Predicting the probability of osteoporotic fracture risk in men versus women: do we need specific reference interventional thresholds for men?

Abstract

Background

Gender differences in the diagnosis and treatment of osteoporosis is a relatively common phenomenon, particularly amongst those patients under 80 years of age presenting with fragility fractures. The clinical implications of these findings are that strategies, which tend to focus on osteoporosis management in women, should also pay enough attention to osteoporosis in men. However, there have been questions whether there is a gender difference when setting intervention thresholds for osteoporosis management. This work was carried out aiming to determine the fracture probabilities calculated by FRAX at which therapeutic intervention in older men and women can be considered.

Results

Assessment of fracture risk probability thresholds in men revealed that for hip fractures, ROC was 0.754 (95% CI: 0.69–0.817). The sensitivity was 80.5% for threshold probabilities of 2.5%. For the major osteoporosis fracture, ROC was 0.828 (95% CI: 0.694–0.963). The sensitivity was 87.5% for threshold probabilities of 10%. Assessment of fracture risk probability threshold in women ROC was 0.760 (95% CI: 0.691–0.83). The sensitivity was 76.1% for threshold probabilities of 3%. For major osteoporosis fracture, ROC was 0.848 (95% CI: 0.784–0.912). The sensitivity was 87.3% for threshold probabilities of 15%.

Conclusion

Operational aspects of osteoporosis management should consider gender specific fracture thresholds. Interventional thresholds were found to be different in men compared to women. This helps to optimise fracture prevention in older men.

Background

With the increased ageing population and longer life expectancy, osteoporosis has become an important public health problem. As osteoporosis does not manifest until a fracture occurs, it has been well known as the silent epidemic [1]. Fragility fractures, which occur as a major consequence to osteoporosis, have been associated with increased mortality and negative impact on the patients’ quality of life as well as functional ability [2,3,4].

It is well known that when considering bone health, there is a gender difference, and that osteoporosis is a disease which predominantly affects women. It has been reported that women 50 years of age or older have a 4 times higher rate of osteoporosis and 2 times higher rate of osteopenia in comparison to men [5]. Earlier published studies revealed that one out of every 3 women and one out of every 5 men will experience fragility fractures at least once in their lifetime [6]. In the over 50 population, approximately, 75% of all hip fractures occur in females [7], whereas 25% of the hip fractures occur in men [8]. For both sexes, the peak of hip fractures number occurred at 75–79 years of age. In contrast, for all other fractures, the peak number occurred at 50–59 years and decreased with age [8]. In contrast, the incidence of vertebral fractures increases with age in both sexes. However, several studies reported that, up to the ages of 50 or 60 years, the prevalence of vertebral fractures in males is similar to, or even more than, that seen in females, but is to some extent greater in women than men thereafter [9, 10]. Over 55% of patients with hip fracture have evidence of a previous vertebral fracture [11]. It has also been shown that a 10% loss of bone mass in the vertebrae can double the risk of vertebral fractures, whereas a 10% loss of bone mass in the hip can result in a 2.5 times greater risk of hip fractures [12].

A series of meta-analyses carried out over the past years have identified clinical risk factors that could be implemented in the case finding approach, with or without BMD assessment. This paved the way for developing FRAX®, a tool which integrates the data derived from the individual subject’s clinical risk factors as well as BMD [13]. The use of FRAX (fixed or age-dependent thresholds) as the gateway to evaluation recognises people at high risk more efficiently than using BMD. However, given the heterogeneity of fracture risk and mortality in different nations, the intervention threshold (i.e. the fracture threshold) is also expected to differ. Furthermore, there have been questions whether there is a gender difference when setting intervention thresholds for osteoporosis management. This study was carried out aiming to determine the FRAX calculated fracture probabilities at which therapeutic intervention can be considered in men as well as women and whether gender specific reference interventional thresholds are required.

Methods

Study design

This was a multi-centre cross-sectional epidemiological study, which involved Egyptian postmenopausal women and men above the age of 40 years, seeking to determine the characteristics and interventional thresholds of hip fractures. The patients were retrieved from the national Fracture Liaison Service (FLS) data base during the period from February 2022 to February 2023.

Ethical consideration

This study is in agreement with the Declaration of Helsinki and it follows Tanta Faculty of Medicine ethical standard. Informed written consent, in accordance with the local ethical committee, was obtained from all patients. The institution’s ethics board approval number 33997/8/20. A code number was given for every patient file included in the work to ensure full privacy of all patients’ data.

Participants

Target population

Inclusion criteria: All patients in both sex presenting with hip fragility fractures or major osteoporosis fracture were assessed as part of the local FLS services.

Case definition

Egyptian patients, either male or postmenopausal female, above 40 years of age, who developed fragility fracture. Fragility fracture is defined as fracture after falling from standard height or less.

• Exclusion criteria: Patients presenting with pathological fractures or had history of violent trauma were excluded from the study.

Patients’ assessment

All patients completed a baseline questionnaire which included of (a) full history taking, including previous fracture, medication, alcohol use, smoking, or family history of fracture in a first-degree relative; (b) anthropometric assessment, including body weight and height with calculation of the individual patient’s BMI; (c) clinical evaluation of the musculoskeletal system; and (d) identifying fracture linked risk factors (specifically FRAX risk factors) [12] as well as any other fracture risks, including parental hip fracture, low trauma fracture in the last 2 years, rheumatoid arthritis, steroid therapy, consumption of more than 3 units of alcohol/day smoking, and chronic diseases such as chronic kidney or liver illnesses, cancer breast on hormone antagonist therapy, cancer prostate on androgen depletion therapy malabsorption syndrome, thyroid diseases, hypogonadism, and epilepsy (on anticonvulsant therapy). In addition, patients with history of joint replacement were assessed regarding the cause and site, of joint replacement.

Evaluation of bone mineral density

Assessment of the BMD was carried out using by dual X-ray absorptiometry (DXA) Hologic scanners at the non-dominant hip (total hip, femoral neck, and trochanter) as well as the lumbar spine (L1 through L4) with calculation of both T-scores and Z-scores.

To calculate the T-scores NHANES Caucasian-based Hologic reference values for females in the age range of 20–29 was utilised. This was carried out using a DXA machine auto analysis system. Then, the T- and Z-scores are calculated using a specific DXA software. Whenever needed, and to double check, a manual analysis of hip, spine, and forearm scans was executed. Before the densitometer would allow scans to be completed, the spine phantom must be calibrated daily. This is carried out by daily quality control (QC) check for the Hologic Anthropomorphic Spine Phantom (HASP).

Database registry

The national database registry was the hub where all the recorded data were collected and saved electronically [14]. This included 9 divisions: patient’s data and demographics, outcomes of the patients’ survey, results of the DXA scan (BMD, bone mineral content, T- and Z-scores), blood tests results, estimated risk for fracture, falls and sarcopenia, reports, and statistical parameters. The electronic data recording not only enables the collection of the patients’ information in their initial review, but also all their data in the follow-up visits. Its statistical analysis tool facilitates the evaluation of the services and the auditing process provided against the national guidelines for osteoporosis management as well as the golden clinical standards.

The Egyptian FRAX model

The Egyptian FRAX model was developed based on the incidence of hip fractures in Egypt. Calculation of the incidence of hip fracture per 100,000 was carried out by dividing the estimated number of fractures by the estimate population at risk at different sex and age group classified as 5 years age intervals from 40 till 75 years and above. Details of the process and how the incidence is calculated and the incidence results were given in full details in a previous study [15].

Development of intervention thresholds

The 10-year probabilities of hip fracture as well as the major osteoporotic fracture (MOF) were calculated without the inclusion of BMD using the Egyptian model of FRAX. Given the assumption that a prior fracture was considered to carry a sufficient risk to recommend treatment, the intervention threshold for men/women without a prior fracture was set at the age-specific 10-year probability of a hip equivalent to men/women with a previous fragility fracture, respectively, but without any other clinical risk factors. At all ages, the body mass index (BMI) was set to 25 kg/m².

To calculate the ROC (receiver operating curve), the data from the patients with previous fracture were set against data from 270 subject of matched sex and age subjects without history of previous fracture attending the osteoporosis clinic.

Statistical analysis

The SPSS computer program was used to tabulate and analyse the collected data.

Chi-square test for association was used to compare risk factors in men versus women. p value was always set at 0.05.

Results

Demographics

One hundred and fifty-four men and 229 women with previous history of osteoporotic fragility fracture were included in this work. Mean age of the men included in the study was 69.07 years ± 11.29, whereas mean age of the postmenopausal women was 65.4 years ± 12.4. The prevalence of hip fractures in men was 89.4% whereas in women it was 78.8% (p < 0.01).

BMD assessment

In general, bone mineral content was higher in men versus postmenopausal women presenting with hip fractures. Mean BMD in men was 0.72 (±0.18), whereas mean BMD in women was 0.69 (±0.15). BMD was normal in 15.4% of the men included in this study, whereas 53.8% had osteopenia and 30.8% had osteoporosis. In contrast, 18.3% of the postmenopausal women had normal BMD, 36.7% had osteopenia, and 45% had osteoporosis. However, in both genders, there was similar gradual loss in BMD with age; this was paralleled by increase in the incidence of hip fractures in both sexes.

Fracture risk factors

Assessment of fracture risk factors in men and postmenopausal women revealed increased prevalence of history of fracture and smoking amongst men vs women, whereas there was increased prevalence of steroids and rheumatoid arthritis in women (Table 1).

Table 1 Fracture risk factors in men and women

Absolute fracture risk

Assessment of fracture thresholds in men versus postmenopausal women (Table 2) revealed significantly higher values amongst postmenopausal women for both major osteoporosis fracture as well as hip fractures. This was applicable at all the age groups (Tables 3 and 4).

Table 2 Mean of the fracture thresholds in men versus postmenopausal women

Table 3 Distribution of the absolute 10-year probability major osteoporosis fracture risk in men versus postmenopausal women

Table 4 Distribution of the 10-year probability hip fracture risk in men versus postmenopausal women

Interventional threshold of FRAX in men

Assessment of fracture risk probability thresholds in men revealed that for the 10-year probability of hip fractures, ROC was 0.754 (95% CI: 0.69–0.817). The sensitivity was 80.5% for threshold probabilities in the range of 2.5%. For the 10-year probability of major osteoporosis fracture, ROC was 0.828 (95% CI: 0.694–0.963). The sensitivity was 87.5% for threshold probabilities of 10% (Fig. 1).

Fig. 1

Area under the curve of hip fracture probability in men: area: 0.754 (std. error: 0.032), 95% CI 0.69–0.817

Assessment of 10-year hip fracture risk probability threshold in women revealed that ROC was 0.760 (95% CI: 0.691–0.83). The sensitivity was 76.1% for threshold probabilities in the range of 2.95–3.05%. For the 10-year probability of major osteoporosis fracture, ROC was 0.848 (95% CI: 0.784–0.912). The sensitivity was 87.3% for threshold probabilities of 15% (Fig. 2).

Fig. 2

Area under the curve of hip fracture probability in women: area: 0.76 (std. error: 0.036), 95% CI 0.691–0.83

For the 10-year probability of major osteoporosis fracture in males, AUCROC was 0.828 (95% CI: 0.694–0.963) (Fig. 3). The sensitivity was 87.5% for threshold probabilities of 10%. For the 10-year probability of major osteoporosis fracture in females, AUCROC was 0.848 (95% CI: 0.784–0.912). The sensitivity was 87.3% for threshold probabilities of 15% (Fig. 2).

Fig. 3

ROC curve displaying MOP Egyptian FRAX in males and females (AUC male = 0.828 SE 0.828; AUC female = 0.848 SE 0.033)

Discussion

In general, osteoporosis has been considered a “women’s disease”. This has been attributed to the high prevalence of osteoporosis and the high-risk of sustaining a fragility fracture reported in postmenopausal women [14]. However, older men do suffer from osteoporotic fractures too with consequent negative impacts on health, physical and functional ability as well as independence. These post-fracture poor health outcomes were rated to be worse in males than females and that males are twice likely to die after fracture hips in comparison to females [16,17,18,19]. Furthermore, a smaller percentage of males at high risk of fractures receive treatment in comparison to high-risk women [20]. Based on these gender differences in osteoporosis and fracture risks as well as its impact on the patients’ management, queries have been raised whether osteoporosis management and particularly its interventional thresholds should be considered differently in men compared to women. This aim of this work was to determine the FRAX calculated fracture probabilities at which therapeutic intervention can be considered in both sexes.

The diagnostic threshold for osteoporosis has been identified by WHO as a bone mineral density (BMD) T-score of −2.5 or below the young adult mean. This definition applies to both men and women and signals the loss of bone mass with consequent high risk for developing a fragility fracture. However, earlier published research revealed that implementing such diagnostic threshold as treatment cut off point, in fact, impedes fracture prevention [21]. Therefore, there have been calls to abolish the use of BMD as an intervention threshold and adopt the age-dependant fracture risk probability instead [22]. Results of this work revealed that whilst high risk for fracture is signalled by the existence of a T-score BMD in the osteoporosis range, a significant percentage of high-risk men who sustained fragility fracture(s) had BMD in the normal or osteopenic, and therefore, unlikely to receive any form of osteoporosis therapy. These findings indicate the poor sensitivity of BMD as men tend to fracture at higher BMD. Consequently, it can be suggested that the fracture risk can be different in men compared to women. This was supported by the finding of significantly higher prevalence of osteoporotic hip fractures in men compared to women. This, in turn, is in favour of the notion that the diagnostic thresholds based on BMD alone are not optimal intervention thresholds.

Results of this work revealed that facture risk factors as well as incidence of hip fractures at specific ages differed in men versus women. Smoking and recent history of fragility fracture were the most prevalent osteoporosis risk factors in men. This highlights the importance of risk factors screening in men and the high likelihood of having secondary osteoporosis. This agrees with the outcomes of earlier studies which revealed that the diagnosis of osteoporosis in men is inclined more often towards a secondary osteoporosis than in women. Research studies carried out by Ebeling and Dy et al. [23, 24] reported that in contrast to females, where 20–40% of osteoporosis is secondary to extra-skeletal illnesses, this percentage goes up to 65% in men. The combination of osteoporosis and sarcopenia was reported in a previous study [25] on the same cohort of patients. Sarcopenia facilitates the reduction of body balance, attributed to the loss of muscular strength, which in turn paves the way for falling and consequently increases the risk of fractures. The importance of identifying these risk factors is based on the fact that they represent the reversible option of the fracture occurrence and treatment.

In contrast to the common concept which identified that fracture and intervention thresholds in women can also be implemented in men too [26], results of this work revealed that interventional thresholds are different in men compared to women. In comparison to women where the interventional thresholds were set at 15% for major osteoporosis fracture and 3% for hip fractures [15], this study identified interventional thresholds of 10% for major osteoporosis fracture and 2.5% for hip fracture in men. Though these figures are lower than that reported in the USA (major osteoporosis fracture risk probability of 20% and hip fracture probability of 3%) [20], it agrees with regional fracture probability rates. In Lebanon [27], the intervention threshold has been set at ≥10% for the 10-year probability of a major fracture for individuals up to age 70 years, whereas “a ‘biphasic’ economic probability threshold was proposed in Greece, using 10-year probabilities for hip and major osteoporotic fractures of 2.5 and 10%, respectively, under the age of 75 years” [28]. Similar findings were reported in other countries: where a 15% and a 7% risk of major osteoporotic fracture is recommended to permit a cost-effective access to therapy in Switzerland and UK, respectively [29, 30]. Interestingly, whilst it is logic to accept that the interventional thresholds identified by the USA are included in the American guidelines, a systematic review of interventional thresholds for osteoporosis management [21] revealed that in several countries (25 countries), there was no rational for using the thresholds of 20% and 3% other than the fact that this was the thresholds used by the US-National Osteoporosis Foundation. Our findings provide more evidence of the challenge in osteoporosis care in men. The implementation of FRAX in standard day to day practice requires the consideration of the appropriate sex-matched probability of developing a fracture at which to advise therapy.

Limitation to this study is the relatively small number of the patients included; however, the inclusion of all individuals at the national level and the agreement of the study outcomes with other studies carried out in the same geographic region suggest that our results will be broadly generalisable to clinical practice and that the study findings are likely to be relevant in many settings outside Egypt.

In conclusion, gender disparities in osteoporosis are mirrored by disparities in interventional thresholds for management. A significant proportion of men at high risk of sustaining a fragility fracture have BMD in the normal or osteopenic range but are unlikely to be treated. Therefore, initiation of osteoporosis management in older men at high risk for fragility fractures can be suboptimal. Using gender specific thresholds are vital to ensure optimum screening for osteoporosis and commencing timely appropriate management.