Depressive Symptoms Can Attenuate Rehabilitation Effects in Fracture Patients

ABSTRACT

Objectives:

This study aimed to clarify the relationship between depressive symptoms on admission and improvement in activities of daily living (ADLs) among patients with fractures admitted to a recovery rehabilitation unit.

Methods:

This retrospective observational study enrolled patients with fractures who were admitted to a rehabilitation hospital between August 2018 and October 2020 in Okinawa City, Japan. Based on Geriatric Depression Scale scores, patients were divided into two groups: one group with depressive symptoms and another group without depressive symptoms. Associations between depressive symptoms and the motor Functional Independence Measure (FIM) gain were identified using multivariate linear regression analyses.

Results:

Out of 127 patients (range, 40.0–83.0 years; male: 27.6%), 71 had depressive symptoms on admission. Multiple linear regression analysis revealed that depressive symptoms on admission were significantly associated with motor FIM gain (β coefficient: −5.022, 95% confidence interval −9.551 to −0.494, P = 0.030). Body Mass Index on admission, motor FIM on admission, Mini-Mental State Examination on admission, and use of antidepressants were also associated with motor FIM gain.

Conclusions:

Among patients with fractures admitted to a recovery rehabilitation unit, depressive symptoms on admission were associated with less improvement in ADLs. The use of antidepressants was also an independent factor affecting the improvement of ADLs in these patients.

INTRODUCTION

Bone fragility is observed in the elderly, with 53% of fractures in individuals aged over 50 years being attributed to bone fragility.¹⁾ Fragility fractures include hip and spine fractures,²⁾ and recent reports have described pelvic ring fractures.^3,4,5) With the increase in the aging population in Japan, the increasing number of patients with these fractures has become a social issue, along with the financial pressure of medical costs. In 2007, the incidence of hip fractures in Japan was 5.1 per 10,000 men and 18.1 per 10,000 women, representing approximately 130,000 patients. This number is 2.8 times the rate of the first survey in 1987 and is still increasing. Hip fractures increase the mortality risk by approximately 10%–20% in the year following the fracture.⁶⁾ Among patients aged 65 years and older with hip fractures, only approximately 30% regain their pre-fracture mobility and activities of daily living (ADLs) ability.⁷⁾ The prevalence of spine fractures is approximately 30% and 40% in women in their 70s and 80s, respectively.⁶⁾ Spine fractures can lead to morbidity rates comparable with or exceeding those of hip fractures.^8,9) Among adults aged 65 years or older, spine fractures are associated with a significantly reduced long-term survival rate, with a mortality risk six times higher than that of the general population.¹⁰⁾ Pelvic ring fractures have also been reported to have a high incidence and mortality rate.^11,12)

Depression in older individuals is a common mental disorder that affects their quality of life. Major depression occurs in 2% of adults aged 55 years and older and its prevalence increases with age. Additionally, 10%–15% of older adults have depressive symptoms, even in the absence of major depression.¹³⁾ Although depression is the most common mental health problem among the elderly, it is often underdiagnosed and undertreated.¹⁴⁾ Studies involving community-dwelling older adults have demonstrated that depressive symptoms are associated with subsequent decline in physical functioning.^15,16,17)

The association between fractures and depression has been reported, with 9% to 47% of patients with hip fractures having depression,¹⁸⁾ and most patients with fragility fractures having depression.¹⁹⁾ The impact of depressive symptoms on rehabilitation after hip fracture surgery has been evaluated repeatedly. In one study, moderate to severe depressive symptoms were associated with adverse events, such as low improvement in ambulation after rehabilitation, death after 1 year, and institutionalization.²⁰⁾ However, another study reported no significant association between depressive symptoms and rehabilitation effectiveness.^21,22) One possible explanation for these discrepancies is the variation in assessment methods for depressive symptoms, because different studies employ different scales and cutoff values.²⁰⁾ In addition, differences in rehabilitation intensity and length of hospital stay may also have an impact. Furthermore, the incidence of fragility fractures is increasing in accordance with the increase in average life expectancy.²³⁾

Although depressive symptoms comorbid with fractures may attenuate rehabilitation effectiveness, most previous studies have focused on patients with hip fractures, whereas research on other types of fractures, such as spine and pelvic ring fractures, remains extremely limited. Therefore, examining the impact of depressive symptoms on rehabilitation effectiveness in a broader population is essential. Clarifying the relationship between comorbid depressive symptoms and improvement in ADLs during hospitalization of patients with fractures is important. This understanding may facilitate the prediction of prognosis during rehabilitation and lead to appropriate therapeutic interventions. Specifically, it is necessary to examine the impact of depressive symptoms on improvement in ADLs, focusing on patients with fractures other than hip fractures, such as those of the spine and pelvic ring, which have been under-researched. To address this gap, this study aimed to examine the relationship between depressive symptoms on admission and improvement in ADLs in patients with fractures admitted to a rehabilitation unit.

MATERIALS AND METHODS

Study Design and Participants

This longitudinal study was conducted between August 2018 and October 2020 at a rehabilitation hospital in Okinawa City, Japan. All patients admitted for treatment of fractures were consecutively enrolled in this study. The following patients were excluded: (1) those discharged because of the onset of acute diseases; (2) those with no data on the Geriatric Depression Scale (GDS) because of examination refusal or other reasons; and (3) those with missing data.

Rehabilitation Program

In the rehabilitation wards, all participants underwent daily rehabilitation as instructed by medical doctors and through face-to-face sessions with therapists. The program included muscle strength and ADLs training, as well as sitting and standing training, balance training, gait exercises, stepping exercises, and training to improve their ability to live their daily lives. Depending on individual needs, participants used corsets, orthoses, and canes while exercising or whenever necessary. Depending on their condition, some participants underwent rehabilitation with only partial or no body weight load on the fractured part. The participants were advised not to perform rehabilitation activities when their circulation or breathing was impaired, or if their condition worsened.

Ethical Considerations

This study was approved by the Institutional Review Board of the hospital (approval ID: 21–32). Regardless of whether they could be included in this study, all patients were required to undergo GDS assessment in this hospital to evaluate the effectiveness of the rehabilitation programs. Given the retrospective design of the study, an opt-out procedure was used to give all participants the opportunity to exclude their data from the study analysis. All experimental procedures were performed following the principles of the Declaration of Helsinki.

Data Collection

Clinical data on admission, including age, sex, Body Mass Index (BMI), Mini Nutritional Assessment Short Form (MNA-SF), Charlson Comorbidity Index (CCI), Mini-Mental State Examination (MMSE), GDS, Functional Independence Measure (FIM), and fracture type, were retrospectively collected from the clinical database. The FIM data were obtained on admission and at discharge to determine the FIM gain. In addition, information regarding the length of hospital stay for rehabilitation, the period of rehabilitation per day, and the use of antidepressants were obtained from the clinical database. The BMI was calculated as the weight in kilograms divided by the height in meters squared. The nutritional status was assessed using the MNA-SF, which provides a score that may range from 0 to 14. Nutritional status was categorized according to the following score cutoffs: 12 or greater, normal nutritional status; 8–11, risk of malnutrition; less than 8, malnutrition.²⁴⁾ The CCI remains the most widely validated and used comorbidity score and includes 17 conditions, each assigned a score from 1 to 6 based on severity.^25,26,27) The MMSE has been a standard tool for evaluating cognitive impairment. It contains 11 items that assess five areas of cognitive function: orientation, registration, attention and calculation, recall, and language. MMSE scores range from 0 to 30, with lower scores indicating more severe cognitive impairment.^28,29) The ability to perform ADLs upon admission and discharge was evaluated using the motor FIM, a subscale of the FIM that includes four domains: self-care, sphincter control, transfers, and locomotion. The FIM consists of 18 items (13 in motor domain and 5 in cognitive domain). Each item was scored from 1 (total dependence) to 7 (complete independence).³⁰⁾ The amount of energy intake was obtained from records of oral, intravenous, or enteral nutrition. Nurses or registered dieticians evaluated the remaining energy intake after each meal and calculated the energy intake accordingly.

Symptoms of Depression

The presence of depressive symptoms was assessed on admission using the 15-item GDS.^31,32) This scale was translated into Japanese. A validation study was then conducted on Japanese participants. As a result, the Japanese version of the GDS proved to be a clinically applicable screening instrument for depression, regardless of age or sex.^33,34) The GDS is a commonly used screening tool in geriatric settings, with higher scores indicating a worse status. A GDS score between 0 and 5 denoted the absence of depressive symptoms, scores between 6 and 9 indicated mild depressive symptoms, and scores between 10 and 15 indicated moderate to severe depressive symptoms.³⁵⁾ Based on previous research,³⁵⁾ the patients were divided into two groups: a group with depressive symptoms and a group without depressive symptoms. The group with depressive symptoms included patients with scores between 6 and 15 on the GDS, whereas the group without depressive symptoms included patients with scores between 0 and 5 on the GDS. The GDS was assessed upon admission by trained therapists.

Statistical Analysis

All statistical analyses were performed using EZR (version 1.61; Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria). Specifically, this is a modified version of the R commander designed to add statistical functions frequently used in biostatistics. All continuous variables were presented as a mean ± standard deviation or medians with interquartile ranges, and categorical variables were expressed as the number of patients and percentages. A descriptive analysis was conducted. The normality of continuous variables was determined using the Shapiro–Wilk normality test; normally distributed variables were compared using the t-test, and skewed variables were compared using the Mann–Whitney U test. A Chi-square test was performed for categorical variables. In addition, univariate and multivariate linear regression analyses were performed to identify associations with motor FIM gain. Variables with P < 0.2 in the univariate linear regression analysis were used for inclusion in the multivariate linear regression analysis as possible predictors of motor FIM gain. For all comparisons, the statistical significance level was set at P < 0.05.

RESULTS

A total of 183 consecutive patients with fractures were initially eligible for enrollment in this study, of whom 56 met the exclusion criteria (Fig. 1). Of these, 6 participants were discharged because of the onset of acute diseases, and 50 participants had no GDS data on admission. A total of 127 patients were included in the final analysis and divided into two groups: those with depressive symptoms (n=71) and those without depressive symptoms (n=56).

Fig. 1.

 Flowchart of participant recruitment. Group with depressive symptoms, GDS score between 6 and 15; group without depressive symptoms, GDS score between 0 and 5.

Table 1 shows the patient characteristics on admission. The median age was 76.5 (range: 40.0–83.0) years. The median GDS score on admission was 6.0 (range: 3.0–10.0). Table 2 shows patient data for each group. The motor FIM score on admission and the motor FIM gain were significantly lower in the group with depressive symptoms (40.0 and 28.0, respectively) than in the group without depressive symptoms (45.0, P = 0.026 and 34.0, P < 0.001, respectively). The CCI on admission was significantly higher in the group with depressive symptoms (2.0) than in the group without depressive symptoms (1.5, P < 0.001). Moreover, there were significant differences between the two groups in terms of BMI, MNA-SF, and MMSE scores on admission, as well as motor FIM score at discharge. In contrast, no significant difference was observed in age, sex, fracture type, antidepressant use, period of rehabilitation, length of hospital stay, or energy intake.

Table 1.  Patient characteristics on admission.

Characteristic
Number of patients	127 (100)
Age, years	76.5 (40.0–83.0)
Sex
Male	35 (27.6)
Female	92 (72.4)
Fracture type
Femoral fracture	62 (48.8)
Spine fracture	43 (33.9)
Others	22 (17.3)
BMI, kg/m2	22.7 (19.7–25.1)
MNA-SF	7 (3–9)
CCI	1 (0–2)
MMSE	20.0 (16.0–25.5)
Motor FIM	42.2±12.8
GDS	6.0 (3.0–10.0)
Depressive symptoms group a	71 (55.9)
No depressive symptoms group b	56 (44.1)

Values are presented as number (percentage), mean ± standard deviation, or median with interquartile range.

^a GDS score between 6 and 15; ^b GDS score between 0 and 5.

Table 2.  Comparison of variables between depressive symptoms group and no depressive symptoms group.

Factor	Depressive symptoms a (n=71)	No depressive symptoms b (n=56)	P value
Age, years	83.0 (79.0–87.0)	83.0 (75.0–89.3)	0.771
Sex
Male	21 (29.6)	14 (25.0)	0.709
Female	50 (70.4)	42 (75.0)
On admission
Fracture type
Femoral fracture	35 (49.3)	27 (48.2)	1.000
Spine fracture	25 (35.2)	18 (32.1)	0.862
Others	11 (15.5)	11 (19.6)	0.706
Use of antidepressants	4 (5.6)	2 (3.6)	0.694
No antidepressants	67 (94.4)	54 (96.4)
BMI, kg/m2	21.8 (18.5–24.3)	23.9 (20.7–26.3)	0.022
MNA-SF	8.0 (6.0–10.0)	9.0 (8.0–10.0)	0.045
CCI	2.0 (1.5–3.0)	1.5 (1.0–2.0)	<0.001
MMSE	18.0 (15.0–23.0)	24.5 (18.8–27.0)	<0.001
Motor FIM	40.0±12.4	45.0±12.8	0.026
At discharge
Period of rehabilitation, min/day	122.4 (114.8–139.6)	124.1 (114.0–136.8)	0.957
Length of hospital stay, days	56.9±20.4	58.7±19.2	0.605
Motor FIM	68.0 (59.0–84.0)	85.0 (75.3–89.0)	<0.001
Motor FIM gain	28.0 (21.0–37.5)	34.0 (29.0–39.5)	<0.001
Energy intake, kcal	1400 (1300–1600)	1400 (1300–1600)	0.947

Values are presented as median with interquartile range, number (percentage), or mean ± standard deviations.

^a GDS score between 6 and 15; ^b GDS score between 0 and 5.

Univariate and multivariate linear regression analyses were performed to assess the associations between motor FIM gain and each factor (Tables 3 and 4). Univariate linear regression analysis showed that BMI on admission, depressive symptoms on admission, motor FIM on admission, MMSE on admission, and antidepressant use were significantly associated with motor FIM gain (P < 0.05). Sex, fracture type (femoral fracture), period of rehabilitation, and length of hospital stay were not significantly associated with motor FIM gain (P ≥ 0.05) but were included in a multivariate linear regression analysis (P < 0.2). Multivariate linear regression analysis revealed that BMI on admission [β coefficient = −0.366, 95% confidence interval (CI) −0.725 to 0.010; P = 0.044], depressive symptoms on admission (β = −5.022, 95% CI −9.551 to −0.494; P = 0.030), motor FIM on admission (β = −0.432, 95% CI −0.652 to −0.212; P < 0.001), MMSE score on admission (β = 0.724; 95% CI 0.325 to 1.123; P < 0.001), and antidepressant use (β = −12.643, 95% CI −22.436 to −2.850; P = 0.012) were significantly associated with motor FIM gain. However, sex, fracture type (femoral fracture), period of rehabilitation, and length of hospital stay were not significantly associated with motor FIM gain. Variance inflation factor (VIF) values (1.025–1.867) indicated no multicollinearity.

Table 3.  Univariate linear regression analysis for motor FIM gain.

Univariate analysis	Coefficient	SE	95% CI	P value
Age	−0.137	0.118	−0.247 to 0.220	0.908
Sex a	3.686	2.827	−1.909 to 9.281	0.195
BMI on admission	−0.736	0.173	−1.078 to −0.393	<0.001
Depressive symptoms on admission b	−5.835	2.507	−10.797 to −0.872	0.022
Motor FIM on admission	−0.380	0.094	−0.566 to −0.194	<0.001
MMSE on admission	0.527	0.207	0.117 to 0.936	0.012
Femoral fracture c	−4.628	2.510	−9.596 to 0.340	0.068
Spine fracture d	2.725	2.676	−2.572 to 8.021	0.311
Use of antidepressants e	−15.488	5.832	−27.029 to −3.946	0.009
CCI on admission	−0.979	0.960	−2.879 to 0.924	0.310
Energy intake	0.000	0.004	−0.008 to 0.008	0.940
Period of rehabilitation	0.126	0.072	−0.017to 0.268	0.083
Length of hospital stay	0.110	0.064	−0.016 to 0.236	0.086
MNA-SF on admission	−0.112	0.655	1.409 to 1.184	0.864

SE, standard error.

^a Female coded 0, male coded 1; ^b Group with depressive symptoms (GDS score between 6 and 15) coded 1, group without depressive symptoms (GDS score between 0 and 5) coded 0; ^c Femoral fracture coded 1, no femoral fracture coded 0; ^d Spine fracture coded 1, no spine fracture coded 0; ^e Use of antidepressants coded 1, no use of antidepressants coded 0.

Table 4.  Multivariate linear regression analysis for motor FIM gain.

Multivariate analysis	β coefficient	SE	95% CI	P value	VIF
Sex	4.281	2.436	−0.543 to 9.105	0.081	1.081
BMI on admission	−0.366	0.181	−0.725 to 0.010	0.044	1.430
Depressive symptoms on admission a	−5.022	2.286	−9.551 to −0.494	0.030	1.193
Motor FIM on admission	−0.432	0.111	−0.652 to −0.212	<0.001	1.867
MMSE on admission	0.724	0.201	0.325 to 1.123	<0.001	1.356
Use of antidepressants b	−12.643	4.945	−22.436 to −2.850	0.012	1.025
Femoral fracture c	−2.041	2.132	−6.263 to 2.181	0.340	1.049
Period of rehabilitation	0.061	0.061	−0.059 to 0.182	0.316	1.040
Length of hospital stay	−0.004	0.061	−0.125 to 0.118	0.949	1.355

SE, standard error; VIF, variance inflation factor.

^a Group with depressive symptoms (GDS score between 6 and 15) coded 1, group without depressive symptoms (GDS score between 0 and 5) coded 0; ^b Use of antidepressants coded 1, no use of antidepressants coded 0; ^c Femoral fracture coded 1, no femoral fracture coded 0.

DISCUSSION

This study aimed to determine whether depressive symptoms on admission compromise motor FIM gain among patients with fractures admitted to a rehabilitation unit. We found that depressive symptoms on admission reduced the degree of improvement in motor FIM gain. In addition, patients taking antidepressants showed less improvement in their ADLs.

Lieberman et al.²²⁾ found no significant association between depressive symptoms and rehabilitation outcome. However, in this study, depressive symptoms were associated with lower improvements in ADLs. One possible explanation for this discrepancy is the difference in the length of hospital stay and rehabilitation intensity. In the study by Lieberman et al.,²²⁾ the mean length of stay was 22.1 days, whereas the mean length of stay was 56.9 days in this study. In addition, patients in the lowest rehabilitation intensity group in the present study received daily rehabilitation of 122.4 min, suggesting that rehabilitation was provided at an intensive level. Depression is a mood disorder characterized by apathy, loss of motivation, and decreased activity.^{36,37,38,39,40,41)} Depressive symptoms, such as apathy and loss of motivation, can hinder rehabilitation effectiveness.⁴²⁾ In intensive rehabilitation, apathy and loss of motivation may have had a cumulative negative impact on rehabilitation effectiveness. Similarly, decreased activity may have resulted in inadequate improvement in physical functioning and, subsequently, compromised the improvement in ADLs. However, biological mechanisms have been proposed to explain the impact of depression on physical disability.^43,44) Depression may have negative effects on the endocrine, nervous, and immune systems by increasing sympathetic tone, decreasing vagal tone, and causing immunosuppression.⁴⁴⁾ These depression-related biochemical pathways may even have a direct effect on neuromuscular functional decline,^45,46,47) potentially reducing rehabilitation effectiveness.

Patients taking antidepressants on admission showed less improvement in ADLs. Although several modalities exist for treating depressive disorders, pharmacotherapy remains the most common first-line treatment strategy.⁴⁸⁾ However, antidepressants are associated with adverse events, such as sedation, nocturia, and impaired sleep, which increase the risk of falls.^{49,50,51,52,53)} Selective serotonin reuptake inhibitors (SSRIs) and related drugs affect sleep duration and quality. SSRIs can also cause insomnia, which may result in nocturia and daytime sleepiness.⁵⁰⁾ In addition, trazodone and mirtazapine have been reported to have a high frequency of somnolence in clinical trials⁵⁴⁾ and cause daytime sleepiness.⁵⁰⁾ Daytime sleepiness is a predictor of lower rates of physical function recovery in the rehabilitation of older patients, and it is associated with lower improvement in ADLs.⁵⁵⁾ The antidepressants used in this study included paroxetine, duloxetine, trazodone, and mirtazapine. In this study, antidepressant users may have experienced lower rates of physical function recovery in rehabilitation because of daytime sleepiness and somnolence, which may have resulted in low improvement in ADLs. Furthermore, previous studies have reported that approximately 50% of patients with depression experience an inadequate response to antidepressant therapy, and medication adjustments, such as changing the type of antidepressant or increasing the dosage, are often required for these patients.^56,57,58) In this study, such medication adjustments may not have been sufficiently implemented. As a result, the therapeutic effects of antidepressants may have been inadequate, which could have contributed to the lower improvement in ADLs among antidepressant users.

In addition to depressive symptoms and antidepressant use, lower MMSE scores and higher BMI on admission were also associated with reduced improvement in ADLs. Cognitive impairment on admission has been reported to decrease rehabilitation outcomes in older patients with hip fractures.^59,60) Regarding BMI, previous studies have reported a negative correlation between BMI and FIM score improvement.^61,62) The results of this study are consistent with these previous reports, indicating that both cognitive function and BMI may influence rehabilitation outcomes.

The findings of this study highlight the importance of properly assessing depressive symptoms in patients with fractures undergoing rehabilitation and implementing appropriate treatment interventions when necessary. Despite the clinical importance of depression in the geriatric population, only 10%–20% of patients are diagnosed and adequately treated.⁶³⁾ In this study, 71 patients (55.9%) had depressive symptoms on admission; 4 patients (5.6%) in the group with depressive symptoms and 2 patients (3.6%) in the group without depressive symptoms used antidepressants. This suggests that depression diagnosis and treatment may have been inadequate in this study. Therefore, early assessment of depression at the start of rehabilitation and appropriate interventions may enhance improvement in ADLs. At that time, it is necessary to proactively adjust the dosage and type of medication while observing daytime sleepiness and activity level. In some cases, it may be necessary to consult a psychiatrist.

This study has some limitations. First, there is no information about the pre-injury functional level. In previous studies, pre-injury functional level was reported as a factor influencing short-term rehabilitation outcomes for proximal hip fractures.⁶⁴⁾ Future studies should evaluate the effect of pre-injury functional level on the improvement of ADLs in patients with fractures. Second, there might be some biases such as selection bias and information bias because we had no information on patients who refused GDS. It is possible that severely depressed or more confused patients refused the GDS. However, it was difficult to obtain such information. Furthermore, differences in the type of rehabilitation performed may influence the effectiveness of rehabilitation after a fracture. The patients underwent rehabilitation, including strength training and ADLs training, and the length of hospital stay and rehabilitation period were evaluated. However, it was difficult to evaluate whether all patients were subjected to a comparable extent of rehabilitation exercises. Therefore, further validation with emphasis on these indicators is needed in future studies.

CONCLUSION

Among patients with fractures admitted to a recovery rehabilitation unit, depressive symptoms on admission were associated with less improvement in ADLs. In addition, antidepressant use was an independent factor affecting improvement in ADLs. However, treatment for depressive symptoms in older patients may have been inadequate. Therefore, patients being admitted with fractures and showing depressive symptoms upon admission should receive adequate treatment of depressive symptoms in conjunction with rehabilitation because it can potentially improve ADLs in these patients.