Loading...
 
Toggle Health Problems and D

Hip fractures are predicted by 10 factors – low Vitamin D is the biggest – Aug 2023

A machine learning-based scoring system and ten factors associated with hip fracture occurrence in the elderly

Bone https://doi.org/10.1016/j.bone.2023.116865   PDF is behind a $31 paywall
Masaru Uragami a 1, Kozo Matsushita a 1, Yuto Shibata a, Shu Takata a, Tatsuki Karasugi a, Takanao Sueyoshi a, Tetsuro Masuda a, Takayuki Nakamura a, Takuya Tokunaga a, Satoshi Hisanaga a, Masaki Yugami a, Kazuki Sugimoto a, Ryuji Yonemitsu a, Katsumasa Ideo a, Yuko Fukuma a, Kosei Takata a, Takahiro Arima a, Jyunki Kawakami a, Kazuya Maeda a, Naoto Yoshimura a…Takeshi Miyamoto a

Hip fractures are fragility fractures frequently seen in persons over 80-years-old. Although various factors, including decreased bone mineral density and a history of falls, are reported as hip fracture risks, few large-scale studies have confirmed their relevance to individuals older than 80, and tools to assess contributions of various risks to fracture development and the degree of risk are lacking. We recruited 1395 fresh hip fracture patients and 1075 controls without hip fractures and comprehensively evaluated various reported risk factors and their association with hip fracture development. We initially constructed a predictive model using Extreme Gradient Boosting (XGBoost), a machine learning algorithm, incorporating all 40 variables and evaluated the model's performance using the area under the receiver operating characteristic curve (AUC), yielding a value of 0.87. We also employed SHapley Additive exPlanation (SHAP) values to evaluate each feature's importance and ranked the top 20. We then used a stepwise selection method to determine key factors sequentially until the AUC reached a plateau nearly equal to that of all variables and identified the top 10 sufficient to evaluate hip fracture risk.
For each, we determined the cutoff value for hip fracture occurrence and calculated scores of each variable based on the respective feature importance.
Individual scores were:

  • serum 25(OH)D levels (<10 ng/ml, score 7),
  • femoral neck T-score (<−3, score 5),
  • Barthel index score (<100, score 3),
  • maximal handgrip strength (<18 kg, score 3),
  • GLFS-25 score (≥24, score 2),
  • number of falls in previous 12 months (≥3, score 2),
  • serum IGF-1 levels (<50 ng/ml, score 2),
  • cups of tea/day (≥5, score −2),
  • use of anti-osteoporosis drugs (yes, score −2), and
  • BMI (<18.5 kg/m2, score 1).

Using these scores, we performed receiver operating characteristic (ROC) analysis and the resultant optimal cutoff value was 7, with a specificity of 0.78, sensitivity of 0.75, and AUC of 0.85. These ten factors and the scoring system may represent tools useful to predict hip fracture.

Introduction
Risk of hip fracture increases dramatically with advancing age in both women and men [1]. In Japan, the proportion of elderly individuals over 65 years old is the highest in the world, and that number increased from 14.6 % in 1995 to 26.6 % in 2015. Specifically, the annual incidence of hip fractures in Japan steadily increased from 76,600 in 1992 to 193,400 in 2017 [2]. Accordingly, the annual incidence of hip fracture worldwide is expected to rise significantly due to population aging, with estimates projecting an increase from 1.26 million cases in 1990 to at least 4.5 million by 2050 [3]. Osteoporotic fracture is a major public health problem associated with increased dependency, morbidity, and mortality. Among all osteoporotic fractures, hip fracture occurs with the greatest morbidity, mortality, and costs. Population aging further contributes to an elevated average age of onset for hip fractures [4], leading to higher mortality rates, medical costs, and prolonged hospitalization durations [5,6]. Therefore, reducing its incidence in the elderly is a pressing concern in developed countries.

The Fracture Risk Assessment Tool (FRAX; The University of Sheffield, Sheffield, UK) [7] is currently used to predict probability of hip fractures and aid in making treatment decisions. This tool has been incorporated into over 100 clinical practice guidelines [8]. However, clinical risk factors included in FRAX, as well as bone mineral density, decrease in magnitude of risk as individuals age, suggesting the presence of other risk factors in the very old [7]. Therefore, the increase in average fracture age worldwide requires comparative analysis of numerous risk factors among the very old, including those included in the FRAX algorithm. Most cohort studies of individuals with a mean age of 80 years or older, however, lack data relevant to important risk factors for hip fracture, such as bone mineral density (which is typically measured by dual-energy X-ray absorptiometry) [9], [10], [[11]]. As a result, our current understanding of risk factors for hip fracture among the very old is limited.

In addition to those proposed in the FRAX algorithm, various factors have been proposed as hip fracture risks, such as: previous non-fragility fractures [12], diabetes [13], stroke [14], hyperthyroidism [15], respiratory disease [16], cardiovascular disease [17], Parkinson's disease [18], chronic liver disease [19], chronic kidney disease [20], proton pump inhibitor use [21], gastric surgery [22], hypertension [23], and low serum levels of insulin-like growth factor-1 (IGF-1) [24] or of 25-dihydroxy vitamin D (25(OH)D) [25]. Furthermore, the most common cause of hip fractures is reportedly a simple fall [26], requiring consideration of fall history [27], ambulatory ability [28], walking time [29], certification for long-term care or support (LTCI), the seven categories of long-term care or support in Japan [30], poor vision [31], nocturia [32], dementia [33], hypnotic drug use [34], lumbar spinal canal stenosis [35], grip strength [36], bed [37], locomotive syndrome [38], Barthel index [39], and living arrangement [40]. Finally, hip fracture risk is reportedly decreased by tea consumption [41], statin use [42], and use of anti-osteoporosis drugs [43]. However, the strength of relationships between each of these factors and hip fracture incidence remains unclear in the very old, and their analysis by traditional approaches is complicated by the presence of non-linear relationships, variable interactions, and risk of overfitting. Machine learning methods, on the other hand, could be advantageous in addressing these issues and are increasingly used by medical researchers to determine significant health-related factors and construct good performance prediction models using high-dimensional and complex datasets [44], [45], [46], [[47]].

Here, we analyzed 1395 cases of hip fracture patients in 23 hospitals as well as 1075 non-hip fracture control subjects. After decreasing subject number based on exclusion criteria, and matching subjects with controls by age and gender, we enrolled 604 subjects each in either hip fracture or control subject groups, with a mean age of hip fracture and control subjects of 81.8 and 81.6 years old, respectively. We then conducted comprehensive association analysis between fractures and reported risk factors including FRAX and bone mineral density, and selected significantly associated factors using a machine learning algorithm. Here, we show that 33 factors were significantly associated with hip fracture incidence after adjustment for age and gender. We identified 10 out of 40 factors sufficient to predict hip fracture incidence. Therefore, we propose a new hip fracture risk assessment tool potentially useful to predict hip fracture incidence or identify individuals at high risk or in urgent need of attention to prevent hip fractures.

Section snippets
Study design and participants
The Kumamoto STop OsteoPorotic hip fractures (K-STOP) group was organized to conduct studies of prevention of new hip fractures. K-STOP participants include 23 hospitals associated with the Department of Orthopedic Surgery of Kumamoto University, and hip fracture cases were recruited from patients admitted to 21 of those hospitals. Each facility registered participants during one year period between March 2021 and December 2022. Excluded were fracture patients who: (1) were under age 20, (2) . . . . .

Participant enrollment
Initially we recruited 1395 cases (group 1) to the study and 1075 controls (338 in group 2 and 737 in group 3). Among controls, 27 were eliminated from group 3 based on exclusion criteria (Fig. 1). Subjects with missing data related to living arrangement and those living in nursing homes or hospitalized at the time of enrollment were excluded (460 cases and 11 controls). Controls were then matched to cases based on age (±2 years) and gender in a 1:1 ratio. Final sample sizes for analysis were . . . . .

Discussion
The increasing number of hip fracture patients is a global concern, particularly in developed countries with large populations of the aged. To prevent hip fractures in the elderly requires an understanding of their pathogenesis and risk factors, although currently, there is limited research relevant to risk factors in the very elderly. Here, we evaluated participants with an average age of >80 years and identified ten promising risk factors from a pool of previously reported candidates or . . . . .

CRediT authorship contribution statement
Masaru Uragami: Writing – original draft, Project administration, Investigation, Formal analysis, Data curation, Conceptualization. Kozo Matsushita: Investigation, Data curation. Yuto Shibata: Investigation, Data curation. Shu Takata: Investigation, Data curation. Tatsuki Karasugi: Supervision, Project administration, Investigation. Takanao Sueyoshi: Investigation. Tetsuro Masuda: Investigation. Takayuki Nakamura: Investigation. Takuya Tokunaga: Investigation. Satoshi Hisanaga: Investigation.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements
T. Miyamoto was supported by a grant-in-aid for Scientific Research in Japan and a grant from the Japan Agency for Medical Research and Development. This study was supported in part by the grand from the Japanese Orthopaedic Association (#2020-3) and the Japan Agency for Medical Research and Development2 (22dk0207053h0002).

63 References online

VitaminDWiki – Falls and Fractures category contains

Falls

Fracture

75+ Hip fracture items in VitaminDWiki

This list is automatcially updated

Items found: 77
Title Modified
Hip fractures 60% more likely in people who drank a lot of milk - Greger Feb 2024 14 Feb, 2024
Hip fracture risk was reduced by 4 times when Vitamin D was added to SERMS - Jan 2024 25 Jan, 2024
Risk of Hip fracture cut in half if Vitamin D was prescribed - Aug 2023 26 Aug, 2023
Hip fractures are predicted by 10 factors – low Vitamin D is the biggest – Aug 2023 11 Aug, 2023
Hip fractures greatly reduced by sunshine, vitamin D, and vitamin K – meta-analysis Sept 2012 23 Dec, 2022
Hip fractures requiring hospitalization cut in half by Vitamin K1 (100 mcg per day) – Sept 2022 17 Dec, 2022
More hip fractures if low vitamin D (2.1X in case-controlled studies) – meta-analysis May 2022 01 Jun, 2022
Less than 3 percent of hip fracture patients prescribed enough Vitamin D to make a difference (Malta in this case) – July 2021 26 Jul, 2021
Recommends vitamin D after hip fracture ( but not enough and no-bone building co-factors) – May 2021 21 May, 2021
A call to action: Vitamin D for hip fracture (50,000 weekly for 8 weeks, then bi-weekly – Holick July 2020 11 Dec, 2020
Risk factors for death after hip fracture surgery – 7 of the 8 are associated with low vitamin D – Aug 2020 09 Aug, 2020
Guideline following hip fracture – 50000 IU vitamin D daily for 7 days – Jan 2013 07 Jun, 2020
3.5 X higher risk of death 2 years after hip fracture surgery if low vitamin D – Jan 2020 07 Mar, 2020
Hip surgery with multiple doses of 50,000 IU of vitamin D weekly both before and after – RCT 2023 16 Jan, 2020
Hip fractures not prevented by Vitamin D (800 IU daily or large quarterly or annual doses) – meta-analysis – Dec 2019 21 Dec, 2019
Hip fractures reduced 16 to 33 percent by any amount of Vitamin D and Calcium – Oct 2019 05 Nov, 2019
Hip fracture rate increases in winter (subtropical Australia too) – Aug 2019 10 Aug, 2019
Warning - High intake of Vitamin B12 and B6 found to increase risk of hip fracture by 47 percent - 2019 19 May, 2019
Following hip fracture 2,000 IU of vitamin D daily improved quality of life – Feb 2019 23 Mar, 2019
Hip fractures worse if both high PTH and low Vitamin D – Jan 2019 01 Feb, 2019
Hip fracture recovery best with home exercise plus Vitamin D – RCT Dec 2018 12 Dec, 2018
Death of older hip fracture patients at least 1.6X more likely if low vitamin D – May 2018 19 Nov, 2018
After hip surgery Vitamin D levels dropped by 32 percent – Sept 2018 26 Sep, 2018
France wants to reduce hip fractures by drinking 10 glasses of milk daily – April 2018 01 Apr, 2018
Hip fractures rates have been increasing since 2012 – Feb 2018 12 Feb, 2018
Hip fracture 58 percent more likely if low vitamin D – meta-analysis March 2017 20 Dec, 2017
Hip fracture risk reduced 6 percent for each daily dairy serving (more if full-fat dairy) – Oct 2017 28 Oct, 2017
France is planning to reduce hip fractures by dairy vitamin D fortification (plan will not work) – 2017 17 Oct, 2017
Vitamin D and exercise after hip fracture surgery – far fewer deaths – July 2016 23 Aug, 2017
Most Hip fracture patients had low vitamin D and protein (needed for strong bones) – May 2017 18 May, 2017
1 in 3 died after hip fracture but only 1 in 14 if add Vitamin D and exercise – RCT April 2017 19 Apr, 2017
Vitamin D loading dose after hip fracture surgery was great – RCT Aug 2016 06 Apr, 2017
Following a stroke, a hip fracture is 6X more likely if low vitamin D – July 2001 04 Mar, 2017
Gene makes Hip Fracture 2.5X more likely unless get more vitamin D - Aug 2015 12 Nov, 2016
Partial (proximal) hip fracture strongly associated with low Vitamin D – Aug 2016 08 Nov, 2016
24 years ago hip fractures were reduced 24 percent in a large trial using Vitamin D and Ca – Oct 2016 22 Oct, 2016
Kaiser working to decrease hip and other fractures – Nov 2011 22 Oct, 2016
Hip Fracture 2.5 times more likely if poor Vitamin D Binding Protein gene - Aug 2015 25 Sep, 2016
Hip fracture risk increased 2.5X if problem with Vitamin D Receptor gene (GC) – March 2014 25 Sep, 2016
Hip fracture 50 percent more likely if low in both vitamin D and vitamin K1 – Dec 2015 26 Dec, 2015
Hip fracture 3X more likely if low vitamin D – Dec 2015 26 Dec, 2015
5X fewer hip fractures from Parkinson – Alzheimer – Stroke with enough sun – June 2011 17 Jul, 2015
7 X more likely to get hip fracture if have highest level of vitamin A – Jan 2003 31 May, 2015
Delirium 2.7 X more likely after hip fracture and low vitamin D – May 2015 27 May, 2015
Hip fractures reduced 30 percent with 800 IU of vitamin D – meta-analysis July 2012 16 Mar, 2015
Hip fracture outcome 5X more likely to be poor if low vitamin D – Jan 2015 16 Mar, 2015
Hip fracture rate varies with latitude (and thus vitamin D) in Sweden – Nov 2013 20 Feb, 2015
Hip Fracture – 5X more likely to have bad outcome if have less than 20 ng of vitamin D – Dec 2014 22 Jan, 2015
Hip replacement 2X more likely if hip bone surgery was during winter (low vitamin D) – Dec 2014 05 Dec, 2014
Fewer heart attacks, hip fractures and deaths if more skin cancer – Sept 2013 09 Nov, 2014
Hip fracture – not getting enough vitamin D – 2013 28 May, 2014
Hip fractures reduced 2X to 6X with just 10 minutes of sunlight daily – RCT 2003-2010 04 Dec, 2013
Men with hip fracture were 1.6X more likely to have low vitamin D – Aug 2013 09 Nov, 2013
5X increase in hip fracture rates in 65 years - 1998 27 Aug, 2013
Hip fracture reduced 38 percent with Calcium and just 400 IU of vitamin D – RCT Feb 2013 17 Aug, 2013
20 percent fewer male hip fractures if more Magnesium in the water – July 2013 11 Jul, 2013
Half as many hip fractures if take Calcium, hormones, and a tiny amount of vitamin D – July 2013 26 Jun, 2013
Male Hip fracture 1.65X more likely if low vitamin D – May 2013 01 Jun, 2013
Vitamin D needed to prevent and treat hip fracture, but only 20 percent got any – May 2013 31 May, 2013
Tiger got surgery for hip arthritis – not much sun in the cage – March 2013 23 Mar, 2013
Having less than 25 ng of vitamin D was associated with severe hip fracture – March 2011 30 Jan, 2013
Hip fracture 19 percent less likely with just 4 ng higher level of vitamin D – Dec 2012 20 Dec, 2012
Hip fracture and vitamin D – Health ABC study – May 2012 28 Sep, 2012
Hip surgery followed by 100000 IU then 1000 IU of vitamin D daily – June 2010 31 Aug, 2012
2X to 3X higher hip fracture rate in Asia in past 30 years – Aug 2012 18 Aug, 2012
Cataract surgery resulted in 30 percent fewer hip fractures – July 2012 02 Aug, 2012
Hip fractures in India - editorial with recommendation - Sept 2010 06 Jul, 2012
Women with hip fractures very low on vitamins D3 and K – Mar 2011 06 Jul, 2012
Upper body bones fractured along with hip when extremely low on vitamin D – Sept 2010 24 Jun, 2012
30 percent less chance of dying after hip fracture if take vitamin D and other drugs – 2011 24 Jun, 2012
75 percent of hip fractures associated with vitamin D deficiency - Jan 2011 24 Jun, 2012
Vitamin D 2nd most recommended way to prevent next hip fracture – Nov 2010 23 Jun, 2012
Hip fractures worse if very low on vitamin D – Mar 2011 23 Jun, 2012
Hip fracture patients had only 16 ng of vitamin D – May 2011 23 Jun, 2012
Declining hip fracture rates in the United States – unknown reason – June 2010 23 Jun, 2012
Vitamin K1 reduced hip fracture but Vitamin K2 did not – Aug 2011 23 Jun, 2012
Calcium and vitamin D supplements after hip fracture reduced death rate by 25% – Feb 2011 19 Nov, 2011

VitaminDWiki – Bone Mineral Density and Vitamin D - many studies

See also in VitaminDWiki

Hip fractures are predicted by 10 factors – low Vitamin D is the biggest – Aug 2023        
1464 visitors, last modified 11 Aug, 2023,
(Cached) Printer Friendly Follow this page for updates
Falls and Fractures 253
Page actions
Refresh Print
Edit