Exploring factors affecting activities of daily living in patients with osteoporotic vertebral fractures managed conservatively: a post-hoc analysis of a prospective cohort study
Article information
Abstract
Study Design
A post-hoc analysis of a prospective cohort study.
Purpose
This study aimed to identify factors at the time of injury associated with declining activities of daily living (ADLs) in the chronic phase of osteoporotic vertebral fractures (OVFs) managed conservatively.
Overview of Literature
Although a conservative approach is the treatment of choice for OVFs, ADLs do not improve or eventually decrease in some cases. However, the risk factors for ADL decline after the occurrence of OVFs, particularly the difference between those with or without initial bed rest, are unknown.
Methods
A total of 224 consecutive patients with OVFs aged ≥65 years who received treatment within 2 weeks after the occurrence of injury were enrolled. The patients were followed up for 6 months thereafter. The criteria for evaluating the degree of independence were applied to evaluate ADLs. Multivariable analysis with a logistic regression model was performed to evaluate the risk factors for ADL decline.
Results
In total, 49/224 patients (21.9%) showed a decline in ADLs. Of these, 23/116 patients (19.8%) in the rest group and 26/108 patients (24.1%) in the no-rest group experienced a decline in ADLs. In the logistic regression analyses, a diffuse low signal on T2-weighted magnetic resonance imaging (MRI) (odds ratio, 5.78; 95% confidence interval, 2.09–16.0; p=0.0007) and vertebral instability (odds ratio, 3.89; 95% confidence interval, 1.32–11.4; p=0.0135) were identified as independent factors in the rest and no-rest groups, respectively.
Conclusions
In patients with acute OVFs, a diffuse low signal on T2-weighted MRI and severe vertebral instability were independently associated with ADL decline in patients treated with and without initial bed rest, respectively.
Introduction
Osteoporotic vertebral fractures (OVFs) are the most common fragility fractures [1]. Although conservative treatment is the first choice for OVFs, no treatment has been standardized because of insufficient clinical evidence [2]. In clinical practice, early ambulation without bed rest is generally performed for OVFs; however, the effects of a proper non-weight-bearing period during the initial stages of conservative treatment of OVFs were reported [3]. A previous study showed that strict hospital bed rest for 2 weeks during the initial stages of conservative treatment yielded favorable results, with 79% of patients achieving bone union [4]. Another study showed that 2-week hospital bed rest exerts a therapeutic effect by reducing the need for surgery among patients with poor prognostic magnetic resonance imaging (MRI) findings. This study also showed that 2-week hospital bed rest exerts a therapeutic effect by reducing the progression of vertebral collapse and kyphosis, regardless of the MRI findings [5].
One of the goals of conservative treatment for older patients with OVFs is to maintain independence in performing activities of daily living (ADLs). However, in clinical practice, clinicians often encounter patients whose ADLs do not improve or worsen during conservative treatment.
Several studies have evaluated ADL decline in patients with OVFs. Suzuki et al. [6] reported that ADLs improved in the first 3 months but subsequently plateaued. Yasuda et al. [7] reported that ADLs declined similarly in patients with bone union and delayed union after conservative treatment of OVFs. Wakao et al. [8] did not find an association between pseudoarthrosis after OVFs and ADL outcomes, whereas Takahashi et al. [9] reported that delayed union is associated with ADL decline. These studies have focused on the association between fracture healing and ADL outcomes. Conversely, only one study analyzed patients’ demographic factors at injury and ADL decline after conservative treatment. Matsumoto et al. [10] reported that posterior wall injury and not performing daily exercise before the occurrence of injury were risk factors for ADL decline after the occurrence of OVFs. To date, the effects of the differences in conservative management methods, including the presence or absence of bed rest, on ADL decline after conservative treatment of OVFs have not been investigated.
We hypothesized that if we could identify risk factors at the time of injury that affect ADLs in the chronic phase of OVFs in patients managed conservatively, therapeutic intervention for these factors after fractures could help prevent ADL decline. In addition, identifying these factors would help determine the appropriate treatment strategy for patients experiencing ADL decline. Therefore, this study aimed to identify factors at the time of injury associated with ADL decline in the chronic phase of OVFs managed conservatively with and without bed rest.
Materials and Methods
This post-hoc analysis of a previous prospective cohort study compared the therapeutic effects of 2-week bed rest with no bed rest on the outcome of OVFs treated conservatively [5]. This study was approved by the review board of authors’ affiliated institutions (H30-172) and was conducted in accordance with the tenets of the Declaration of Helsinki. The hospital obtained from the patients a comprehensive agreement for academic use of information such as type of treatments, treatment progress, or any other data acquired during treatments at the time of hospitalization, and no identifiable information of the participants is included in the manuscript.
Summary of the original study [5] (Fig. 1)
Flow diagram for the study sample selection. OVF, osteoporotic vertebral fracture; ADL, activity of daily living.
A total of 224 of 329 consecutive patients with OVFs aged ≥65 years who received treatment within 2 weeks after the occurrence of injury between December 2018 and December 2020 were enrolled in the analysis. The patients were followed up for 6 months after the occurrence of injury. All patients who could not be managed at home because of severe pain were routinely recruited. Patients with pathological fractures caused by metastatic spinal tumors and pyogenic spondylitis, fractures within diffuse idiopathic skeletal hyperostosis, comorbidities requiring priority treatment, poorly controlled diabetes (hemoglobin A1c ≥7.5%), secondary osteoporosis caused by long-term oral corticosteroid use and dialysis, and degenerative lumbar scoliosis (Cobb angle >20°) were excluded. Patients without pain and those who did not agree to be included in the study were also excluded.
The study participants were patients from two general hospitals in different geographic areas. Of the 153 patients in the rest group, 116 participated in the study. Of the 176 patients in the no-rest group, 108 participated in the study. The rest group enforced strict hospital bed rest during the first 2 weeks of conservative treatment of OVFs, as previously described [4], and the patients were only allowed to leave the bed for walking rehabilitation thereafter. The no-rest group was allowed to leave their beds as soon as possible for walking rehabilitation according to their pain tolerance.
A Jewett-type rigid orthosis was used in all patients (both groups) for 3 months after leaving the bed. The conservative approach was transitioned to surgical treatment when continued rehabilitation did not improve the ADLs because of pain. Osteoporosis medications were prescribed by the attending physician.
Data collection
Patient age, sex, injured spinal level, body mass index, comorbidities, spinal level within the thoracolumbar junction (T11–L2), and prior OVFs were recorded. The bone mineral density (BMD) of the proximal femur, tartrate-resistant acid phosphatase (TRACP)-5b, total procollagen type-I N-terminal propeptide (P1NP), and 25-hydroxy vitamin D (25OH-vitD) levels were measured through blood tests.
Radiographic assessment
Poor prognostic image findings of fracture healing and conservative treatment-resistant image findings were assessed based on previous studies [11–16]. Lateral dynamic radiographs at the injury level were taken with the patient standing or sitting (loaded position) and lying supine (nonloaded position). The vertebral collapse rate (%) was defined by the following equation: [1−(anterior vertebral wall height/posterior vertebral wall height)]×100. Vertebral instability (%) was defined as the difference in vertebral collapse rate between the loaded and nonloaded images [11].
The presence or absence of localized high signals and diffuse low signals on T2-weighted sagittal MRI sections [12] and the presence or absence of injuries to the middle column, comprising the posterior vertebral wall and the pedicles, in horizontal computed tomography (CT) images were assessed [11].
Outcome measures
To evaluate the ADLs in each enrolled patient, the criteria for evaluating the degree of independence (severity of the bedridden state) proposed by the Long-term Care Insurance System in the Japanese Health and Welfare Ministry were applied (Table 1). The criteria were previously used in several studies on OVF treatment [5,7,10,17,18]. These criteria have four different ranks: J levels indicate an ability to move about independently (4 points), A levels indicate housebound status but an ability to move with assistance (3 points), B levels indicate chairbound status that usually indicates wheelchair use (2 points), and C levels indicate bedbound status (1 point).
Activities of daily living categories
The ADL rank was recorded both before the occurrence of injury and after 6 months. If the treatment transitioned to surgical treatment because ADL did not improve, the preoperative ADL rank was also recorded. The decrease in ADL score (in points), number of patients with lower ADL ranks, and number of patients with a decrease in ADL rank of ≥1 rank were recorded. According to the clinical experiences and standpoints, an ADL decline was defined as a reduction of ≥1 rank by the 6-month follow-up after fracture and as a need to transition to surgical treatment because ADLs did not improve with rehabilitation.
Data analysis
Unpaired t-tests and chi-square tests were applied to compare groups in patient demographics. Subsequently, the associations between patients’ background data at the time of injury and ADL decline after conservative treatment of OVFs were investigated in the univariable and multivariable analyses.
First, univariable analyses were performed using Fisher’s exact probability test. Continuous variables were divided by the median and converted to categorical variables. Second, multivariable analysis was performed using multiple logistic regression models for the whole cohort and each group separately. The explanatory variables were selected according to the results of the univariable analysis and clinical background knowledge. Although no consensus on the appropriate sample size for multiple regression analysis has been established, 10 events per variable were considered clinically reasonable in some studies [19,20]. This study followed this concept. Subsequently, the odds ratios (OR) and 95% confidence intervals (CI) for ADL decline were calculated. In addition, the variance inflation factor (VIF) was calculated as an index of multicollinearity, and items with VIF values of ≤5 were considered independent variables.
All statistical analyses were performed using EZR software ver. 1.3.0 (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [21]. All p-values of <0.05 were considered statistically significant.
Results
Patient demographics and events
Table 2 shows the baseline characteristics of the patients. The mean patient age was 80.4 years in the rest group and 81.5 years in the no-rest group. Both groups showed no significant differences in all categories at baseline.
Patient demographics
In total, 23/116 patients (19.8%) in the rest group experienced a decline in ADLs. Of these patients, 20 showed a reduction of ≥1 rank, and three transitioned to surgical treatment. Among these patients, two underwent balloon kyphoplasty and one underwent vertebroplasty and percutaneous pedicle screw fixation. Conversely, 26/108 patients (24.1%) in the no-rest group experienced a decline in ADLs. Of these patients, 17 showed a reduction of ≥1 rank, and nine transitioned to surgical treatment. Among these patients, five underwent balloon kyphoplasty, three underwent vertebroplasty and percutaneous pedicle screw fixation, and one underwent posterior and anterior fusion. Finally, 49/224 patients (21.9%) in both groups experienced ADL decline, and these cases were used as objective variables in this study (Fig. 1).
Factors contributing to ADL decline after the occurrence of OVFs
Univariable analyses revealed that vertebral instability (p=0.001), diffuse low signal on T2-weighted MRI (p<0.001), posterior vertebral wall injury on CT (p=0.003), and pedicle injury on CT (p=0.003) were significantly associated with ADL decline (Table 3). These four variables were consistent with the clinical background knowledge.
Univariable analysis among all patients
For the multivariable logistic regression analysis for the whole cohort, all four variables were selected as explanatory variables according to the level of ADL decline (n=49). In addition, for multivariable logistic regression analysis for the rest and no-rest groups, two of four variables with smaller p-values, namely, vertebral instability and diffuse low signal on T2-weighted MRI, were selected as explanatory variables according to the number of cases with ADL decline (23 and 26 patients, respectively).
Among all patients, a diffuse low signal on T2-weighted MRI (OR, 3.08; 95% CI, 1.47–6.46; p=0.0028; VIF, 1.094) was identified as an independent factor associated with an ADL decline (Table 4). In addition, a diffuse low signal on T2-weighted MRI (OR, 5.78; 95% CI, 2.09–16.0; p=0.0007; VIF, 1.006) and vertebral instability (OR, 3.89; 95% CI, 1.32–11.4; p=0.0135; VIF, 1.126) were identified as independent factors in the rest and no-rest groups, respectively (Table 5).
Multivariable logistic regression analysis: independent factors contributing to ADL decline among all patients
Multivariable logistic regression analysis: independent factors contributing to ADL decline among each group
Discussion
In this study, the risk factors for ADL decline after conservative treatment of OVFs were analyzed. In the univariable analyses, vertebral instability, diffuse low signal on T2-weighted MRI, posterior vertebral wall injury on CT, and pedicle injury on CT were significantly associated with ADL decline. In the multivariable logistic regression analyses, a diffuse low signal on T2-weighted MRI was identified as an independent factor associated with an ADL decline among all patients and the rest group, whereas vertebral instability was an independent factor in the no-rest group. To the best of our knowledge, this study is the first to show the heterogeneity of factors that affect ADL decline depending on the presence or absence of bed rest from the conservative management strategy.
The frequency of ADL decline was reported to range from 20.4% [7] to 21.3% [10]. In this study, the frequency of ADL decline was 21.9% in the whole cohort, which was similar to those in previous reports. Matsumoto et al. [10] reported that posterior wall injury and lack of daily exercise before the occurrence of injury are risk factors for ADL decline after the occurrence of OVFs. This study did not examine differences in conservative treatment methods, including the presence or absence of bed rest and the type of brace. In the present study, the differences in outcomes with the presence or absence of bed rest after the type and duration of brace were unified were investigated.
Although univariable analyses revealed that posterior vertebral wall injury was one of the significant factors associated with ADL decline, only a diffuse low signal on T2-weighted MRI was an independent risk factor for ADL decline in the entire cohort and the rest group in the multivariable logistic regression analyses. Conversely, severe vertebral instability was an independent risk factor in the no-rest group, which means that the factors influencing ADL decline differed depending on the presence or absence of initial bed rest.
A diffuse low signal on T2-weighted MRI is a known predictor for nonunion and resistance to conservative treatment of OVFs [12,13,16]. This finding illustrates that the damage to the cancellous bone of the fractured vertebral body is widespread, and the degree of damage inside the body would be correspondingly large [12]. On the contrary, large vertebral body instability has also been reported as one of the imaging findings that predicts nonunion [9] and resistance to conservative treatment [11]. Without bed rest, a large vertebral instability is expected to cause significant collapse. In contrast, vertebral instability was not an independent risk factor in the rest group. We believe that this was due to the therapeutic effect of 2-week bed rest on reducing the progression of vertebral collapse and kyphosis from the original study [5].
Based on the present results, to avoid ADL decline, clinicians must pay maximum attention to a diffuse low signal on T2-weighted MRI when bed rest is chosen and to large vertebral instability when early ambulation without bed rest is chosen.
This study has some limitations. First, this is a post-hoc analysis of a subset of a limited number of patients from an original cohort study [5]. Not all initially enrolled patients were evaluated. Some patients were lost to follow-up because several patients entered geriatric homes after discharge or were unable to revisit the hospital. Second, patient-reported outcome measures were not assessed; as ADLs and quality of life are closely related, several studies have reported using a questionnaire of patient-reported outcome measures for the conservative treatment of OVFs [22,23]. Although patient-reported outcome measures are important, they are not possible to obtain at the pre-injury period. Third, pain assessment was not performed in this study. According to a previous study, >30% of patients with OVFs have decreased cognitive function or dementia [24]. Although pain is an important factor in the selection of treatment options in OVFs, there is a great concern on the reliability of pain assessments for people with dementia [25–27].
Despite these limitations, the selection biases were low because patients’ demographic data at the time of injury were not different between the two groups (Table 2). The results of this study have significant clinical implications. Moreover, the event (ADL decline) sample size was deemed sufficient based on the 10 events per variable concept for multiple regression analysis [19,20]. Four variables among the whole cohort and two variables among each group were selected for the multivariable logistic regression analysis. Finally, the VIFs of the extracted risk factors after multivariable logistic regression analyses were all <2, which indicates that all of them were dependent variables. Therefore, we believe that our findings apply to clinical practice.
Recently, several studies have described an association between malnutrition and ADL decline [28,29] or potentially inappropriate medications and ADL decline in patients with OVFs [30]. Future studies including these factors should evaluate the effects of bed rest or absence thereof on ADL decline. Finally, a multicenter randomized controlled trial and longitudinal study are warranted to establish optimized, standardized conservative treatment protocols for patients with OVFs.
Conclusions
When conservative treatments were selected for acute OVFs, a diffuse low signal on T2-weighted MRI was independently associated with ADL decline in patients treated regardless of the initial bed rest. Moreover, a diffuse low signal on T2-weighted MRI and large vertebral instability were independently associated with ADL decline in patients treated with initial bed rest and early ambulation without initial bed rest, respectively.
Key Points
Among 224 consecutive patients with osteoporotic vertebral fractures, 49/224 (21.9%) showed a decline in activities of daily living (ADLs).
Of these, 23/116 (19.8%) patients in the rest group and 26/108 (24.1%) in the no-rest group experienced a decline in ADLs.
A diffuse low signal on T2-weighted magnetic resonance imaging and severe vertebral instability were independently associated with ADL decline in patients treated with and without initial bed rest, respectively.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Author Contributions
TF, MT, and KF designed the study. YS, SO, KS, TS, KI, HG, KM, and HN co-treated the patients and analyzed the data. TF wrote the manuscript. TF, HT, and MK revised the manuscript. MY supervised the manuscript. All authors approved the final manuscript.