Is health-related quality of life associated with the risk of low-energy wrist fracture: a case-control study

Low-energy fractures are common in the elderly population, and the wrist is one of the most common fracture sites [1‐5]. By definition, low-energy fracture results from minimal trauma, e.g., falling from a standing position [6]. It is important to characterize and identify patients at high risk of low-energy fracture based on the presence of risk factors to understand better the potential complex mechanisms involved and thereby to develop preventive strategies to reduce fracture risk.

Low-energy fractures are related mainly to the event of falling. However, the reasons for falling may be explained by several factors. Obviously, a fall may relate to the activities people undertake and the environmental conditions in which activities are performed, as well as an individual's physical and mental capacities to meet the challenges of the activities. In ageing subjects health status may be impaired [1, 4, 5, 7, 8], and falling may relate to impaired physical health (e.g., reduced balance, inability to perform ordinary daily activities, or other diseases) or impaired mental health (e.g., cognition, inattentiveness, depression or anxiety). Moreover, because bone mineral density (BMD) declines with age [1], bones may have less strength to tolerate stress caused by falls. The importance of BMD at the time of fracture in explaining wrist fractures among elderly people is well documented [1, 2, 9, 10], whereas physical and mental health only have been used as outcome parameters to assess recovery after fracture [7, 8].

Health-related quality of life (HRQOL) may be seen as a patient's evaluation of his or her health status. Thus, HRQOL is defined as the individual's experience of his or her general state of health, such as physical, social and mental functioning, and well-being [11]. Self-reported health is supposed to capture the full array of a person's illness and possibly even symptoms of as yet undiagnosed diseases in preclinical stages [12, 13]. Traditionally, HRQOL is measured using a questionnaire comprising a broad variety of health aspects such as bodily pain, general health, physical and social functioning, role functioning, mental health and vitality [14]. To understand the context of low-energy fractures in a broader sense, HRQOL evaluation may give a more complex perspective on the issue of possible fracture prevention. Strong risk factors for low-energy wrist fracture have been difficult to identify and are not widely studied, as wrist fractures often occur in relatively healthy elderly persons [1]. To our knowledge, no studies have examined self-reported health status as a potential risk factor before fracture, together with objective risk factors (e.g., BMD) in patients with a low-energy wrist fracture. Several studies reporting pre-injury HRQOL among patients with other types of low-energy fractures (e.g., hip fractures) do not include assessments of BMD and therefore lack an important health parameter [15‐17].

This study examined pre-fracture HRQOL in low-energy wrist fracture patients by including self-reported physical and mental health and BMD. We first examined whether HRQOL shortly before fracture in elderly patients with low-energy wrist fracture differed from that of age- and sex-matched controls recruited randomly from the general population. We also explored whether HRQOL was independently associated with wrist fracture after adjusting for known risk factors for fracture such as age, weight, osteoporosis and falls.

The wrist fracture patients reported HRQOL before fracture at a level similar to the controls. Several dimensions of HRQOL were independently associated with increased risk of wrist fracture. However, the direction of the associations with wrist fracture varied between the different sub-dimensions of HRQOL. Higher scores on the SF-36 domains of general health, bodily pain and mental health, and lower scores on physical role limitation and social functioning were associated with an increased risk of wrist fracture. Our data indicate that patients with favourable and unfavourable HRQOL measures may be at increased risk of wrist fracture.

Our finding that favourable health in some SF-36 domains is associated with increased risk of wrist fracture contradicts the current opinion that only individuals with impaired health are at increased risk of low-energy fracture. One explanation may be that elderly people in good health lead a more active lifestyle, which may increase the risk of falling [1, 4, 5]. However, our HRQOL results seem to show a duality because we also found that individuals with impaired physical function were at increased risk of fracture. Kelsey et al. [4] concluded that wrist fractures often occur as a result of a fall in women with low BMD who are relatively healthy and active and have good neuromuscular function.

Our study identified BMI, osteoporosis and falls as independent risk factors for wrist fracture and agrees with the results of other studies [1, 2, 4, 10, 24]. Furthermore, we found that living alone was independently associated with increased risk of wrist fracture. As in our study, a lower rate of cohabitation in fracture patients has been reported previously [25]. This suggests that behavioural and psychological factors associated with living alone, which are also reflected in the HRQOL, might influence the risk of falls and fractures [26‐28].

Our data suggest that a complex mix of circumstances and factors can increase or decrease the risk of low-energy fracture in the elderly. The clinical implications of our findings are equivocal because the association between the SF-36 domains and wrist fractures were inconsistent and use of SF-36 is not feasible in routine daily care. We emphasize that our results should not be used to recommend restriction of physical activity in elderly persons because the overall beneficial health effects of physical activity and an active lifestyle in elderly individuals are indisputable [29].

The significance of some differences in SF-36 domains between the wrist fracture patients and controls was not striking. Thus, the observed differences may be attributed to random or selection biases for both fracture patients and controls. To validate our control group, we compared the controls with normative Norwegian national data for the SF-36 domains [30]. Across all the SF-36 domains, controls had higher age- and sex-adjusted scores compared with the Norwegian normative data (p < 0.05). One could argue that the national population-based SF-36 norms might have provided better control values for our wrist fracture patients. However, this was a case-control study with patients and controls gathered from the same geographical area, and we used data from the matched controls for comparisons. Regardless of whether we compare with national norms or with the case controls, our results run counter to the hypothesis that low HRQOL is associated with low-energy wrist fractures.

Our study has some limitations that should be considered when interpreting the findings. The collection of HRQOL data retrospectively, i.e., after fracture, is a possible limitation because ideally such data would be collected before the fracture occurred. However, it is almost impossible to collect HRQOL data prospectively for a population with a specific injury, and alternative methods rely on pre-injury recall, as used in this trauma study and other studies [15, 17, 31, 32]. Retrospective evaluations can be biased by recall problems and response shifts due to fracture [21, 33, 34]. To minimize this problem, it is recommended that HRQOL assessments should be performed with the shortest possible time lag after the fracture event, which we aimed for in our study [21, 23]. The elapsed time from fracture to assessment was relatively short and most patients completed the pre-fracture HRQOL questionnaire within the first two weeks after fracture. Thus, it seems unlikely that the patients would have forgotten about their HRQOL immediately before and at the time of the fracture [35]. Patients who have experienced a recent change in health are more likely to make accurate responses than are controls [21]. To minimize further the limitations of our study design, the questionnaire was administered with a clear instruction that the patient should think of the period before the fracture. For our study, a prospective study design would also have limitations because HRQOL may change between the time of data capture and fracture and thus may no longer be valid for the time of fracture.

The controls were asked to report their HRQOL as it was prior to their visit to the osteoporosis centre. We could argue that controls also should have had a time lag of reporting their HRQOL, e.g., relate the questions to the time preceiding two weeks ago. On the other hand, the elapsed time between fracture and examination at the osteoporosis centre in the patients varied, and it was difficult to achieve a perfect match to reach the same time lag [34]

Furthermore, both patients and controls were asked to report objective data based on the preceding period or year before fracture or inclusion (controls), and both groups might be influenced by recall bias [21, 34]. On the other hand, objective data are less likely to be influenced by events such as a fracture, and objective data in both patients and controls therefore would be quite comparable.

Both patients and controls self-reported their co-morbidities and other diseases and conditions. The participants' hospital records might have been a more reliable source, but hospital records also have their shortcomings. Furthermore, self-report was considered the most ethically correct approach. BMD measurements were carried out after the fracture, which is unlikely to influence the BMD measurement. The sites of BMD measurements were hip and spine, and BMD would not change significantly over a period of two weeks.

Only 56% of the wrist fracture patients referred to the hospital were included in this study. Inclusion of only those patients willing or capable of self-reporting their health status may have underestimated the importance of HRQOL as a risk factor for wrist fracture and the difference in HRQOL between wrist fracture patients and controls. Furthermore, the controls included in our study also comprised elderly persons willing to participate in the study. A selection bias towards the healthiest elderly might arguably occur. However, our data give us limited information regarding the controls who were unwilling to participate, and the assumption that it was the healthiest among the eligible controls who chose to participate in the study cannot be verified. Despite these limitations, a major strength of our study is that the patients included were recruited from the only referral centre for orthopaedic trauma in the region. Thus, we were able to account for all patients who were referred and treated at the hospital.

The inconsistent results observed in our study might be caused by the design, the retrospective recall used to assess HRQOL and covariates, and differences in sampling fractions for cases and controls [21]. It is possible that our findings should be regarded as restricted to elderly persons within the geographical area of our cases and controls. On the other hand, it must be regarded as a strength that patients and controls have been recruited from the same geographical area. As previously discussed, both retrospective and prospective designs to assess HRQOL and covariates have their shortcomings, and achieving identical sampling fractions for both cases and controls is difficult. Therefore, the design and methods used in this study seem to be appropriate to answer our research questions.

Our data support the contention that wrist fracture patients perceive themselves to be quite healthy compared with controls from the general population. The association between wrist fracture and HRQOL was inconsistent, and favourable scores for some SF-36 domains were associated with an increased risk of wrist fracture, whereas other SF-36 domains exhibited the opposite pattern. Although several dimensions of HRQOL were related to an elevated risk for low-energy wrist fracture, our data suggest that the utility of HRQOL as an indicator of risk of fracture is limited in clinical practice because of the inconsistent direction of the associations for different sub-dimensions. Our study underlines the role of biological mechanisms, including low BMI, osteoporosis and previous falls, in increasing the risk of wrist fracture. In addition, behavioural and psychological risk factors associated with living alone should be considered. Further studies of HRQOL in wrist fracture patients are warranted, especially studies that include larger sample sizes, older patients and patients with greater impairment of physical and mental health.