Abstract
Background
The aim of this study was to evaluate the association between the degree of religiosity and subsequent fractures and a decrease in bone mineral density in a Japanese population.
Methods
We conducted a retrospective longitudinal study at St. Luke’s International Hospital in Tokyo, Japan, from 2005 to 2018. All participants who underwent voluntary health check-ups were included. Our outcomes were any fractures and the change in T-score from baseline to each visit. We compared these outcomes by the self-reported degree of religiosity (not at all; slightly; somewhat; very) and adjusted for potential confounders.
Results
A total of 65,898 participants were included in our study. Their mean age was 46.2(SD:12.2) years, and 33,014(50.1%) were male. During a median follow-up of 2,500 days (interquartile range (IQR):987–3,970), 2,753(4.2%) experienced fractures, and their mean delta T-score was -0.03%(SD:18.3). In multivariable longitudinal analyses, the slightly religious group had a statistically lower adjusted odds ratio (AOR) for a fracture than the nonreligious group(AOR:0.81,95% confidence interval(CI):0.71 to 0.92).
Conclusions
We demonstrated that slightly religious people, but not somewhat or very religious people, had a lower incidence of fracture than nonreligious individuals, although the T-scores were similar regardless of the degree of religiosity.
Supplementary Information
The online version contains supplementary material available at 10.1186/s13030-023-00265-6.
Keywords: Fracture, Japan, Large scale, Longitudinal study, Religion, T score
Key messages
・Slightly religious people had a lower incidence of any fractures.
・Somewhat or very religious people didn’t have a lower incidence of any fractures
・The T-scores were similar regardless of the degree of religiosity.
Introduction
Regardless of the type of religion, religious people have been shown to have better mental and physical health outcomes [1]. As a benefit of religiosity, religious people have also been shown to have longer life expectancies than less religious people [2–5]. However, in terms of each physical disease and its risk factors, the evidence is still mixed. For instance, systematic reviews have indicated that the incidence of cardiovascular disease and its mortality were lower among religious populations [5], but large cohort studies and reviews found the opposite [6, 7]. Among the cardiovascular risk factors, our previous study found a lower risk of future diabetes but did not reveal changes in hypertension or dyslipidemia among religious people [8]. Unfavorable evidence for religiosity is dominant in some unhealthier conditions, such as obesity/overweight [7, 9–11]. Similarly, in terms of mental health, such as depression, the evidence is mixed [1]. Further evidence is required to evaluate these associations.
Although osteoporosis/fracture is considered to be one of the important health conditions of elderly individuals, a very limited number of studies about the association between religion and osteoporosis/fracture have been published. One cross-sectional study showed different prevalences of osteoporosis by the type of religion among women [12]. However, detailed studies on this topic are scarce. For osteoporosis/fractures, a fracture would be considered a better outcome than osteoporosis. The degree of religiosity or religion among different populations should be evaluated.
The aim of this study was to evaluate the association between the degree of religiosity and subsequent fractures and decreases in bone mineral density in a Japanese population.
Methods
We conducted a retrospective longitudinal study at St. Luke’s International Hospital, a large teaching hospital in Tokyo, Japan, from 2005 to 2018. We included all participants who underwent voluntary health check-ups at the center of preventive medicine in the hospital from 2005 to 2012. The purpose of health check-ups is to screen for and manage chronic diseases, including malignancy, hypertension, diabetes and osteoporosis. We excluded participants who had a prior history of any fractures at baseline visits based on their medical records at the hospital and self-reports. Our primary outcome was any fracture during follow-ups up to December 31, 2018, and the secondary outcome was the change in T-score from baseline to each visit. We compared these outcomes by the self-reported degree of religiosity, adjusting for potential confounders.
St. Luke’s Ethics Committee Institutional Review Board approved this study (approval number: 18-R203. Comprehensive approvals for studies about social habits).
Fractures and Delta T-scores
Our primary outcome was any fracture during the follow-up period. Information about a fracture was obtained based on electronic medical records at the hospital and self-reports at the time of each follow-up visit. The fracture information at the hospital contained the site of the fracture and its international classification of disease 10 (ICD-10) codes [13], whereas that from a self-report had only the history of a fracture. Those who had fractures multiple times during the study period only had their first episode of a fracture recorded as an outcome.
The T-scores were calculated based on the bone mineral density (BMD) in comparison to young adult mean (YAM) values based on the World Health Organization (WHO) criteria [14]. Participants had the BMD of their radial bone measured as a part of their health check-ups at each follow-up visit by dual-energy X-ray absorptiometry (DEXA) DCS-6EX-3 by ALOKA (Tokyo Japan). The T-scores are expressed not as the standard deviation (SD) but as the percentage (%).T-scores of 100% and 70% in Japan are equivalent to 0 SD and -2.5 SD, respectively, according to the WHO criteria [15]. The delta T-score was defined as the change in the T-score from baseline at each visit.
Degree of religiosity
From 2005 to 2012, all participants were asked about their religiosity by responding to the following questionnaire at each follow-up visit: “Are you religious? (translated from Japanese)”. Participants chose one of the following responses: not at all, slightly, somewhat, and yes. We categorized them into four groups: not religious at all, slightly religious, somewhat religious and very religious according to their response to this question. These responses were obtained at every follow-up visit and involved a time-dependent variable. We considered “none” as the reference group for the analyses.
Confounders
We obtained the following information to consider potential confounders in the analyses: participant demographics, health habits, comorbidities, and treatment for osteoporosis. The demographics included age, sex and body mass index (BMI). BMI was calculated by trained staff who measured the participant’s height and weight at each visit. The BMI was categorized into three groups based on WHO criteria for Asians [16]: underweight (BMI < 18.5 kg/m2), normal (18.5–24.9 kg/m2) or obese/overweight (≥ 25.0 kg/m2). Participants were asked about their health habits at each visit as well: alcohol consumption (abstainer, occasional drinker or regular drinker), smoking status (never, former or current) and exercise habit (almost none, 1–2 times a week, 3–5 times a week, or almost all days). Information about comorbidities, including a current diagnosis of hypertension, diabetes and dyslipidemia, was based on self-reports. The history of depression and its treatment status were also taken. In addition, information about pharmacological treatment status and each medication taken for osteoporosis was obtained based on medical records or self-reports. These potential confounders were collected at every follow-up visit and considered to be time-dependent covariates.
Statistical methods
We obtained descriptive statistics on baseline characteristics and outcomes by baseline degree of religiosity. Next, multivariable longitudinal analyses were performed, adjusting for potential confounders. The generalized estimating equation (GEE) with a binomial distribution and logit link function was used for the incidence of any fracture, and a mixed effects model was used for the delta T-score. We applied different confounders for adjustment for different multivariable models: model 1 was adjusted for time variables, age and sex, and baseline T-score; model 2 was adjusted for health habits (alcohol consumption, cigarette smoking and exercise) and body mass index in addition to the covariates in model 1; model 3 was adjusted for comorbidities (current diagnosis of hypertension, diabetes and dyslipidemia) in addition to the covariates in model 2; model 4 was adjusted for treatment status for osteoporosis. As subanalyses, a similar approach was applied by stratifying by age (< 50 years, ≥ 50 years) and sex. In addition, we performed similar analyses focusing on hip fracture and vertebral fracture as outcomes.
All analyses were performed by using STATA MP 14.2 in 2019 (STATA Corp., College Station, TX, USA).
Results
A total of 65,898 participants were included in our study. Their mean age was 46.2 (SD: 12.2) years, 33,014 (50.1%) were male, and their mean baseline T-score was 96.9 (SD: 13.4) %. Table 1 shows the baseline characteristics by the degree of religiosity. In terms of demographic data, religious people tended to be older, female, and obese/overweight. They had positive health habits, such as less regular drinking, less current cigarette smoking, and higher levels of exercise. More current comorbidities were diagnosed among religious people. The prevalence of depression and its treatment status were similar across all religious groups. Regarding fracture-related factors, religious people were more likely to be diagnosed with osteoporosis and be treated for it, but all groups had similar T-scores.
Table 1.
Baseline characteristics and outcome by baseline degree of religiosity
| Religiosity | ||||||||
|---|---|---|---|---|---|---|---|---|
| Not religious at all | Slightly religious | Somewhat religious | Very religious | |||||
| n = 16,050 | n = 25,378 | n = 17,992 | n = 6,478 | |||||
| Outcomes | ||||||||
| Fracture, n (%) | 518 | (3.2) | 933 | (3.7) | 867 | (4.8) | 435 | (6.7) |
| Delta T score, % (SD) | 0.1 | (17.0) | -0.1 | (16.9) | -0.1 | (16.8) | -0.2 | (16.9) |
| Follow-up periods, days (IQR) | 2,346 | (843- 3,816) | 2,464 | (945—3,963) | 2,572.5 | (1,077—4,032) | 2,553 | (1,076—4,012) |
| Demographics | ||||||||
| Age, years (SD) | 41.7 | (11.2) | 45.3 | (11.5) | 49.3 | (12.1) | 52.1 | (13.0) |
| Male, n (%) | 8,810 | (54.9) | 13,321 | (52.5) | 8,080 | (44.9) | 2,803 | (43.3) |
| Body mass index | ||||||||
| Underweight | 1,657 | (10.3) | 2,385 | (9.4) | 1,691 | (9.4) | 576 | (8.9) |
| Normal | 11,423 | (71.2) | 18,164 | (71.6) | 12,670 | (70.4) | 4,522 | (69.8) |
| Obesity/overweight | 2,968 | (18.5) | 4,829 | (19.0) | 3,631 | (20.2) | 1,380 | (21.3) |
| Health habits, n (%) | ||||||||
| Alcohol consumption | ||||||||
| Abstainer | 6,064 | (37.8) | 9,265 | (36.5) | 7,263 | (40.4) | 3,085 | (47.6) |
| Occasional | 2,704 | (16.9) | 4,662 | (18.4) | 3,276 | (18.2) | 1,098 | (17.0) |
| Regular | 7,282 | (45.4) | 11,451 | (45.1) | 7,453 | (41.4) | 2,295 | (35.4) |
| Cigarette smoking | ||||||||
| Never smoker | 9,151 | (57.0) | 15,074 | (59.4) | 11,317 | (62.9) | 4,248 | (65.6) |
| Former smoker | 3,174 | (19.8) | 5,774 | (22.8) | 4,220 | (23.5) | 1,444 | (22.3) |
| Current smoker | 3,725 | (23.2) | 4,530 | (17.9) | 2,455 | (13.6) | 786 | (12.1) |
| Exercise | ||||||||
| Almost none | 7,176 | (44.7) | 9,649 | (38.0) | 5,953 | (33.1) | 2,057 | (31.8) |
| Once to twice a week | 5,621 | (35.0) | 9,846 | (38.8) | 6,804 | (37.8) | 2,241 | (34.6) |
| 3–5 times a week | 1,966 | (12.3) | 3,704 | (14.6) | 3,236 | (18.0) | 1,196 | (18.5) |
| Almost everyday | 1,287 | (8.0) | 2,179 | (8.6) | 1,999 | (11.1) | 984 | (15.2) |
| Comorbidities, n (%) | ||||||||
| Hypertension | 835 | (5.2) | 1,774 | (7.0) | 1,640 | (9.1) | 745 | (11.5) |
| Diabetes | 280 | (1.7) | 571 | (2.3) | 510 | (2.8) | 270 | (4.2) |
| Dyslipidemia | 434 | (2.7) | 992 | (3.9) | 1,008 | (5.6) | 450 | (7.0) |
| Depression | 447 | (2.9) | 674 | (2.7) | 540 | (2.9) | 212 | (3.0) |
| Medications for depression | 167 | (1.1) | 270 | (1.1) | 207 | (1.1) | 86 | (1.2) |
| Fracture related factors | ||||||||
| Baseline T score, % (SD) | 96.7 | (13.6) | 96.8 | (13.4) | 97.0 | (13.2) | 96.9 | (13.2) |
| Osteoporosis, n (%) | 17 | (0.1) | 56 | (0.2) | 77 | (0.4) | 47 | (0.7) |
| Treatment for osteoporosis, n (%) | 8 | (0.1) | 29 | (0.1) | 33 | (0.2) | 23 | (0.4) |
| Selective estrogen receptor modulator | 2 | (0.0) | 3 | (0.0) | 1 | (0.0) | 1 | (0.0) |
| Vitamin D | 7 | (0.0) | 25 | (0.1) | 31 | (0.2) | 23 | (0.4) |
| Bisphosphonate | 7 | (0.0) | 25 | (0.1) | 31 | (0.2) | 23 | (0.4) |
| Vitamin K | 0 | (0.0) | 0 | (0.0) | 1 | (0.0) | 0 | (0.0) |
| Calcium | 3 | (0.0) | 4 | (0.0) | 5 | (0.0) | 1 | (0.0) |
During a median follow-up of 2,500 days (interquartile range (IQR): 987–3,970), 2,753 (4.2%) experienced fractures, and the mean delta T-score was -0.03 (SD: 18.3). Among them, 36,465 (55.3%) had the same, 16,032 (24.3%) had increased, 12,556 (19.1%) had decreased, and 845 (1.3%) had a fluctuating degree of religiosity through study period. In our multivariable longitudinal analyses, the slightly religious group had significantly lower odds ratios (OR) for a fracture than did the nonreligious group in all models (OR: 0.81, 95% confidence interval (CI): 0.71 to 0.92 in model 4) (Table 2). The somewhat religious group had lower but not statistically significant ORs (OR: 0.88, 95% CI: 0.77 to 1.01 in model 4), and the very religious group had a similar incidence of fracture (OR: 1.07, 95% CI: 0.91 to 1.25 in model 4) compared to the nonreligious group. In terms of delta T score, the religious group tended to have a higher β coefficient, but the difference was not statistically significant (β coefficient: 0.10, 95% CI: -0.08 to 0.27 in model 4). There was no difference in the incidence of fracture between those who change their degree of religiosity over time and those who had same degree of religiosity over time, apart from the baseline degree of religiosity. (OR: 1.00, 95%CI: 0.92 to 1.10 for participants with increased; OR 1.00, 95%CI 0.90 to 1.10 for those with decreased; OR 1.20, 95%CI: 0.87 to 1.64 for those with fluctuated compared to those with same degree of religiosity). Table S1 shows the changes of health habits over time by the change of degree of religiosity over time. Those who had the same degree of religiosity over time (“no change” group in the change of degree of religiosity over time) tended to have same health habits. In contrast, there was no obvious trend in the change of health habits among those with increased, decreased, or both degree of religiosity overtime.
Table 2.
Adjusted odds ratio for development of any fracture and adjusted β coefficient for change of T score from baseline by religiosity from longitudinal analyses
| Adjusted odds ratio (95% confidence interval) | ||||
|---|---|---|---|---|
| Religiosity | ||||
| Not religious at all | Slightly religious | Somewhat religious | Very religious | |
| Any fracture | ||||
| Model 1 | reference | 0.81 (0.72 to 0.90) | 0.91 (0.81 to 1.02) | 1.07 (0.93 to 1.23) |
| Model 2 | reference | 0.80 (0.70 to 0.91) | 0.88 (0.77 to 1.00) | 1.06 (0.90 to 1.24) |
| Model 3 | reference | 0.80 (0.70 to 0.91) | 0.88 (0.77 to 1.00) | 1.06 (0.91 to 1.25) |
| Model 4 | reference | 0.81 (0.71 to 0.92) | 0.88 (0.77 to 1.01) | 1.07 (0.91 to 1.25) |
| Delta T score | Adjustedβ coefficient (95% confidence interval) | |||
| Model 1 | reference |
-0.02 (-0.14 to 0.11) |
0.01 (-0.13 to 0.14) |
0.09 (-0.09 to 0.26) |
| Model 2 | reference |
-0.01 (-0.13 to 0.11) |
0.02 (-0.12 to 0.15) |
0.10 (-0.08 to 0.28) |
| Model 3 | reference |
-0.01 (-0.13 to 0.11) |
0.02 (-0.12 to 0.15) |
0.10 (-0.08 to 0.27) |
| Model 4 | reference |
-0.01 (-0.13 to 0.11) |
0.02 (-0.12 to 0.15) |
0.10 (-0.08 to 0.27) |
Model 1 was adjusted for time variable, participants’ age and gender, and baseline T score; model 2 was adjusted for health habits (alcohol consumption, cigarette smoking and exercise) and body mass index in addition to covariates in model 1; model 3 was adjusted for comorbidities (current history of hypertension, diabetes and dyslipidemia) in addition to covariates in model 2; model 4 was adjusted for depression and its treatment status, and treatment status for osteoporosis in addition to covariates in model 3
Number in bold represents that the p value was less than 0.05
In the subanalyses by age and sex, all groups had significantly lower ORs for a fracture compared to the nonreligious reference group among older women (Table 3). The association between the degree of religiosity and a fracture among older women had a backward J shape (the very religious group had higher ORs than the slightly religious or somewhat religious group). Other subgroups, including younger women, both younger and older men, and the very religious group had higher ORs for a fracture, but most differences were not statistically significant. The results of subanalysis focusing on hip fracture and vertebral fracture as outcomes are shown in Table S2.
Table 3.
Adjusted odds ratio for development of any fracture and adjusted β coefficient for change of T scores from baseline by religiosity stratified by age and sex from longitudinal analyses
| Adjusted odds ratio (95% confidence interval) | ||||||||
|---|---|---|---|---|---|---|---|---|
| Male | Female | |||||||
| Not religious at all | Slightly religious | Somewhat religious | Very religious | Not religious at all | Slightly religious | Somewhat religious | Very religious | |
| Age < 50 years | ||||||||
| Any fracture | ||||||||
| Model 1 | Reference | 0.92 (0.74 to 1.14) | 1.13 (0.89 to 1.44) | 1.39 (0.99 to 1.95) | Reference | 0.94 (0.74 to 1.19) | 1.25 (0.98 to 1.59) | 1.39 (1.02 to 1.89) |
| Model 2 | Reference | 0.89 (0.70 to 1.13) | 0.99 (0.75 to 1.30) | 1.32 (0.91 to 1.90) | Reference | 0.92 (0.70 to 1.22) | 1.13 (0.85 to 1.49) | 1.39 (0.99 to 1.96) |
| Model 3 | Reference | 0.89 (0.70 to 1.13) | 0.98 (0.75 to 1.30) | 1.32 (0.91 to 1.91) | Reference | 0.92 (0.70 to 1.21) | 1.13 (0.85 to 1.49) | 1.39 (0.99 to 1.96) |
| Model 4 | Reference | 0.89 (0.70 to 1.13) | 0.98 (0.75 to 1.30) | 1.32 (0.91 to 1.91) | Reference | 0.92 (0.70 to 1.22) | 1.13 (0.86 to 1.49) | 1.41 (0.99 to 1.98) |
| Delta T score | Adjusted β coefficient (95% confidence interval) | |||||||
| Model 1 | Reference | -0.01 (-0.21 to 0.20) | 0.07 (-0.18 to 0.32) | 0.17 (-0.20 to 0.55) | Reference | -0.03 (-0.23 to 0.18) | 0.10 (-0.12 to 0.33) | 0.15 (-0.17 to 0.47) |
| Model 2 | Reference | -0.01 (-0.20 to 0.21) | 0.10 (-0.15 to 0.35) | 0.21 (-0.17 to 0.59) | Reference | -0.02 (-0.23 to 0.18) | 0.11 (-0.12 to 0.34) | 0.17 (-0.16 to 0.49) |
| Model 3 | Reference | 0.00 (-0.20 to 0.21) | 0.10 (-0.16 to 0.34) | 0.21 (-0.17 to 0.58) | Reference | -0.02 (-0.23 to 0.18) | 0.11 (-0.12 to 0.34) | 0.17 (-0.15 to 0.49) |
| Model 4 | Reference | 0.00 (-0.20 to 0.21) | 0.09 (-0.16 to 0.34) | 0.21 (-0.17 to 0.58) | Reference | -0.02 (-0.23 to 0.18) | 0.11 (-0.12 to 0.34) | 0.17 (-0.15 to 0.49) |
| Age ≥ 50 years | ||||||||
| Any fracture | Adjusted odds ratio (95% confidence interval) | |||||||
| Model 1 | Reference | 0.89 (0.68 to 1.17) | 1.00 (0.76 to 1.30) | 1.13 (0.83 to 1.54) | Reference | 0.71 (0.57 to 0.88) | 0.68 (0.56 to 0.84) | 0.77 (0.61 to 0.97) |
| Model 2 | Reference | 0.90 (0.66 to 1.23) | 1.00 (0.73 to 1.35) | 1.10 (0.78 to 1.57) | Reference | 0.63 (0.49 to 0.81) | 0.63 (0.50 to 0.80) | 0.74 (0.57 to 0.97) |
| Model 3 | Reference | 0.91 (0.67 to 1.23) | 1.00 (0.74 to 1.36) | 1.11 (0.78 to 1.58) | Reference | 0.63 (0.49 to 0.81) | 0.63 (0.50 to 0.80) | 0.75 (0.57 to 0.98) |
| Model 4 | Reference | 0.90 (0.66 to 1.23) | 1.00 (0.73 to 1.35) | 1.11 (0.78 to 1.58) | Reference | 0.63 (0.49 to 0.82) | 0.63 (0.50 to 0.80) | 0.76 (0.58 to 0.99) |
| Delta T score | Adjusted β coefficient (95% confidence interval) | |||||||
| Model 1 | Reference | 0.03 (-0.26 to 0.33) | -0.09 (-0.40 to 0.22) | 0.14 (-0.24 to 0.52) | Reference | -0.20 (-0.55 to 0.15) | -0.23 (-0.58 to 0.11) | -0.24 (-0.63 to 0.15) |
| Model 2 | Reference | 0.05 (-0.25 to 0.34) | -0.06 (-0.37 to 0.25) | 0.17 (-0.21 to 0.55) | Reference | -0.22 (-0.57 to 0.14) | -0.26 (-0.60 to 0.09) | -0.26 (-0.66 to 0.13) |
| Model 3 | Reference | 0.05 (-0.25 to 0.34) | -0.06 (-0.37 to 0.25) | 0.17 (-0.21 to 0.55) | Reference | -0.22 (-0.57 to 0.14) | -0.26 (-0.60 to 0.09) | -0.26 (-0.66 to 0.13) |
| Model 4 | Reference | 0.04 (-0.25 to 0.34) | -0.06 (-0.37 to 0.24) | 0.17 (-0.21 to 0.55) | Reference | -0.22 (-0.57 to 0.14) | -0.26 (-0.60 to 0.09) | -0.26 (-0.66 to 0.13) |
Model 1 was adjusted for time variable, age and sex, and baseline T score; model 2 was adjusted for health habits (alcohol consumption, cigarette smoking and exercise) and body mass index in addition to covariates in model 1; model 3 was adjusted for comorbidities (current history of hypertension, diabetes and dyslipidemia) in addition to covariates in model 2; model 4 was adjusted for depression and its treatment status, and treatment status for osteoporosis in addition to covariates in model 3
Number in bold represents that the p value was less than 0.05
Discussion
We demonstrated that being slightly religious is associated with a lower risk of fracture but not with changes in T-scores when compared to nonreligious individuals in all populations studied. Interestingly, somewhat or very religious people had a similar risk of fracture and delta BMD as nonreligious people in all populations studied. In addition, higher degrees of religiosity were associated with a lower risk of fracture compared to no religiosity among older women, although their T-scores were similar regardless of the degree of religiosity.
To the best of our knowledge, this study is the first to evaluate the association between the degree of religiosity and fractures and changes in BMD among a healthy population. Some previous studies may be slightly related, but most were different. Streeten et al. showed a reduced incidence of hip fracture in Old Order Amish compared to non-Amish whites in the U.S. [17] Although Old Order Amish is a religious population, the results may come from differences in lifestyle, such as very low levels of smoking or alcohol consumption, rather than the degree of religiosity. Another study from India showed different prevalences of osteopenia/osteoporosis across religions [18]. Although one study showed that regular Muslims worshipers had no difference in osteoporosis compared to nonregular worshipers, this study examined the effect of “praying” as a physical exercise rather than its religious aspects [19].
One of the potential reasons the incidence of fracture is decreased but BMD does not change in a specific religious group may be social behavioral reasons. Religious people may be more careful about accidents compared to nonreligious people in their occupation, habits, or life environment. In fact, religious people may avoid risky behaviors based on one of the common Japanese religious views—if you do a bad thing, it will come back to you- [20]. In contrast, religious people may be more active than slightly religious people [21, 22], resulting in an increased accident rate, leading to more fractures, although the increased level of exercise may have benefits for other health conditions, such as diabetes. In these ways, Japanese religious views may have both positive and negative impacts on fracture related social activities. Although we could not assume the magnitudes of each impact on the fracture, the difference of degree of religiosity may result in the different direction to fracture from the sum of each different impact. Our finding that there was no difference in delta BMD regardless of religiosity may support this hypothesis.
We also found that there was no difference in the incidence of any fracture between those who change their degree of religiosity over time and those who had same degree of religiosity over time, apart from baseline degree of religiosity. This may be simply due random error of the participants’ response over time. In fact, a similar proportion of “not religious at all” participants sometimes marked “slightly religious” over time and “very religious” participants sometimes marked “somewhat religious”. A similar phenomenon was observed for “slightly religious” and “somewhat religious” participants. Identically, the change of health habits over time had no obvious trend by the change of degree of religiosity over time.
Interestingly, the decrease in BMD was not different across religious groups. Religious people may consume more calcium, vitamin D or vitamin K. Generally, the source of animal protein in Japan mainly depended on fish hundreds of years ago for religious reasons, such as Shinto and Buddhism [23, 24]. Although current survey data about the association between religiosity and fish consumption in Japan is lacking, this tradition may have somewhat carried over, even in present days. High levels of fish consumption has been said to be associated with higher levels of vitamin D and a greater bone mass [25]. However, the effects of increased mineral intake through the diet may be insufficient for increasing BMD. In fact, calcium or vitamin D supplementation is recommended for the osteoporotic population, but there is no evidence that it increases BMD [26]. In addition, compared to cardiovascular risk factors, such as diabetes, prevention of osteoporosis may be less effective, resulting in similar BMDs across religiosity, even though religious people have a higher health awareness.
Differences in sex in the effect of religiosity on health outcomes have been reported in previous studies in other countries than Japan [27–29]. Interestingly, these studies evaluated gender differences in religiosity on health/well-being [27, 28] and cognitive function [29] and reported that the impact of religiosity on men was stronger than that on women in the U.S. and China, contrasting with our findings that religious elderly women had a lower risk of fracture in Japan. Previous authors hypothesized that this may come from increased opportunity costs of religious attendance among women in the U.S., which meant that women may have difficulty in participating in regular religious attendance due to lack of support or substitution by other social networks [27]. In contrast, because Shinto and Buddhism in Japan, which do not emphasize regular religious attendance, allow Japanese women to enjoy their religiousness, resulting in more benefit from religiosity in terms of fractures.
Our study has some limitations. First, our data contain only self-reported religiosity but not the type of religion. In addition, there was no detailed explanation for the participants about the distinction between the choices in degree of religiosity, therefore, their reported degree of religiosity totally was subjective. We cannot compare the associations across other religions. Second, we were unable to obtain complete information about fractures because a few participants were lost to follow-up. However, because being lost to follow-up may be independent of religiosity, the effect on the results would be negligible. Third, it would be reasonable to divide women into pre- or post-menopause, however, we did not have data about menopause status, so that we divided them at the age of 50, which is close to the age of menopause in Japanese women [30]. Based on these limitations, including the questions about religion, the particular nature of religion in Japan, and the lack of previous studies in this research area, additional research will be required to examine the association between the degree of religiosity and fractures.
Conclusion
We demonstrated that slightly religious people, but not somewhat or very religious people, had a lower incidence of fracture than nonreligious people, although the T-scores were similar regardless of the degree of religiosity. We hypothesize that this population may be more careful about avoiding behaviors that might cause fractures.
Supplementary Information
Abbreviations
- IQR
Interquartile range
- AOR
Adjusted odds ratio
- CI
Confidence interval
- ICD-10
International classification of disease 10
- BMD
Bone mineral density
- WHO
World Health Organization
- YAM
Young adult mean
- SD
Standard deviation
- BMI
Body mass index
- GEE
Generalized estimating equation
- OR)
Odds ratios
Authors’ contributions
DK led and organized this study. DK analyzed data, interpreted the results and wrote draft. HK contributed to discussion. TS confirmed the analyses and interpretation and contributed to discussion. All authors read and approved the final manuscript.
Funding
The study was funded by departmental resources.
Availability of data and materials
The data underlying this article cannot be shared publicly due to patient privacy. The data will be shared on reasonable request to the corresponding author.
Declarations
Ethics approval and consent to participate
St. Luke’s Ethics Committee Institutional Review Board approved this study on March 27, 2019. (approval number: 18-R203).
Competing interests
Daiki Kobayashi, Hironori Kuga, and Takuro Shimbo declare that they have no conflict of interest.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Kobayashi D, First MB, Shimbo T, Kanba S, Hirano Y. Association of self-reported religiosity with the development of major depression in multireligious country Japan. Psychiatry Clin Neurosci. 2020;74(10):535–541. doi: 10.1111/pcn.13087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bruce MA, Martins D, Duru K, et al. Church attendance, allostatic load and mortality in middle aged adults. PLoS ONE. 2017;12(5):e0177618. doi: 10.1371/journal.pone.0177618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Zimmer Z, Chiu CT, Saito Y, Lin YH, Ofstedal MB, Jagger C. Does religious activity distinguish the mortality experiences of older Taiwanese? an analysis using eighteen years of follow-up data. J Relig Health. 2020;59(1):289–308. doi: 10.1007/s10943-019-00778-x. [DOI] [PubMed] [Google Scholar]
- 4.McCullough ME, Hoyt WT, Larson DB, Koenig HG, Thoresen C. Religious involvement and mortality: a meta-analytic review. Health Psychol. 2000;19(3):211–222. doi: 10.1037/0278-6133.19.3.211. [DOI] [PubMed] [Google Scholar]
- 5.Chida Y, Steptoe A, Powell LH. Religiosity/spirituality and mortality. a systematic quantitative review. Psychother Psychosom. 2009;78(2):81–90. doi: 10.1159/000190791. [DOI] [PubMed] [Google Scholar]
- 6.Powell LH, Shahabi L, Thoresen CE. Religion and spirituality. Linkages to physical health Am Psychol. 2003;58(1):36–52. doi: 10.1037/0003-066x.58.1.36. [DOI] [PubMed] [Google Scholar]
- 7.Feinstein M, Liu K, Ning H, Fitchett G, Lloyd-Jones DM. Burden of cardiovascular risk factors, subclinical atherosclerosis, and incident cardiovascular events across dimensions of religiosity: the multi-ethnic study of atherosclerosis. Circulation. 2010;121(5):659–666. doi: 10.1161/CIRCULATIONAHA.109.879973. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kobayashi D, Shimbo T, Takahashi O, Davis RB, Wee CC. The relationship between religiosity and cardiovascular risk factors in Japan: a large-scale cohort study. J Am Soc Hypertens. 2015;9(7):553–562. doi: 10.1016/j.jash.2015.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bharmal NH, McCarthy WJ, Gadgil MD, Kandula NR, Kanaya AM. The association of religious affiliation with overweight/obesity among South Asians: the Mediators of Atherosclerosis in South Asians Living in America (MASALA) Study. J Relig Health. 2018;57(1):33–46. doi: 10.1007/s10943-016-0290-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Cline KM, Ferraro KF. Does religion increase the prevalence and incidence of obesity in adulthood? J Sci Study Relig. 2006;45(2):269–281. doi: 10.1111/j.1468-5906.2006.00305.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dodor BA, Robinson MA, Watson R, Meetze D, Whicker R., Jr The impact of religiosity on substance abuse and obesity in African Americans. J Relig Health. 2018;57(4):1315–1328. doi: 10.1007/s10943-017-0501-2. [DOI] [PubMed] [Google Scholar]
- 12.Zimmer Z, Jagger C, Chiu CT, Ofstedal MB, Rojo F, Saito Y. Spirituality, religiosity, aging and health in global perspective: a review. SSM Popul Health. 2016;2:373–381. doi: 10.1016/j.ssmph.2016.04.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Organization WH. International Statistical Classification of Diseases and Related Health Problems 10th Revision. 2016; https://www.who.int/classifications/icd/icdonlineversions/en/. Accessed 9 May 2019.
- 14.Kanis JA, Melton LJ, 3rd, Christiansen C, Johnston CC, Khaltaev N. The diagnosis of osteoporosis. J Bone Miner Res. 1994;9(8):1137–1141. doi: 10.1002/jbmr.5650090802. [DOI] [PubMed] [Google Scholar]
- 15.Soen S, Fukunaga M, Sugimoto T, et al. Diagnostic criteria for primary osteoporosis: year 2012 revision. J Bone Miner Metab. 2013;31(3):247–257. doi: 10.1007/s00774-013-0447-8. [DOI] [PubMed] [Google Scholar]
- 16.Consultation WHOE. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363(9403):157–163. doi: 10.1016/S0140-6736(03)15268-3. [DOI] [PubMed] [Google Scholar]
- 17.Streeten EA, McBride DJ, Lodge AL, et al. Reduced incidence of hip fracture in the old order Amish. J Bone Miner Res. 2004;19(2):308–313. doi: 10.1359/JBMR.0301223. [DOI] [PubMed] [Google Scholar]
- 18.Sharma S, Tandon VR, Mahajan A, Kour A, Kumar D. Preliminary screening of osteoporosis and osteopenia in urban women from Jammu using calcaneal QUS. Indian J Med Sci. 2006;60(5):183–189. doi: 10.4103/0019-5359.25679. [DOI] [PubMed] [Google Scholar]
- 19.Yilmaz S, Kart-Koseoglu H, Guler O, Yucel E. Effect of prayer on osteoarthritis and osteoporosis. Rheumatol Int. 2008;28(5):429–436. doi: 10.1007/s00296-007-0466-7. [DOI] [PubMed] [Google Scholar]
- 20.Kobayashi T. How have Japanese attitues and behabior on Religion Changed?from the ISSP Survey on Religion: survey results in Japan (Article in Japanese) The NHK monthly report on broadcast research. 2019;69(4):52–72. [Google Scholar]
- 21.Svensson NH, Larrabee Sonderlund A, Wehberg S, Hvidt NC, Sondergaard J, Thilsing T. The association between individualised religiosity and health behaviour in Denmark: are social networks a mediating factor? J Relig Health. 2022;61(6):4738–4757. doi: 10.1007/s10943-022-01650-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Saito T, Kondo K, Murata C, et al. Gender and regional differences in going-out, social, and leisure activities among older adults Findings from the JAGES Project (Article in Japanese) Jpn J Public Health. 2015;62(10):596–608. doi: 10.11236/jph.62.10_596. [DOI] [PubMed] [Google Scholar]
- 23.Yokota Y. Fishery and consumption of the sea urchin in Japan. Sea Urchin: From Basic Biology to Aquaculture Balkema Publ, Rotterdam. 2002:129–138.
- 24.Kiuchi K. Miso & Natto. food. Culture. 2001;3:7–10. [Google Scholar]
- 25.Calderon-Garcia JF, Moran JM, Roncero-Martin R, Rey-Sanchez P, Rodriguez-Velasco FJ, Pedrera-Zamorano JD. Dietary habits, nutrients and bone mass in Spanish premenopausal women: the contribution of fish to better bone health. Nutrients. 2012;5(1):10–22. doi: 10.3390/nu5010010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Institute for Clinical Systems Improvement (ICSI). Diagnosis and treatment of osteoporosis, ninth edition.
- 27.Mukerjee S, Venugopal A. Religiosity and health through the decades: is there a gender difference? Am J Health Promot. 2018;32(4):1028–1041. doi: 10.1177/0890117116687886. [DOI] [PubMed] [Google Scholar]
- 28.Maselko J, Kubzansky LD. Gender differences in religious practices, spiritual experiences and health: results from the US general social survey. Soc Sci Med. 2006;62(11):2848–2860. doi: 10.1016/j.socscimed.2005.11.008. [DOI] [PubMed] [Google Scholar]
- 29.Zhang W. Religious participation, gender differences, and cognitive impairment among the oldest-old in China. J Aging Res. 2010;2010:160294. doi: 10.4061/2010/160294. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Kono S, Sunagawa Y, Higa H, Sunagawa H. Age of menopause in Japanese women: trends and recent changes. Maturitas. 1990;12(1):43–49. doi: 10.1016/0378-5122(90)90059-F. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article cannot be shared publicly due to patient privacy. The data will be shared on reasonable request to the corresponding author.