The Impact of Persistent Low Weight Status on the Occurrence of Vertebral Fractures: A Nationwide Population-Based Cohort Study

Sang-Min Park; Jiwon Park; Sangsoo Han; Hae-Dong Jang; Jae-Young Hong; Kyungdo Han; Ho-Joong Kim; Jin S. Yeom

doi:10.3346/jkms.2023.38.e48

Journal List > J Korean Med Sci > v.38(7) > 1516081315

Go to TopGo to Top Go to BottomGo to Bottom

TOOLS

Park, Park, Han, Jang, Hong, Han, Kim, and Yeom: The Impact of Persistent Low Weight Status on the Occurrence of Vertebral Fractures: A Nationwide Population-Based Cohort Study

Original Article

Musculoskeletal Disorders, Rehabilitation & Sports Medicine

Journal of Korean Medical Science 2023; 38(7): e48.

Published online: 3 February 2023

DOI: https://doi.org/10.3346/jkms.2023.38.e48

The Impact of Persistent Low Weight Status on the Occurrence of Vertebral Fractures: A Nationwide Population-Based Cohort Study

Sang-Min Park^1,^*

, Jiwon Park^2,^*

, Sangsoo Han³

, Hae-Dong Jang⁴

, Jae-Young Hong²

, Kyungdo Han⁵

, Ho-Joong Kim¹

, Jin S. Yeom¹

¹Spine Center and Department of Orthopaedic Surgery, Seoul National University College of Medicine and Seoul National University Bundang Hospital, Seongnam, Korea.

²Department of Orthopedic Surgery, Korea University Ansan Hospital, Ansan, Korea.

³Department of Emergency Medicine, Soonchunhyang University Bucheon Hospital, Bucheon, Korea.

⁴Department of Orthopaedic Surgery, Soonchunhyang University Bucheon Hospital, Bucheon, Korea.

⁵Department of Statistics and Actuarial Science, Soongsil University, Seoul, Korea.

Address for Correspondence: Jae-Young Hong, MD, PhD. Department of Orthopedics, Korea University Hospital, Ansan, 123 Jeokgeum-ro, Danwon-gu, Ansan 15355, Republic of Korea. osspine@korea.ac.kr

Equal contributor:

^*Sang-Min Park and Jiwon Park equally contributed to this work.

Received 25 August 2022 Accepted 30 October 2022

The Korean Academy of Medical Sciences

https://creativecommons.org/licenses/by-nc/4.0/

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

Although, being underweight is commonly associated with osteoporosis and sarcopenia, its association with vertebral fractures (VFs), is less well researched. We investigated the influence of cumulative, chronic periods of low weight and changes in body weight on VF development.

Methods

We used a nationwide, population-based database with data on people (> 40 years) who attended three health screenings between January 1, 2007, and December 31, 2009 to assess the incidence of new VFs. Cox proportional hazard analyses were used to establish the hazard ratios (HRs) for new VFs based on the degree of body mass index (BMI), the cumulative numbers of underweight participants, and temporal change in weight.

Results

Of the 561,779 individuals in this analysis, 5,354 (1.0%) people were diagnosed three times, 3,672 (0.7%) were diagnosed twice, and 6,929 (1.2%) were diagnosed once. The fully adjusted HR for VFs in underweight individuals was 1.213. Underweight individuals diagnosed only once, twice, or three times had an adjusted HR of 0.904, 1.443, and 1.256, respectively. Although the adjusted HR was higher in adults who were consistently underweight, there was no difference in those who experienced a temporal change in body weight. BMI, age, sex, and household income were significantly associated with VF incidence.

Conclusion

Low weight is a risk factor for VFs in the general population. Given the significant correlation between cumulative periods of low weight and the risk of VFs, it is necessary to treat underweight patients before a VF to prevent its development and other osteoporotic fractures.

Graphical Abstract

Keywords: Claim Database, Vertebral Fracture, Osteoporosis, Underweight

INTRODUCTION

Vertebral fractures (VFs) are one of the most prevalent types of fractures and are becoming increasingly common in an ageing population.1 2 3 4 Over 700,000 VFs occur in the United States each year, with more than 40% of women experiencing at least one VF during their lifetime.5 VFs cause a lower quality of life and increased social expenses, particularly in older adults with osteoporosis.6 As a result, it is critical to identify and improve the risk factors for VFs.

Previous research has identified several risk factors for osteoporosis, such as weight, body mass index (BMI), drinking, smoking, and physical activity.7 8 9 In addition to bone mass density (BMD), BMI and body weight are the most important parameters in the World Health Organization (WHO) Fracture Risk Assessment tool.10 11 According to previous studies, BMI is either a risk or a protective factor of bone fractures.12 13 14 It has been found that decreased activity and muscular function in obese people leads to frequent falls, which contributes to an increase in fractures.14 On the other hand, a lower BMI has been linked to an increased risk of muscle loss, known as sarcopenia and osteoporosis, leading to an increase in the risk of fractures.12 13 However, only a few studies have investigated the relationship between low body weight and VFs.15 In addition, there is no study on the incidence of VFs in relation to the status change of low body weight, nor is there a large-scale study in East Asia examining the association between low body weight and VFs. Using data from the population-based, nationwide Korean National Health Insurance Service (KNHIS) database, we previously studied the association between the severity of underweight and all types of fractures,16 as well as the relationship between changes in underweight status and hip fractures.17 18

To the best of our knowledge, no study has used a nationwide population-based database to investigate the impact of being underweight on VFs. The purpose of this study was to investigate how being underweight affects the incidence of VFs, as well as the influence of cumulative, longitudinal periods of low BMI and changes in body weight on VF development.

METHODS

Data source and study design

We used nationwide, population-based KNHIS database. The KNHIS database includes all Korean population health information, such as patient diagnoses (International Classification of Disease, 10^th revision [ICD-10]), prescriptions, and procedures. Therefore, this database contains all medical information for the entire Korean population (approximately 50 million people). Furthermore, all Korean adults over the age of 40 years attend general health screening once every two years.19 In this health screening data, not only are regular health screening records, such as anthropometric measurements, lifestyle questionnaires, and basic laboratory results recorded, but also socioeconomic information, prescription records, hospitalization records, outpatient records, and the date of death of the insured Korean population. Based on this database, we created a longitudinal cohort using health screening data from January 1, 2002, to December 31, 2017, and extracted data on people over the age of 40 years who underwent three serial general health screenings between January 1, 2007, and December 31, 2009. To eliminate the effects of previous fractures, we excluded people who had a history of osteoporotic fractures before the health screening date. A one-year lag period after health screening was also applied to increase the effect of being underweight. Individuals with missing data values were excluded. Ultimately, 561,779 participants were included in this study (Fig. 1). In this cohort, we followed up instances of VFs after health screening until the cohort termination date or death. A VF was defined as a fracture for which a claim for hospitalization or outpatient treatment was received by fracture code after the general health-screening date.

Fig. 1

Flow chart of study population.

Evaluation of underweight and VF

Weight and height were extracted from the results of the general health screening. The BMI was calculated as the ratio of weight (kg) to height (m) squared (kg/m²). Based on the WHO Asia-Pacific regional guidelines, BMI was classified as underweight (< 18.5 kg/m²), overweight (≥ 23.0 kg/m²), obese (≥ 25.0 kg/m²), and severely obese (≥ 30.0 kg/m²), respectively.20 21

During each health screening, the patients’ underweight status would be noted. The number of underweight individuals was defined as the cumulative number of underweight patients diagnosed at each health screening. To estimate the influence of temporal trends in BMI changes on VFs, we compared the diagnosis of an underweight status at the first health screening with that at the third (last) health screening. The study population was divided into four groups according to changes in underweight status: underweight to underweight (U-to-U), underweight to non-underweight (U-to-N), non-underweight to underweight (N-to-U), and non-underweight to non-underweight (N-to-N).

To identify VFs, we used ICD-10 codes (S22.0 Fracture of thoracic vertebra, S22.1 Multiple fracture of thoracic spine, S32.0 Fracture of lumbar vertebra, S32.8 Multiple fractures of lumbar spine, T08 fracture of spine, level unspecified) to find the claims in the database.22 23 24 VF was defined as each fracture code with admission, repeated outpatient clinic claims, or emergency clinic visits.

Covariates and measurements

Baseline demographic data in this study were defined as that in the last health screening data. These basic characteristics included socioeconomic data, laboratory results (total cholesterol, glucose, blood pressure), answers to lifestyle questionnaires (regular exercise, smoking, alcohol drinking), anthropometric measurements (height, weight, waist circumference), and medical histories, which included hypertension, diabetes, dyslipidemia, and chronic kidney disease.25 Regarding medical history, comorbidities were described if a record confirmed this at the health screening or treatment was received in the past medical claim data.

Smoking status was classified as non-smoker, past smoker, or current smoker. Alcohol consumption was classified as non-drinkers, mild drinkers (less than 30 g/day), or heavy drinkers (more than 30 g/day) according to the amount of alcohol consumed per day. Regular exercise was defined as the level of physical activity that included at least 20 minutes of vigorous physical activity over three or more days per week or 30 minutes of moderate to intense physical activity five or more days per week. Income status was divided into low as income in the bottom 20% of the insured’s annual income, and normal. The ICD-10 codes used in this study are listed in Appendix 1.

Statistical analysis

Baseline characteristics of this study population are presented as mean ± standard deviation or counts and percentages in parentheses according to the cumulative number of underweight patients. The incidence rate (IR) was defined as the IR per 1,000 person-years (PY). The risk of VF development was analyzed using Cox proportional regression analysis to calculate the hazard ratios (HRs) and 95% confidence intervals (95% CIs) between VFs and the number of underweight patients. To reduce the bias of co-variates, we analyzed the HRs for unadjusted and three adjusted models: Model 1 was adjusted for age and sex; Model 2 was adjusted for age, sex, and other environmental factors such as smoking, alcohol consumption, physical activity, and household income; Model 3 was fully adjusted for age, sex, other environmental factors (smoking, alcohol consumption, physical activity, and household income), and comorbidities (diabetes, hypertension, dyslipidemia, and chronic kidney disease). Subgroup analysis was performed based on several factors, including age (< 65 years, ≥ 65 years), sex, smoking, alcohol consumption, physical activity, and household income.

Statistical analysis was performed using the SAS software 9.3 (SAS Institute, Cary, NC, USA). Student’s t-test for continuous variables and the chi-square test for categorical variables were used, and two-tailed statistical significance was considered at P < 0.05.

RESULTS

Baseline characteristics

The baseline characteristics of the adults according to the cumulative number of underweight participants at each health screening examination are summarized in Table 1. Of the 561,779 individuals included in this analysis, 545,824 (97.2%) were never diagnosed as underweight. Regarding those underweight, 5,354 (1.0%) people were diagnosed three times, 3,672 (0.7%) people were diagnosed twice, and 6,929 (1.2%) people were diagnosed once of the three screenings. Except for age, the baseline characteristics of individuals in the four groups who were never diagnosed, were diagnosed once, twice, and three times revealed statistically significant variations in all categories evaluated. Individuals in the underweight group were more likely than those in the non-underweight group to be current smokers, have no alcohol consumption, engage in regular physical activity, and have low household income, regardless of the duration the participant was underweight.

Table 1

Baseline characteristics of this study according to the cumulative number of the presence of underweight

Variables		The cumulative number of the presence of underweight^a at each health screening				P value
Variables		0	1	2	3	P value
Participants		545,824	6,929	3,672	5,354
Age, yr		49.69 ± 7.14	49.73 ± 7.77	49.67 ± 7.78	49.58 ± 7.87	0.682
Sex						< 0.001
	Men	403,926 (74)	4,270 (61.63)	2,373 (64.62)	3,495 (65.28)
	Women	141,898 (26)	2,659 (38.37)	1,299 (35.38)	1,859 (34.72)
Height, cm		165.72 ± 8.03	163.99 ± 8.23	164.4 ± 7.9	164.81 ± 7.86	< 0.001
Weight, kg		66.39 ± 10.19	51.26 ± 6.2	49.62 ± 5.22	47.35 ± 4.99	< 0.001
BMI, kg/m²		24.1 ± 2.73	19.02 ± 1.46	18.32 ± 0.87	17.39 ± 0.77	< 0.001
Smoking						< 0.001
	Non	257,094 (47.1)	3,699 (53.38)	1,808 (49.24)	2,683 (50.11)
	Ex	127,920 (23.44)	960 (13.85)	515 (14.03)	580 (10.83)
	Current	160,810 (29.46)	2,270 (32.76)	1,349 (36.74)	2,091 (39.05)
Alcohol consumption^b						< 0.001
	Non	234,442 (42.95)	3,738 (53.95)	1,916 (52.18)	2,817 (52.61)
	Mild to moderate	264,341 (48.43)	2,826 (40.79)	1,539 (41.91)	2,247 (41.97)
	Heavy	47,041 (8.62)	365 (5.27)	217 (5.91)	290 (5.42)
Physical activity^c		122,611 (22.46)	1,010 (14.58)	476 (12.96)	655 (12.23)	< 0.001
Low household income^d		112,089 (20.54)	1,570 (22.66)	823 (22.41)	1,222 (22.82)	< 0.001
Comorbidities
	Diabetes	51,701 (9.47)	409 (5.9)	205 (5.58)	244 (4.56)	< 0.001
	Hypertension	157,011 (28.77)	1,052 (15.18)	490 (13.34)	641 (11.97)	< 0.001
	Dyslipidemia	103,324 (18.93)	620 (8.95)	291 (7.92)	352 (6.57)	< 0.001
	CKD	42,114 (7.72)	378 (5.46)	202 (5.5)	315 (5.88)	< 0.001

Numeric parameters are expressed as mean ± standard deviation and categorical parameters are expressed as counts and percentages in parentheses.

BMI = body mass index, CKD = chronic kidney disease.

^aUnderweight was defined as body mass index under 18.5 kg/m².

^bAlcohol consumption was divided into 3 categories; Non (no alcohol consumption), Mild (under 30g/day consumption), and heavy (over 30 g/day consumption).

^cPhysical activity is defined as performing over 30 minutes moderate intensity exercise over 5 times per a week or over 20 minutes vigorous intensity exercise over 3 times per a week.

^dLow household income is defined as total household monthly income belongs to lower 20% group among Korean entire population.

The incidence and risk of VFs according to BMI

The IR was 2.04/1000 PY in the underweight group. The unadjusted and adjusted HRs (models 1, 2, and 3) were also significantly higher in the underweight group. The fully adjusted HR (model 3) for VFs according to BMI were 1.213 (1.037–1.417), 0.982 (0.925–1.042), 1.053 (0.994–1.117), and 1.081 (0.929–1.259), respectively (Table 2).

Table 2

The risk of vertebral fracture according to body mass index using Cox regression analysis

Body mass index, kg/m²	No. of fracture	IR^a	Unadjusted			Model 1			Model 2			Model 3
Body mass index, kg/m²	No. of fracture	IR^a	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value
< 18.5	167	2.04	1.33	1.137–1.555		1.232	1.054–1.441		1.217	1.040–1.423		1.213	1.037–1.419
18.5 ≤ and < 23	2,545	1.54	1		0.002	1		0.015	1		0.022	1		0.020
23 ≤ and < 25	1,923	1.47	0.954	0.899–1.012		0.974	0.918–1.034		0.979	0.923–1.039		0.982	0.925–1.042
25 ≤ and < 30	2,232	1.53	0.995	0.940–1.053		1.043	0.985–1.105		1.048	0.990–1.110		1.053	0.994–1.117
30 ≤	182	1.56	1.014	0.872–1.178		1.074	0.924–1.248		1.073	0.923–1.247		1.081	0.929–1.259

IR = incidence rate, HR = hazard ratio, 95% CI = 95% confidence interval.

^aIncidence rate is defined as incidence rate per 1,000 person-year.

Model 1 was adjusted by age, and sex; Model 2 was adjusted by age, sex, and other environmental factors such as smoke, alcohol consumption, physical activity, household income; Model 3 was fully adjusted by age, sex, other environmental factors (smoke, alcohol consumption, physical activity, household income), and comorbidities (diabetes, hypertension, dyslipidemia, chronic kidney disease).

The incidence and risk of VFs according to the cumulative number of underweight

A total of 7,049 VFs were observed (1.2%). The IR of VFs was 1.53/1000 PY in the once diagnosed underweight group, 2.35/1000 PY in the twice diagnosed underweight group, and 1.98/1000 PY in the three times diagnosed underweight group, with the IR being greater in the underweight group overall. In contrast, there was no statistically significant serial increase in fractures according to the number of underweight diagnoses; nonetheless, it was significantly greater in the groups with two or more underweight diagnoses. In the multivariate-adjusted analysis, this correlation was statistically significant. Underweight individuals who were diagnosed only once, twice, or three times had an adjusted HR (Model 3) of 0.904 (0.731–1.117), 1.443 (1.140–1.827), and 1.256 (1.028–1.482) for VFs, respectively (Table 3).

Table 3

The risk of vertebral fracture according to the cumulative number of the presence of underweight using Cox regression analysis

Cumulative number of underweight	No. of fracture	IR^a	Unadjusted			Model 1			Model 2			Model 3
Cumulative number of underweight	No. of fracture	IR^a	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value
0	6,806	1.51	1		< 0.001	1		0.003	1		0.005	1		0.005
1	87	1.53	1.016	0.822–1.255		0.912	0.738–1.126		0.904	0.731–1.117		0.901	0.729–1.114
2	70	2.35	1.554	1.228–1.967		1.464	1.157–1.853		1.443	1.140–1.827		1.438	1.136–1.821
3	86	1.98	1.312	1.061–1.623		1.268	1.037–1.509		1.256	1.028–1.482		1.241	1.002–1.480

IR = incidence rate, HR = hazard ratio, 95% CI = 95% confidence interval.

^aIncidence rate is defined as incidence rate per 1,000 person-year.

The risk of VF according to temporal trends in BMI changes

The IR was 1.51/1000 PY in the N-to-N group, 2.02/1000 PY in the N-to-U group, 1.72/1000 PY in the U-to-N group, and 2.05/1000 PY in the U-to-U group. After multivariate-adjusted analysis, adults in the U-to-U group had a significantly higher risk of VF (HR, 1.239; 95% CI, 1.022–1.502). Although the adjusted HR was higher in adults who were consistently underweight, there was no difference in those who experienced a temporal change in body weight (Table 4).

Table 4

The risk of vertebral fracture according to temporal trends in body mass index changes using Cox regression analysis

Underweight status^a	No. of fracture	IR^b	Unadjusted			Model 1			Model 2			Model 3
Underweight status^a	No. of fracture	IR^b	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value
N to N	6,827	1.51	1		0.001	1		0.060	1		0.098	1		0.108
N to U	61	2.02	1.338	1.040–1.721		1.160	0.901–1.493		1.146	0.890–1.474		1.142	0.887–1.470
U to N	55	1.72	1.143	0.876–1.490		1.098	0.842–1.432		1.089	0.835–1.420		1.086	0.832–1.416
U to U	106	2.05	1.359	1.122–1.646		1.263	1.043–1.531		1.242	1.025–1.505		1.239	1.022–1.502

IR = incidence rate, HR = hazard ratio, 95% CI = 95% confidence interval, N = non-underweight (body mass index ≥ 18.5 kg/m²), U = underweight (body mass index < 18.5 kg/m²).

^aTemporal changes of underweight status (first to 3rd health screening) are divided into four groups: non-underweight to non-underweight, non-underweight to underweight, underweight to non-underweight, and underweight to underweight.

^bIncidence rate is defined as incidence rate per 1,000 person-year.

Subgroup analysis

VFs were more likely to occur in underweight adults who were under the age of 65 (1.203; 95% CI, 1.006–1.440), men (1.289; 95% CI, 1.046–1.588), and low-income (1.364; 95% CI, 1.077–1.727) (Fig. 2). In the stratified analysis, none of the covariates had a statistically significant association with the risk of VFs in groups defined by the cumulative number of underweight diagnoses (P for interaction > 0.05) (Fig. 3).

Fig. 2

Incidence and hazard ratios of vertebral fractures according to several subgroups and body mass index.

BMI = body mass index, CI = confidence interval.

Fig. 3

Incidence and hazard ratios of vertebral fractures according to several subgroups and the cumulative numbers of underweight participants.

CI = confidence interval.

Ethics statement

The study protocol was approved by the Institutional Review Board of Korea University Ansan Hospital (approval no. K2021-2601-001). The ethics committees of Korea University Ansan Hospital have waived the requirement to obtain informed consent as the register data analysed in this study are in anonymised and deidentified format. This study was performed in accordance with the tenets of the Declaration of Helsinki, and all research methods were carried out in accordance with appropriate regulations and guidelines.

DISCUSSION

To the best of our knowledge, this is the first study to determine the risk of VFs associated with the cumulative burden of low body weight in a large nationwide population-based cohort study. Through various analyses, we confirmed the following: 1) Underweight status increased the risk of VFs. 2) The cumulative burden of being underweight increased the risk of VFs. 3) There was no increase in the risk of VFs if the cumulative burden of being underweight was small or if there was a temporal change in body weight. 4) Among several factors, under 65-year, male sex and low household income were significantly more affected.

Osteoporosis is a risk factor for fractures, particularly when combined with physical activity, smoking, alcohol consumption, and body weight.7 8 The relationship between fractures and body weight varies according to the location of the fracture. Weight gain is associated with a considerable reduction in hip fractures.26 VF research continues to be contentious, with studies indicating that being underweight is both a risk and protective factor for VFs.13 27 However, there is no conclusive evidence of the association between low weight and VFs.

Although the mechanism by which low weight increases the frequency of VF is unknown, this study revealed that underweight is a risk factor for increased VFs. In humans, being underweight is often related to malnutrition, which is hypothesized to lead to osteoporosis.28 Malnutrition causes bone loss and, eventually, osteoporosis.29 30 In addition, low BMI is highly associated with the development of sarcopenia. Previous research has shown that malnourished individuals are more prone to sarcopenia.31 Physical capacity and muscular function are diminished as a result of sarcopenia, resulting in falls that increase the likelihood of VF.32 33 In conclusion, lower BMI is thought to be associated with lower BMD levels and decreased muscle strength. However, this study was a population-based study using the ICD-10 diagnostic code, and the actual skeletal muscle index and BMD scores of the patients could not be determined. Although this study cannot clearly elucidate the relationship between low BMI, BMD, and skeletal muscle index, the fact that low BMI is related to VFs was confirmed using the large population-based database.

Low body weight was investigated as a risk factor for VFs after controlling for several variables. In a further analysis, transient underweight status (cumulative number = 1) did not increase the risk of VF (HR, 0.901; 95% CI, 0.729–1.114). There was no increase in risk, even when changing from normal weight to underweight or from underweight to normal weight. In other words, it can be considered that VFs increase only when the body weight is continuously low. This suggests that being underweight does not increase fractures directly; rather, fractures occur as a result of low bone density or muscle loss following underweight.

In the subgroup analysis, the effect of being underweight was greater, especially those who were under 65 years of age, men, and who had low household income. Thus, the importance of being underweight as a risk factor for VFs varies with age, sex, and household income. Although the specific mechanism of VFs in young individuals remains unknown in this study, a possible explanation is that younger underweight individuals have quicker bone loss due to a relatively rapid change in metabolism compared to older individuals. The finding that underweight men had a greater risk of VFs than women after correcting for age is consistent with previous research indicating that men have a greater risk of fractures than women.5 34 Low household income is also associated with malnutrition, which may have contributed to the low BMI and skeletal muscle index, which may have led to an increase in fractures.

To our knowledge, this is the only study to evaluate the risk of VF in the underweight population using a nationwide database. The major strength of this study is that it used national health insurance data in which all citizens were enrolled. This is a large amount of data, and the database is maintained constantly. Consequently, it produces significant results that are representative of the real world. However, this study had several limitations. First, the T-scores of the BMD results could not be directly verified. Being underweight is thought to have had an effect on the BMD score, but the direct effect was unknown in this study. In addition, grip strength and muscle mass for analyzing sarcopenia could not be further analyzed. This limitation of further data analysis is due to the limitation of KNHIS database, which consists of only basic demographics, diagnostic, radiographic, and treatment codes. Second, it was difficult to determine the exact number of VFs. Because VFs are frequently asymptomatic or weak, some patients do not visit outpatient clinics or hospitals. Third, this study used a nationwide database from national health insurance services in one nation, so it is difficult to apply it to multiple ethnicities. Finally, because the diagnosis of VFs in this investigation was made using the fracture diagnostic code, we were unable to confirm that all VFs were accurately diagnosed. To identify VFs, an algorithm used in this study identical to that employed in prior studies was applied.22 23 24 Validation studies are the best method to confirm the proposed algorithm of diagnostic codes.24 According to our recent study, the sensitivity was 62.5% and the positive predictive value of operational definition was 59.7%, which was lower than that of hip35 or wrist fractures,36 but higher than that of other VFs.37 To diagnose VFs as accurately as possible, we used a one-year lag time period after being diagnosed as underweight and excluded patients with previous VFs. Because this study used the most conservative algorithm, it is quite likely that the incidence rate of VFs was slightly underestimated, as previously stated.

This study investigated whether being underweight is an important factor that increases the risk of VF in the Korean population using a nationwide population-based cohort. The risk of VFs was particularly high among people who were persistently underweight, and men under the age of 65 and those with low household income had a higher risk of VFs. Therefore, patients with these risk factors require treatment to reduce the risk of VF.

ACKNOWLEDGMENTS

We would like to thank Editage (www.editage.co.kr) for English language editing.

Notes

Funding: This work was partially supported by CGBio research grants (800-20200433). The funder played no role in the study design, data collection and analysis, decision to publish, or manuscript preparation.

Disclosure: The authors have no potential conflicts of interest to disclose.

Author Contributions:

Conceptualization: Park SM, Park J, Han K, Hong JY.
Data curation: Han K, Hong JY.
Formal analysis: Han K, Hong JY.
Funding acquisition: Park SM.
Investigation: Park SM, Park J.
Methodology: Han S, Jang HD.
Software: Han K, Hong JY.
Validation: Kim HJ, Yeom JS.
Visualization: Park SM.
Writing - original draft: Park SM.
Writing - review & editing: Park J, Kim HJ, Yeom JS.

Appendix

Appendix 1

Definitions of covariates and measurements

CKD = chronic kidney disease, N/A = not applicable.

^aCombination: 1 = ICD-10-CM code and medication; 2 = number of diagnosis; and 3 = diagnosis test or treatment.

References

1. Warriner AH, Patkar NM, Curtis JR, Delzell E, Gary L, Kilgore M, et al. Which fractures are most attributable to osteoporosis? J Clin Epidemiol. 2011; 64(1):46–53. PMID: 21130353.

2. Jang HD, Kim EH, Lee JC, Choi SW, Kim K, Shin BJ. Current concepts in the management of osteoporotic vertebral fractures: a narrative review. Asian Spine J. 2020; 14(6):898–909. PMID: 33373513.

3. Umehara T, Inukai A, Kuwahara D, Kaneyashiki R, Kaneguchi A, Tsunematsu M, et al. Physical functions and comorbidity affecting collapse at 4 or more weeks after admission in patients with osteoporotic vertebral fractures: a prospective cohort study. Asian Spine J. 2022; 16(3):419–431. PMID: 33940771.

4. Cho ST, Kim SJ, Nam BJ, Kim KW, Lee GH, Kim JH. Absolute bed rest duration of 3 days for osteoporotic vertebral fractures: a retrospective study. Asian Spine J. 2022; 16(6):898–905. PMID: 35527538.

5. Johnell O, Kanis JA, Odén A, Sernbo I, Redlund-Johnell I, Petterson C, et al. Fracture risk following an osteoporotic fracture. Osteoporos Int. 2004; 15(3):175–179. PMID: 14691617.

6. Kim TY, Jang S, Park CM, Lee A, Lee YK, Kim HY, et al. Trends of incidence, mortality, and future projection of spinal fractures in Korea using nationwide claims data. J Korean Med Sci. 2016; 31(5):801–805. PMID: 27134505.

7. Abrahamsen B, Brask-Lindemann D, Rubin KH, Schwarz P. A review of lifestyle, smoking and other modifiable risk factors for osteoporotic fractures. Bonekey Rep. 2014; 3:574. PMID: 25228987.

8. Lee CY, Back GY, Lee SH. Relationship between type 2 diabetes mellitus and lumbar bone mineral density in postmenopausal women. Asian Spine J. 2021; 15(6):721–727. PMID: 34551503.

9. Lee J, Chang G, Kang H, Ham DW, Lee JS, Jung HS, et al. Impact of bone mineral density on the incidence of age-related vertebral fragility fracture. J Korean Med Sci. 2020; 35(17):e116. PMID: 32356418.

10. Kanis JA, Oden A, Johansson H, Borgström F, Ström O, McCloskey E. FRAX and its applications to clinical practice. Bone. 2009; 44(5):734–743. PMID: 19195497.

11. World Health Organization. WHO Scientific Group on the Assessment of Osteoporosis at Primary Health Care Level. Summary Meeting Report. Geneva, Switzerland: World Health Organization;2004. p. 5–7.

12. Armstrong ME, Spencer EA, Cairns BJ, Banks E, Pirie K, Green J, et al. Body mass index and physical activity in relation to the incidence of hip fracture in postmenopausal women. J Bone Miner Res. 2011; 26(6):1330–1338. PMID: 21611971.

13. Tanaka S, Kuroda T, Saito M, Shiraki M. Overweight/obesity and underweight are both risk factors for osteoporotic fractures at different sites in Japanese postmenopausal women. Osteoporos Int. 2013; 24(1):69–76. PMID: 23229467.

14. Beck TJ, Petit MA, Wu G, LeBoff MS, Cauley JA, Chen Z. Does obesity really make the femur stronger? BMD, geometry, and fracture incidence in the women’s health initiative-observational study. J Bone Miner Res. 2009; 24(8):1369–1379. PMID: 19292617.

15. Cho MJ, Moon SH, Lee JH, Lee JH. Association between osteoporotic vertebral compression fractures and age, bone mineral density, and European quality of life-5 dimensions in korean postmenopausal women: a nationwide cross-sectional observational study. Clin Orthop Surg. 2021; 13(2):207–215. PMID: 34094011.

16. Han S, Park J, Nah S, Jang HD, Han K, Hong JY. Severity of underweight and risk of fracture: a Korean nationwide population-based cohort study. Sci Rep. 2022; 12(1):10153. PMID: 35710927.

17. Han S, Park J, Jang HD, Nah S, Boo J, Han K, et al. Incidence of hip fracture in underweight individuals: a nationwide population-based cohort study in Korea. J Cachex Sarcopenia Muscle;2022.

18. Han S, Park J, Jang HD, Han K, Lee C, Kim W, et al. Changes in underweight status and risk of hip fracture: a Korean nationwide population-based cohort study. J Clin Med. 2022; 11(7):11.

19. Choi EK. Cardiovascular research using the Korean national health information database. Korean Circ J. 2020; 50(9):754–772. PMID: 32725984.

20. World Health Organization. The Asia-Pacific Perspective: Redefining Obesity and Its Treatment. Geneva, Switzerland: World Health Organization;2000.

21. World Health Organization. Food and Nutrition Surveillance Systems: A Manual for Policy-Makers and Programme Managers. Geneva, Switzerland: World Health Organization;2014.

22. Ahn SH, Park SM, Park SY, Yoo JI, Jung HS, Nho JH, et al. Osteoporosis and osteoporotic fracture fact sheet in Korea. J Bone Metab. 2020; 27(4):281–290. PMID: 33317231.

23. Park SM, Ahn SH, Kim HY, Jang S, Ha YC, Lee YK, et al. Incidence and mortality of subsequent vertebral fractures: analysis of claims data of the Korea National Health Insurance Service from 2007 to 2016. Spine J. 2020; 20(2):225–233. PMID: 31589928.

24. Yu MH, Hong N, Lee S, Kim HY, Park HS, Park SM, et al. Operational definition identifying osteoporotic vertebral fractures in the claims database. J Korean Med Sci. 2022; 37(32):e249. PMID: 35971763.

25. Kwon H, Yun JM, Park JH, Cho BL, Han K, Joh HK, et al. Incidence of cardiovascular disease and mortality in underweight individuals. J Cachexia Sarcopenia Muscle. 2021; 12(2):331–338. PMID: 33619889.

26. De Laet C, Kanis JA, Odén A, Johanson H, Johnell O, Delmas P, et al. Body mass index as a predictor of fracture risk: a meta-analysis. Osteoporos Int. 2005; 16(11):1330–1338. PMID: 15928804.

27. Nevitt MC, Cummings SR, Stone KL, Palermo L, Black DM, Bauer DC, et al. Risk factors for a first-incident radiographic vertebral fracture in women > or = 65 years of age: the study of osteoporotic fractures. J Bone Miner Res. 2005; 20(1):131–140. PMID: 15619679.

28. Xue S, Kemal O, Lu M, Lix LM, Leslie WD, Yang S. Age at attainment of peak bone mineral density and its associated factors: The National Health and Nutrition Examination Survey 2005-2014. Bone. 2020; 131:115163. PMID: 31760214.

29. Coin A, Perissinotto E, Enzi G, Zamboni M, Inelmen EM, Frigo AC, et al. Predictors of low bone mineral density in the elderly: the role of dietary intake, nutritional status and sarcopenia. Eur J Clin Nutr. 2008; 62(6):802–809. PMID: 17637603.

30. Coin A, Sergi G, Benincà P, Lupoli L, Cinti G, Ferrara L, et al. Bone mineral density and body composition in underweight and normal elderly subjects. Osteoporos Int. 2000; 11(12):1043–1050. PMID: 11256896.

31. Sieber CC. Malnutrition and sarcopenia. Aging Clin Exp Res. 2019; 31(6):793–798. PMID: 31148100.

32. Landi F, Liperoti R, Russo A, Giovannini S, Tosato M, Capoluongo E, et al. Sarcopenia as a risk factor for falls in elderly individuals: results from the ilSIRENTE study. Clin Nutr. 2012; 31(5):652–658. PMID: 22414775.

33. Tokeshi S, Eguchi Y, Suzuki M, Yamanaka H, Tamai H, Orita S, et al. Relationship between skeletal muscle mass, bone mineral density, and trabecular bone score in osteoporotic vertebral compression fractures. Asian Spine J. 2021; 15(3):365–372. PMID: 32872758.

34. Kanis JA, Johansson H, Harvey NC, Gudnason V, Sigurdsson G, Siggeirsdottir K, et al. The effect on subsequent fracture risk of age, sex, and prior fracture site by recency of prior fracture. Osteoporos Int. 2021; 32(8):1547–1555. PMID: 33537845.

35. Park C, Jang S, Jang S, Ha YC, Lee YK, Yoon HK, et al. Identification and validation of osteoporotic hip fracture using the National Health Insurance Database. J Korean Hip Soc. 2010; 22(4):305–311.

36. Lee YK, Park C, Won S, Park JW, Koo KH, Ha YC, Jung HS. Validation of an operational definition to identify distal radius fractures in a National Health Insurance Database. J Hand Surg Am. 2021; 46(11):1026.e1–e7.

37. Curtis JR, Mudano AS, Solomon DH, Xi J, Melton ME, Saag KG. Identification and validation of vertebral compression fractures using administrative claims data. Med Care. 2009; 47(1):69–72. PMID: 19106733.

TOOLS

Similar articles