examining the relationship between FIM^TM scores at discharge and each of the assessment items. In the total FIM^TM score at discharge, the total FIM^TM score at admission was a significant explanatory variable (p<0.01, β=0.86). The motor FIM^TM score at admission (p<0.01, β=0.79) and NT-proBNP concentration (p=0.02, β= −0.21) were significant explanatory variables in the motor FIM^TM score at discharge. Furthermore, the cognitive FIM^TM score at admission was a significant explanatory variable in the cognitive FIM^TM score at discharge (p<0.01, β=0.72). Tables 5, 6 shows the results of a multiple regression analysis exploring the relationship between FIM^TM effectiveness and each of the assessment items. NT-proBNP concentration was a significant explanatory variable in both total FIM^TM effectiveness and motor FIM^TM effectiveness (p<0.01, β=−0.45; p<0.01, β=−0.50). However, no significant results were obtained for cognitive FIM^TM effectiveness using the stepwise method.

Table 1. Evaluation item values.

Evaluation items		Values
Sex (n)	Male	6
	Female	29
Age (years)		87.3 ± 6.4
Fracture type (n)	Femoral neck fracture	18
	Trochanteric fracture	17
Type of surgery (n)	Osteosynthesis	19
	Arthroplasty	16
Period from injury to hospitalization (days)		25.3 ± 10.2
Length of hospital stay (days)		46.8 ± 21.9
NT-proBNP concentration at admission	<125 pg/mL	8
	≥125 pg/mL, <400 pg/mL	13
	≥400 pg/mL, <900 pg/mL	9
	≥900 pg/mL	5
eGFR at admission (mL/min/1.73 m2)		67.3 ± 18.9
GNRI at admission		94.8 ± 7.9
HDS-R at admission (score)		25.9 ± 2.7
FAC before injury (n)	4	12
	5	23
Total FIMTM at admission (score)		78.5 ± 18.7
Motor FIMTM at admission (score)		48.9 ± 15.6
Cognitive FIMTM at admission (score)		28.6 ± 5.1
Total FIMTM at discharge (score)		99.7 ± 14.5
Motor FIMTM at discharge (score)		68.9 ± 12.2
Cognitive FIMTM at discharge (score)		30.8 ± 4.3
Total FIMTM effectiveness		0.46 ± 0.16
Motor FIMTM effectiveness		0.48 ± 0.16
Cognitive FIMTM effectiveness		0.29 ± 0.36

eGFR: estimated glomerular filtration rate; GNRI: geriatric nutritional risk index; HDS-R: Hasegawa’s dementia scale; FAC: functional ambulation categories; FIM^TM: functional independence measure.

Table 2. Results of multiple regression analysis with Total FIM^TM scores at discharge as the dependent variable.

	Unstandardized coefficient	Standardized coefficient	p-value	95% CI		VIF
				Lower limit	Upper limit
Total FIMTM at discharge
Constant	47.21		<0.001	36.09	58.33
Total FIMTM at admission	0.67	0.86	<0.001	0.53	0.81	1.00
	R2=0.75	ANOVA p<0.001

ANOVA: analysis of variance; FIM^TM: functional independence measure; VIF: variance inflation factor.

Table 3. Results of multiple regression analysis with Motor FIM^TM scores at discharge as the dependent variable.

	Unstandardized coefficient	Standardized coefficient	p-value	95% CI		VIF
				Lower limit	Upper limit
Motor FIMTM at discharge
Constant	44.05		<0.001	34.11	53.98
Motor FIMTM at admission	0.62	0.79	<0.001	0.48	0.76	1.13
NT-proBNP concentration	−2.56	−0.21	0.021	−4.72	−0.41	1.13
	R2=0.79	ANOVA p<0.001

ANOVA: analysis of variance; FIM^TM: functional independence measure; VIF: variance inflation factor.

Table 4. Results of multiple regression analysis with Cognitive FIM^TM scores at discharge as the dependent variable.

	Unstandardized coefficient	Standardized coefficient	p-value	95% CI		VIF
				Lower limit	Upper limit
Cognitive FIMTM at discharge
Constant	13.46		<0.001	7.37	19.54
Cognitive FIMTM at admission	0.61	0.72	<0.001	0.40	0.81	1.00
	R2=0.51	ANOVA p<0.001

ANOVA: analysis of variance; FIM^TM: functional independence measure; VIF: variance inflation factor.

Table 5. Results of multiple regression analysis with Total FIM^TM effectiveness as the dependent variable.

	Unstandardized coefficient	Standardized coefficient	p-value	95% CI		VIF
				Lower limit	Upper limit
Total FIMTM effectiveness
Constant	0.63		<0.001	0.50	0.76
NT-proBNP concentration	−0.07	−0.45	0.006	−0.13	−0.02	1.00
	R2=0.20	ANOVA p=0.006

ANOVA: analysis of variance; FIM^TM: functional independence measure; VIF: variance inflation factor.

Table 6. Results of multiple regression analysis with Motor FIM^TM effectiveness as the dependent variable.

	Unstandardized coefficient	Standardized coefficient	p-value	95% CI		VIF
				Lower limit	Upper limit
Motor FIMTM effectiveness
Constant	0.66		<0.001	0.54	0.78
NT-proBNP concentration	−0.08	−0.50	0.002	−0.13	−0.03	1.00
	R2=0.25	ANOVA p=0.002

ANOVA: analysis of variance; FIM^TM: functional independence measure; VIF: variance inflation factor.

DISCUSSION

In this study, we investigated whether admission NT-proBNP concentration could predict discharge FIM^TM outcomes among elderly postoperative patients with proximal femur fractures aged 75 years or older. Results showed that NT-proBNP concentrations at admission were predictive of motor FIM^TM scores at discharge, as well as total FIM^TM effectiveness and motor FIM^TM effectiveness. A study describing the effect of heart failure on FIM^TM effectiveness after hip fracture reported lower FIM^TM effectiveness in patients with a history of heart failure than in those without⁸⁾. Another study on patients with hip fractures aged 65 years and older found that higher BNP concentrations are associated with lower FIM^TM scores and FIM^TM effectiveness at discharge⁹⁾. Specifically, patients with BNP concentrations above 100 pg/mL demonstrate significantly lower FIM^TM gains and FIM^TM effectiveness¹⁰⁾. Consistent with these findings, the present study suggests that NT-proBNP concentration at admission affects FIM^TM effectiveness and FIM^TM scores at discharge.

The first difference from previous studies is that this study excluded patients with cognitive decline. Research indicates that patients with post-fracture femur who have MMSE scores of 20 or lower tend to show poor improvement in motor FIM^{TM 21)} and reduced ADL independence, particularly among those with dementia, aged 85 or older, and with multiple comorbidities²²⁾. This suggests that cognitive impairment may have a considerable impact on improvement in ADL after femur fracture surgery. The second difference lies in the age criteria, where this study focused on patients aged 75 years and older. Prior studies either did not set specific age criteria or included a wide age range starting from 65 years, suggesting that age might have contributed to the decline in ADL capacity. In this study, we excluded patients with cognitive impairment (HDS-R score of 20 or less) and limited the age to 75 years and older, which we believe suppressed the effects of cognitive function and age, thereby providing a more accurate assessment of the impact of NT-proBNP concentration on FIM^TM scores at discharge.

This study, which specifically excluded dementia in elderly patients aged 75 years and older, also suggested that heart failure severity was associated with motor scores at discharge and motor FIM^TM effectiveness.

We believe that the NT-proBNP concentration was selected as an explanatory factor for the motor FIM^TM score at discharge, total FIM^TM effectiveness, and motor FIM^TM effectiveness in this study for several reasons. First, this choice is rooted in the pathophysiology of heart failure. In heart failure, elevated left ventricular end-diastolic pressure and decreased cardiac output are known to promote the secretion of natriuretic peptides^{23, 24)}. Elevated left ventricular end-diastolic pressure, in turn, causes increased left atrial pressure and pulmonary artery wedge pressure²⁵⁾. Obokata et al.²⁶⁾ reported that higher pulmonary capillary wedge pressure in patients with heart failure leads to hyperventilation and an increased respiratory rate during exercise compared with those without heart failure. In this study, NT-proBNP concentrations at admission were classified into four groups based on the cutoff values established by the “Guidelines for the Diagnosis of Acute and Chronic heart failure”¹⁹⁾ by the Japanese Society of Cardiology and the Japan Heart Failure Society. However, elevated NT-proBNP concentrations may have caused hyperventilation during exercise, potentially hindering patients’ ability to achieve the adequate level of physical activity.

Second, poor exercise tolerance has been reported to significantly reduce ADL in patients with heart failure²⁷⁾. Peripheral factors, such as skeletal muscle function, are considered more important than central circulating factors, such as cardiac function, in determining exercise intolerance. Specifically, heart failure is associated with a decline in oxidative enzymes, a decrease in type I muscle fibers alongside a relative increase in type II fibers, reduced mitochondrial content, and muscle atrophy²⁸⁾. The association found in this study between NT-proBNP concentrations and motor FIM^TM score at discharge, as well as total FIM^TM effectiveness and motor FIM^TM effectiveness, may be due in part to the effect of heart failure on exercise intolerance.

Third, we considered the effect of heart failure on fracture and soft tissue healing. The general process of fracture healing involves three phases: an inflammatory phase, a repair phase where temporary bone is formed, and a remodeling phase²⁹⁾. Adequate blood circulation is considered important for bone heading³⁰⁾. According to Gurlt³⁰⁾, the average healing time for the femoral neck to tolerate exercise load is approximately 12 weeks. However, factors, such as older age, undernourishment, complications like diabetes, osteoporosis, and a history of steroid use, are believed to prolong bone healing³⁰⁾. In heart failure, increased sympathetic nerve activity and elevated levels of inflammatory cytokines accelerate bone resorption^31,32,33), resulting in decreased bone mineral density in the extremities and trunk compared with patients without heart failure³⁴⁾. This reduction in bone mineral density may have affected bone healing. In addition, post-fracture surgery involves the invasion of soft tissues, such as muscle and fascia, whose regeneration requires adequate blood circulation and an acute inflammatory response³⁵⁾. Thus, blood circulation is not only vital for bone healing but also for tissue repair^{30, 35)}. We can speculate that high NT-proBNP concentrations may have affected bone healing and soft tissue repair by reducing tissue perfusion due to decreased cardiac output, thus delaying the recovery of physical function. The results of this study suggest that physical therapists should focus on NT-proBNP concentrations because the recovery of motor FIM^TM scores in elderly postoperative proximal femur fracture patients without dementia is greatly influenced by heart failure severity.

This study had some limitations. First, it was conducted at a single center, which limits the generalizability of the findings. To build stronger evidence regarding the severity of heart failure in patients with postoperative hip joint, future research should include multicenter studies. Second, the health conditions of elderly patients are characterized by individual variability, multiple chronic diseases, atypical symptoms, and the presence of several comorbidities³⁶⁾. Although this study focused on heart failure severity, examining the comorbidities and complications that affect a patient’s ability to perform ADL is necessary. This approach would allow for more accurate predictions of the FIM^TM and functional recovery at discharge. Finally, heart failure severity in this study was assessed solely through NT-proBNP concentration. A more detailed examination could be achieved using cardiac echocardiographic measurements, such as left ventricular ejection fraction and the cardiothoracic ratio from chest X-rays, to classify patients into pathological states, such as systolic and diastolic dysfunction.

Funding

This research received no grant from any funding agency in the public, commercial or not-for-profit sectors.

Conflict of interest

The authors declare that there is no conflict of interest.