Incidence and risk factors associated with hip fracture in institutionalised elderly people in Japan

SIR—Institutionalized elderly people are considered to form a high-risk group for hip fracture. Sugarman et al. [1] showed that the incidence of a hip fracture in nursing home residents was four times higher than that of non-nursing home residents. Epidemiological studies of hip fractures among institutionalised elderly have been conducted by several researchers [5–8], and have demonstrated that low bone mass, impaired mobility and presence of a previous fracture are important risk factors. However, epidemiological studies in institutionalised elderly people are rare compared to those in general populations, and no studies to date have been conducted among Asian populations. The aim of this study was to clarify the incidence and risk factors of a hip fracture in geriatric homes for long-term care services in Japan. The target population of this study comprised residents living in 201 homes registered to Niigata Council of Institutions for the Elderly, providing long-term care services in Niigata Prefecture, Japan. Two types of nursing homes for long-term care services exist in Japan: nursing homes for frail elderly people who require extensive care (special nursing homes) and nursing homes for elderly people who require some care (other nursing homes). Of the 201 nursing homes (137 special nursing homes) targeted, 140 (95 special nursing homes) participated in this study. The 8,905 elderly people in the 140 homes as of 1 April 2007 constituted the cohort for this study. Of these, 7,654 people who had no incidence of a hip fracture were analysed. The ethics committee of Niigata University School of Medicine approved the protocol for this study. The present cohort study began on 1 April 2007 and finished on 31 March 2008. We asked nurses and caregivers of each home to provide subjects’ information regarding demographic characteristics, levels of ADL and dementia at the start of the study. Age, sex, height, weight and past medical histories were obtained from medical records and body mass index (BMI, kg/m2) was calculated. ADL levels were evaluated by caregivers using the Barthel index [6]. Dementia levels were also assessed by caregivers using the dementia scale provided by the Ministry of Health, Labor and Welfare of Japan [13, 14], which is an observer-rating scale widely used in the long-term care insurance system to evaluate individual care service needs. This scale includes five categories: (1) normal, (2) slight, (3) light, (4) moderate and (5) severe. Details of the Barthel index and dementia scale are shown in Appendix 1 (available as supplementary data at Age and Ageing online). We asked each home to report incident of hip fractures, including fractures at the femoral neck, trochanter and femoral head during the 1-year follow-up period. Hip fracture diagnosis was based on medical records of the hospitals where patients were admitted. Information on dates of discharge from the geriatric homes was also obtained to calculate person-years of this population. The incidence and unadjusted relative risk of a hip fracture were calculated by the number of hip fractures divided by person-years. We used Cox's proportional-hazards model to estimate the adjusted relative risk of hip fracture. Independent predictor variables in a multivariate model were selected using a stepwise method. Statistical Analysis Software (SAS; release 9.13, SAS Institute Inc., Cary, NC, USA) was used for data analysis. A P-value of < 0.05 was considered to be statistically significant. Of the 7,654 subjects, 5,556 were special nursing home residents. Average age and BMI were 85.0 years and 20.7 kg/m2 for women and 80.1 years and 20.5 kg/m2 for men, respectively. Frequency distribution of total scores for the Barthel index and other characteristics at baseline are shown in Appendices 2 and 3 (available as supplementary data at Age and Ageing online), respectively. The incidence (per 1000 person-years) of hip fracture during the 1-year follow-up was 14.9 (11.0 for femoral neck, 3.8 for trochanter and 0.2 for femoral head) in women and 7.9 (5.7 for femoral neck, 1.1 for trochanter and 1.1 for femoral head) in men. The age-specific incidence rate of hip fracture was 8.3, 16.7 and 10.8 for women and 4.4, 7.7 and 11.7 for men in their 70s, 80s and 90s, respectively. The overall incidence of hip fracture in women and men was 12.6/1000 person-years. Totally dependent, bedridden persons are less likely to partake of activities that place them at a fall risk and thus rarely suffer hip fracture. Hip fracture incidence in the group with a Barthel index score of 0, i.e. totally dependent persons (1.0/1000 person-year), was much lower than other groups (17.1/1000 person-year). Because this group was considered heterogeneous, unadjusted relative risks for hip fracture were calculated in subjects with a Barthel index score ranging between 1 and 100, i.e. excluding subjects with a score of 0 (Table 1). Subjects with the following characteristics had a significantly higher risk of hip fracture: female, 80–89 years old, special nursing home resident, lower weight, lower BMI, third-quartile Barthel index score (55–85) and having any level of dementia. Unadjusted relative risk (RR) of predictors for hip fracture in subjects with a Barthel index score ranging between 1 and 100 (1-year follow-up) Q1, first quartile; Q2, second quartile; Q3, third quartile; Q4, fourth quartile. aNursing homes for the frail elderly. bCalcium supplementation use by the subjects was recorded. Fifty-nine homes fortified diets with calcium, and subjects in these homes were classified as calcium supplement users. Unadjusted relative risk (RR) of predictors for hip fracture in subjects with a Barthel index score ranging between 1 and 100 (1-year follow-up) Q1, first quartile; Q2, second quartile; Q3, third quartile; Q4, fourth quartile. aNursing homes for the frail elderly. bCalcium supplementation use by the subjects was recorded. Fifty-nine homes fortified diets with calcium, and subjects in these homes were classified as calcium supplement users. Cox's proportional-hazards model was used to identify independent risk factors for hip fracture in subjects with a Barthel index score ranging between 1 and 100. No linear association was observed between the incidence of a hip fracture and variables of ‘age group’, ‘Barthel index score’ or ‘levels of dementia’ (Table 1); therefore, these variables were considered dummy variables. Significant high-risk categories were coded as 1, and others were coded as 0. The stepwise method selected suboptimal Barthel index scores (55–85), having light to severe dementia and low weight as independent predictors of hip fracture (Table 2). Independent predictors of hip fracture selected by the stepwise method in subjects with a Barthel index score ranging between 1 and 100 (i.e. excluding totally dependent subjects with score 0) aEstimated by the Cox's proportional-hazards model. bCorresponding to Q3 of total score of the Barthel index in Table 1. Independent predictors of hip fracture selected by the stepwise method in subjects with a Barthel index score ranging between 1 and 100 (i.e. excluding totally dependent subjects with score 0) aEstimated by the Cox's proportional-hazards model. bCorresponding to Q3 of total score of the Barthel index in Table 1. The incidence of hip fracture in the present population can be compared to that of other frail populations. Walter et al. [9] showed that the incidence (per 1000 person-years) of a hip fracture was 33 in their overall cohort (average age 79 years) and 45 among disabled white women aged 80–90 years of age. Chandler et al. [3] showed a hip fracture incidence rate of 44 (per 1000 person-years) among white women living in nursing homes (average age 85). In comparison with these data, the incidence of hip fracture among Japanese women in the present study (14.9/1000 person-years) was about one-third of the incidence among white women. It should be noted, however, that such overall comparisons are limited by the inability to control for certain key demographic variables and risk factors. The present study and some previous epidemiological studies have suggested that the incidence of a hip fracture among East Asians is lower than among Caucasians [10]. This may be associated with genetic factors, such as skeletal structure and body size. Also, differences in hip geometry between Asians and Caucasians may explain the ethnic differences in the incidence of hip fractures, i.e. the shorter hip axis or femoral neck length of Asians may contribute to their lower risk [11, 12]. Performance of ADLs may be a key factor for hip fracture among the frail elderly. Previous studies have shown that independence of ambulation is a significant predictor of osteoporotic fracture in nursing home residents. Walter et al. [9] showed that those elderly individuals who are ambulating independently have a 3-fold higher risk of a hip fracture. This result is slightly different from our findings, which suggest that partially independent subjects (Barthel index score between 55 and 85; the second highest group in terms of ADL) may have the highest risk of hip fracture. Partially independent subjects may have a higher risk of falling, which can lead to higher susceptibility to hip fractures than dependent and totally independent people. The results from the current study also revealed that those with dementia have a 1.9-fold higher risk of a hip fracture than those without dementia, independent of ADL levels. It has been reported that elderly people with cognitive impairment had a 1.6- to 2.1-fold increased risk [9, 13], and those with Alzheimer's disease had a 2-fold risk [14] of a hip fracture. The increased risk of a hip fracture associated with dementia may be partially explained by the increased risk of falling [15]. There were limitations in this study. First, we do not know the incidence of hip fracture in nursing homes (∼30%) that did not participate in this study. Second, hip fracture diagnoses were based on medical certificates from hospitals. No independent case review board confirmed the diagnoses, potentially leading to misclassification of cases. Third, the dementia scale used in this study is not used internationally. Therefore, further studies are needed to confirm the association between dementia and hip fracture. This is the first epidemiological study to examine the incidence and risk factors of hip fracture among institutionalised frail elderly people in East Asia. We showed that suboptimal ADLs, the presence of dementia and low weight are independent risk factors for hip fracture, and this information will be useful for caregivers to prevent hip fracture in long-term care facilities. Future studies should focus on determining protective factors so that interventions can be designed in order to reduce hip fracture incidence. The incidence of hip fracture among institutionalised Japanese women was about one-third of the incidence among institutionalised white women. Suboptimal ADLs, the presence of dementia and low weight are independent risk factors for hip fracture in the institutionalised elderly people. We wish to thank Niigata Council of Institutions for the Elderly for help in data collection. We thank Ms A. Kagoshima for help with data analysis. The authors have no conflicts of interest. This study was partly supported by a grant from Chiyoda Mutual Life Foundation, 2006. Supplementary data are available at Age and Ageing online.