Introduction
Osteoporosis is one of the most common diseases in the elderly population. The process of bone tissue loss usually does not give any clear clinical symptoms, and the first manifestation is a low-trauma fracture [1]. Hip and spine fractures are the most harmful to overall health. Osteoporotic fractures are considered a late consequence of osteoporosis; therefore, the main goal of patient treatment is effective prevention. The first osteoporotic fracture is one of the most important risk factors for subsequent fractures [2]. A previous forearm fracture increases this risk 2-fold, and a spine or hip fracture increases the risk of subsequent fractures by up to 4–6 times. Therefore, one of the most important steps is the early identification of fracture risk factors. In general, the risk of fractures is influenced by the condition of the skeletal system and the tendency to fall due to functional status. Low bone mass detected by hip or spine densitometry increases the risk of fractures. Several so-called clinical risk factors have been identified. These include certain medications, comorbidities, smoking, high alcohol consumption, a diet low in calcium, early menopause, and the influence of genetic factors, especially hip fractures in parents. Additionally, all circumstances affecting the functional condition and the risk of falling significantly increase the risk of fractures. All the above-mentioned risk factors for fractures should be taken into account when managing the patient. Notwithstanding this general recommendation, the role of risk factors is not always the same for fractures at different skeletal sites. For example, a spine or rib fracture can occur without a fall, and peripheral fractures of the forearm, arm, or ankle are almost always associated with a fall. Knowing the relative strength of previously outlined risk factors can improve preventive measures.
The aim of the study was to present data on risk factors for fractures at various sites of the skeleton in a 10-year prospective longitudinal observation. It can be hypothesised that fracture risk factors should be different for spine, hip, and peripheral fractures.
Material and methods
The study group was formed by the RAC-OST-POL study cohort recruited in 2010. The population was randomly selected from postmenopausal women over 55 years of age from the entire Racibórz district in southern Poland. At baseline, the total number of randomly selected women was 625. Baseline epidemiological data have been reported previously [3]. However, in addition to randomly selected women invited via letters sent by traditional mail, an additional 353 women who volunteered were included in the study. To present data in a larger cohort, we added these women to the previously described randomly recruited group. We checked whether the baseline incidence of falls was dependent on the method of recruitment (random or non-random). As shown by the chi-square test, the incidence of falls did not differ between them (33.8% in the random cohort and 33.1% in the non-random cohort, p = 0.84).
Therefore, further study analyses were performed in the entire cohort of 978 women [4–6]. The survey described in detail earlier [3] collected several data, including previous falls that occurred in the 12 months preceding the date of the review. Data on falls and osteoporotic fractures were collected at baseline in 2010 and annually thereafter for 10 years. One experienced researcher (WP) called all patients annually to collect the mentioned data. The following potential risk factors for fractures were collected at baseline: smoking, alcohol consumption, parental femoral neck fracture, medications (steroids or anticonvulsants), and comorbidities (type 1 and type 2 diabetes, depression, bronchial asthma, ulcerative colitis, kidney disease, liver cirrhosis, rheumatoid arthritis, thyroid disease, and Alzheimer’s disease).
Statistical analysis
All calculations were performed using Microsoft Office Excel and Statistica software (StatSoft Inc., Tulsa, OK; www.statsoft.com) running on a PC. Descriptive statistics of quantitative variables were presented as means and standard deviations (SD). The normality of the distribution of the analysed data was checked using the Shapiro-Wilk test. To compare continuous parameters between subgroups, the t-test for independent samples was performed. The presentation of qualitative features was made by providing the number of subjects and the percentage in specific subgroups. The frequency of qualitative features between subgroups was compared using the chi-square test. Correlation analysis was performed using the Spearman correlation test. Finally, logistic regression identified the features most strongly associated with fractures. All p values < 0.05 were considered statistically significant.
Results
The mean baseline age, weight, height, and body mass index (BMI) were 65.0 ± 6.9 years, 74.5 ± 14.0 kg, 156.6 ± 6.6 cm, and 30.6 ± 5.4 kg/m2, respectively. During the observation period, 190 low-traumatic fractures were recorded in 129 patients. The following number of fractures was observed: hip — 15 in 14 subjects, spine — 30 in 17 subjects, and non-hip non-spine fractures — 145 in 99 subjects (including 81 forearm fractures in 67 subjects).
Table 1 shows the clinical characteristics in the 4 subgroups according to the fracture site.
|
Parameter |
Spine fracture (n = 16) |
Hip fracture (n = 14) |
Wrist fracture (n = 67) |
Non-spine and non-hip fracture (n = 99) |
|
Age [years] |
67.5 ± 7.1 |
71.6 ± 9.1* |
66.7 ± 7.3 |
66.9 ± 6.8** |
|
Height [cm] |
155.3 ± 5.4 |
158.0 ± 4.4 |
157.3 ± 5.1 |
156.9 ± 5.4 |
|
Weight [kg] |
69.2 ± 10.6 |
70.7 ± 8.6 |
74.5 ± 12.6 |
75.0 ± 13.5 |
|
BMI [kg/m²] |
28.7 ± 4.2 |
28.4 ± 4.0 |
30.2 ± 5.1 |
30.5 ± 5.5 |
The nature of fractures in different parts of the skeleton is not uniform, and the role of falls may vary. To determine the significance of falls, their impact was examined separately for spine, hip, wrist, and all non-spine, non-hip fractures. The effect of falls was insignificant in the case of spine fractures (chi-square test: 3.64; df = 1, p = 0.06). For all other sites, the incidence of fractures increased significantly when falls occurred, with the greatest effect observed for forearm fractures and all non-spine and non-hip (chi-square test for hip fractures, forearm fractures, and all non-spine, non-hip fractures was 6.43, p < 0.05; 42.7, p < 0.0001 and 66.7, p < 0.0001, respectively).
Similar results were obtained when the number of falls recorded during the observation period was additionally taken into account — the results of Spearman’s correlation analysis between the number of falls and the frequency of fractures in various parts of the skeleton, as well as comparisons of the strength of the correlation, are presented in Table 2.
|
Parameter |
Spine fracture (n = 16) |
Hip fracture (n = 14) |
Wrist fracture (n = 67) |
Non-spine and non-hip fracture (n = 99) |
|
r and p value |
0.06, ns |
0.11, p < 0.01 |
0.24, p < 0.0001*,** |
0.28, p < 0.0001#,## |
To determine the factors having a significant impact on the incidence of fractures during the observation period, logistic regression was used separately in subgroups. The following risk factors were considered, selected based on their known clinical significance and/or potential impact on fractures identified in univariate analyses: age, height, body weight, bone mineral density (BMD) of the femoral neck expressed by T-score, rheumatoid arthritis, steroid use, falls reported at baseline, and total number of risk factors. Spine fractures were dependent only on T-score, OR = 0.42 (0.23–0.76), hip fracture only on age OR = 1.15 (1.07–1.24), forearm only on T-score OR = 0.69 (0.51–0.92), and non-hip, non-spine on falls rate OR = 1.86 (1.20–2.87).
Discussion
The current study presents several observations regarding the role of risk factors for fractures at various skeletal sites. The most important finding is the observation that the incidence of fractures increased with the falls rate in long-term follow-up. Additionally, we have also proven that a greater number of falls also increases the risk of fractures. Another valuable finding for daily practice is the observation that each fracture site was associated with only one dominant risk factor. Spine and forearm fractures were dependent on the T-score, hip fractures depended only on age, and non-hip and non-spine fractures depended on the frequency of falls. Another unexpected finding was the observation that a greater cumulative number of clinical risk factors did not increase the incidence of falls.
We believe that the results obtained are important for practitioners. Some medical interventions that may increase BMD or reduce the incidence of falls may reduce the incidence of fractures.
A population-representative cohort and long-term longitudinal observation allowed us to present reliable data. Valuable conclusions emerge from the comparison of current results with those reported in the same population at baseline [4]. Baseline data showed that fractures occurred less frequently in women from the city of Racibórz than in women from rural areas. Baseline analysis also found that type 1 diabetes and depression influence fall rates. Previous analyses also showed a significant role of kidney diseases, rheumatoid arthritis and bronchial asthma, which was not demonstrated in longitudinal observation. Both analyses showed strong associations between fractures and falls.
We believe that the comparison of baseline data (4) and current longitudinal data suggests that the more reliable come from long-term follow-up.
Some studies published so far have also been based on longitudinal observations [7–11]. Only one study reported the results of a 10-year follow-up [11], while others reported shorter follow-ups ranging from 4 months to 3 years. In a study by Balogun et al. [11] the observed fall rate was 64%, which is very close to our result of 60%. The incidence of fractures was lower than in our study (17 vs. 20%), probably due to the lower mean age.
Some studies present data on preventive models (12-14). An interesting concept of fall prevention was established in 647 nursing home patients during a 9-month follow-up period [14]. The authors recommended limiting the physical activity, and the use of this restriction was associated with a higher incidence of fall-related injuries. Vitamin D levels should also be considered, because vitamin D can improve muscle function and reduce the incidence of falls.
In a long-term, 20-year observation of recreational gymnastics, Uusi-Rasi et al. [15] presented some procedures that have the potential to reduce the incidence of falls in older people. However, none of the cited studies included a separate analysis for the risk for fractures in specific skeletal sites. In this aspect, the results of our study are extremely innovative.
In some other studies of cross-sectional design (16-19) the results regarding the role of falls have been reported. We believe that despite some interesting information resulting from these studies, the most valuable data are those based on long-term follow-up. Our previous results, presented earlier, support the statement of priority given by long-term longitudinal studies.
The affected functional condition is especially dangerous for older people. American recommendations for patient management present procedures aimed at reducing the incidence of falls [20]. Polish recommendations also propose directions to improve muscle condition and reduce the risk of falls [21].
Our study has some limitations: we only studied women, and no data were collected on the exact date and place of falls (at home or outside). The role of vitamin D was not established in the current study, and we plan to conduct such an analysis in future studies.
However, the study design, the collection of a representative epidemiological sample, and long-term follow-up allowed us to obtain reliable data on the role of factors increasing the incidence of fractures.
Concluding, fractures at various skeletal sites recorded in long-term follow-up in postmenopausal women were related to different risk factors. There was only one dominant risk factor that increases the incidence of fractures at specific skeletal site. The impact of falls was most pronounced in the case of non-hip, non-spine fractures.
Data availability statement
The data may be available upon request
Ethics statement
The study was approved by the Ethics committee of Medical University of Silesia (KNW/0022/KB1/132/10). At baseline in 2010 year all participants gave their written informed consent.
Author contributions
W.P.: the concept and design of the study, the acquisition of data, an analysis and interpretation of data, drafting of the manuscript, the final approval of the submitted version (the first author). P.A.: an analysis and interpretation of data, drafting of the manuscript, preparing graphical presentation, the final approval of the submitted version. B.D.: an analysis and interpretation of data, a critical revision of the manuscript, the final approval of the submitted version. H.H.: an analysis and interpretation of data, a critical revision of the manuscript, the final approval of the submitted version.
Funding
No funding information.
Conflict of interest
The authors decare they have no conflict of interest