Prevalence of osteoporotic fracture risk factors and antiosteoporotic treatments in the Valencia region, Spain. The baseline characteristics of the ESOSVAL cohort
- First Online:
- Cite this article as:
- Sanfélix-Genovés, J., Sanfélix-Gimeno, G., Peiró, S. et al. Osteoporos Int (2013) 24: 1045. doi:10.1007/s00198-012-2018-6
- 407 Views
This study provides information on the prevalence of the most important risk factors for osteoporosis and osteoporotic fracture in a large sample of women and men from the Valencia region and also provides the FRAX 10-year major and hip fracture risks for this population, as well as data about the use of diagnostic tests and antiosteoporotic treatments.
The purpose of this study was to describe demographic characteristics, osteoporosis risk factors, the 10-year risk of osteoporotic fracture, and the use of densitometry and antiosteoporotic treatments in the Valencia region, Spain.
A cross-sectional study using the ESOSVAL cohort baseline data was conducted. We analyze the data from 5,310 women and 5,725 men aged 50 and over who attended to 272 collaborating primary care centers in 2009–2010. We collected the demographic, anthropometric, clinical, and pharmacy data from the electronic medical record.
The mean age of participants was 64.3 years old for women and 65.6 years old for men. The most frequent fracture risk factors were sedentary life (22.2 %) and previous fracture (15.8 %) in women and low calcium intake (21.4 %) and current smoker (20.9 %) in men. According to FRAX®, the 10-year risk of presenting a major fracture was 5.5 % for the women and 2.8 % for the men. The 10-year risk for hip fracture was 1.9 and 1.1 % for the women and the men, respectively; 23.8 % of the women and 5.2 % of the men had a densitometry test, 27.7 % of the women and 3.5 % of the men were taking calcium and/or vitamin D supplements, and 28.2 % of the women (22.0 % in the 50–64 age group) and 2.3 % of the men were taking antiosteoporotic drugs.
The prevalence of certain fracture risk factors not included in the FRAX tool (sedentary life, falls, low calcium intake) is high. In young women, their low risks estimated by FRAX contrast with the high figures for densitometry testing and treatment.
KeywordsCross-sectionalOsteoporosisOsteoporotic fractureRisk assessmentRisk factors
Osteoporosis is a highly prevalent systemic disease of the skeleton with reported figures in Spain for women aged 50 and over of around 30 % . The incidence in developed countries is rising with improved life expectancy and lifestyle changes such as a more sedentary life and changes in dietary habits, among others. Osteoporosis is a silent disease, yet it has a major clinical impact because of its association with an increased risk of fracture. Osteoporotic fractures constitute a serious health problem not only because of their severe consequences in terms of mortality, pain, and quality of life  but also because of their important social and economic repercussions .
In the 1990s, the scientific community gave a leading role to bone mineral density (BMD) in the development of osteoporotic fractures. The National Institutes of Health (NIH) defined osteoporosis as a disease characterized by a low bone density , and the World Health Organization (WHO) Study Group classified patients as normal, osteopenic, or osteoporotic according to the results of bone densitometry . This “densitometric” definition of osteoporosis probably contributed to raising awareness of the high prevalence of low bone density in the population but—despite the authors' advice to use those criteria exclusively for epidemiological purposes—may also have promoted the pharmacological treatment of low-risk patients, such as young postmenopausal women. In fact, some leading institutions still recommend starting pharmacological treatment with the single criterion of densitometric osteoporosis [6–9], despite the well-known low BMD predictive value for osteoporotic fractures .
In 2001, the NIH redefined the concept of osteoporosis to introduce the notion of bone strength and to incorporate other risk factors—in addition to the densitometric one—to determine bone fragility and predisposition to bone fracture . Today, it is widely accepted that bone strength and several clinical risk factors are strongly related to osteoporotic fractures, and in recent years, several risk assessment tools have been proposed to estimate the probability of osteoporotic fractures and to identify high-risk patients [11–14], FRAX®  being the most widespread of these instruments. Supported by the WHO , the International Osteoporosis Foundation , the NOF , and other institutions, FRAX® has been validated and calibrated for different countries showing in general good clinimetrical properties .
In Spain, there are currently several ongoing studies attempting, among other objectives, to validate FRAX® [17–20] (some of them with recently published preliminary results [21, 22]), including the ESOSVAL study , a large cohort of women and men recruited in the primary health-care centers in the Valencia region in Spain, with patient follow-up using routine electronic health information systems. The validation of risk assessment tools requires collecting a broad set of information about fracture risk factors in a sizeable and representative group of patients that should be followed up for several years. The baseline characteristics of these cohorts could be useful to understanding both the actual risks in the population and the current clinical management of those risks. This information may be valuable for defining practical policies on osteoporotic fracture risks. The aim of this study is to describe demographic characteristics, fracture risk factors, the 10-year risk of osteoporotic fracture, and the use of densitometry tests and antiosteoporotic treatments in the ESOSVAL cohort at its baseline.
Materials and methods
A cross-sectional study was conducted in 2009–2010 in the Valencia region (Spain) using the baseline data of the ESOSVAL risk study, a large cohort with planned follow-up to 2020. The ESOSVAL risk study protocol has been fully described elsewhere .
The sample was recruited among the population receiving care from the Valencia Health Agency (VHA). The VHA, as with other regional health services in Spain, operates an extensive hospital and primary health-care center network which covers about 97 % of the six million inhabitants of the region. All the care in this network is free of charge, except for a pharmaceutical copayment for nonretired people. The VHA has developed some institutional programs for chronic diseases, including the ESOSVAL program for osteoporosis. These programs incorporate several components such as professional training, improved information systems with special attention to electronic medical records, quality improvement activities, and clinical and epidemiological research [23, 24].
Women and men aged 50 and over attended to 272 primary health-care centers from the VHA for any health problem between November 2009 and September 2010. Subjects were recruited opportunistically by 329 general practitioners and 279 primary care nurses participating in the ESOSVAL training program and following precise study criteria (patients should be recruited at different times and days of the week and in predefined age and sex strata, attempting to get the age distribution as close as possible to the distribution of the region's population). Exclusion criteria included nonresidents of the region, individuals with cognitive impairment which could make collection of the study variables difficult, people receiving their usual care through private insurance companies, people physically unable to attend their primary health-care center, and people of Asian or African descent. The criteria for clinicians and patient selection have been fully described previously .
Data sources and study development
The main source of data was the VHA electronic medical record, the so-called ABUCASIS system, which among other items includes demographic and clinical data and information on prescriptions. The electronic medical record ABUCASIS was modified in the context of the ESOSVAL program to include a specific osteoporotic risk sheet. This sheet facilitates the registration of fracture risk factors, patient monitoring, and decision-making about the need for complementary tests or pharmacological treatment (case-finding strategy). If more than one sheet was available in this period, we selected the most recent one. The participating doctors and nurses were trained through classroom activities, online courses, and practical exercises to standardize definitions and improve their knowledge of the management of osteoporosis, including the proper completion of the electronic medical record and the specific osteoporotic risk sheet .
The variables used in the study first include the patient's sociodemographic and clinical characteristics such as age (stratified into 50–64, 65–74, and 75 and over); gender; educational level (no studies, primary studies, and secondary/university studies); body mass index (BMI), categorized into BMI <20 or BMI ≥30; comorbidities registered in the medical record which include diabetes, chronic obstructive pulmonary disease, coronary heart disease, cerebrovascular disease, and other chronic conditions (renal disease, endocrine diseases, rheumatoid arthritis, chronic liver disease, malabsorption syndrome, prolonged immobility, and organ transplantation); and the diagnosis of osteoporosis registered in the medical record. Second, we included osteoporotic and fracture risk factors such as history of hip fracture in parents or siblings, personal history of previous osteoporotic fracture in any location, regular alcohol intake (consumption >17 U/week for women and >28 U/week for men), current smoker (any quantity of tobacco), sedentarism (if a person is sitting or standing for most of the day or have done less than two episodes of physical activity of more than 20 min each in the 2 weeks prior to the measurement) , number of falls in the previous year, calcium intake <500 mg/day estimated using the food composition chart of the Spanish Ministry of Health , use of glucocorticoids (≥5 mg per day of prednisone or equivalent for at least 3 months in the previous year), menopause age, untreated hypogonadism in men or women, and other osteopenic diseases included in the osteoporosis risk sheet (type I diabetes, rheumatoid arthritis, untreated long-standing hyperthyroidism, chronic malnutrition or malabsorption, chronic obstructive pulmonary disease, renal disease, prolonged immobility and organ transplantation, and chronic liver disease). The participating physicians were trained in the management of those variables and were given a manual of procedures containing all the pertinent definitions.
Using the FRAX® tool calibrated for Spain (www.shef.ac.uk/FRAX/index.htm), we estimated the 10-year risk of major osteoporotic fracture (clinical spine, hip, forearm, or humerus fracture) and the 10-year risk of hip fracture for each patient . Data in the FRAX® web were introduced by the research team, and calculations were based on gender, age, BMI, personal history of previous fracture, family history of fracture, current smoking, glucocorticoid use, rheumatoid arthritis, other osteopenic diseases, alcohol intake, and, if available, BMD measurement. In accordance with the FRAX® recommendations, missing values were considered as normal. Although in Spain there are no cutoff points to define populations at high or low risk, we tentatively use the criteria of the Scientific Advisory Council of Osteoporosis in Canada  to classify the FRAX® scores as low risk (10-year risk fracture <10 %), intermediate risk (10–19 %), and high risk (≥20 %) of major osteoporotic fracture and low risk (<3 %) or high risk (≥3 %) of hip fracture.
We recorded densitometric osteoporosis of the lumbar spine established through DXA (T-score L2–L4) and/or DXA in the total hip or the femur neck (T-score) and classified as normal, osteopenia, or osteoporosis using the WHO criteria . We collected densitometry results over 24 months prior to the date of enrolment for each patient up to the date of recruitment of the last patient. If more than one exploration was available, we selected the closest to the end of the recruitment period. If the selected exploration had more than one value (e.g., T-scores from both lumbar spine and femur), the lowest value was used to classify the patient. We also recorded the use of calcium and vitamin D supplements and treatments for osteoporosis including bisphosphonates (alendronate, risendronate, ibandronate), raloxifene, strontium ranelate, parathyroid hormone (1-34 and 1-84), and calcitonin. Patients were considered as treated if they had taken any antiosteoporotic treatment up to 3 months before the time of their inclusion in the ESOSVAL cohort. Information on supplements and antiresorptive treatments was collected from the GAIA system, the electronic prescription system associated with the electronic medical record ABUCASIS, which covers practically all prescriptions reimbursed by the Spanish National Health System (sNHS) in the Valencia region. Figures for antiosteoporotic drugs and calcium and vitamin D supplements do not include over-the-counter treatments or treatments prescribed by private doctors.
The ESOSVAL risk study is a naturalistic, observational study undertaken as part of routine clinical practice, with no intervention (apart from training of participating clinicians) with the patients included in the study. No additional diagnostic tests, visits, evaluations, or treatments were provided apart from what the attending physician deemed appropriate. All patients included in the study signed the informed consent form granting the researchers access to information contained in their medical record for the study purposes. All information relative to the patients was handled according to Spanish laws on confidentiality and patients' rights. The ESOSVAL risk study was reviewed and approved by the Committee for Ethics and Clinical Trials of the Centre for Public Health Research and the Public Health General Directorate (decision March 27, 2009).
The estimated prevalence of risk factors, the FRAX® 10-year fracture risks, densitometric results, and antiosteoporotic treatments are shown as percentages by gender and age groups (50–64, 65–74, and ≥75 years old), with the corresponding 95 % confidence intervals calculated using the binomial approach. To provide estimates for the whole population, totals were weighted according to the age distribution of the women and men over 50 in the Valencia region . Appendix shows the distribution of the 50-and-over population in the Valencia region and in the ESOVAL cohort and the weights used for weighting totals. As a rule, cases with missing data in one variable were eliminated in the analyses using that variable. The FRAX risk estimation is the only exception to this rule because the FRAX tool considers the variables with no information as “normal”. All analyses were performed using the STATA 11.0 (StataCorp, College Station, TX) statistical software.
Baseline characteristics of the ESOSVAL cohort: educational level and morbidity (in percent and 95 %CI)
n = 3,043
n = 1,249
n = 838
n = 5,310
n = 2,983
n = 1,555
n = 1,187
n = 5,725
Obesity (BMI ≥30)
Chronic obstructive pulmonary disease
Coronary heart disease
Other chronic conditions
Baseline characteristics of the ESOSVAL cohort: osteoporotic and fracture risk factors (in percent and 95 %CI)
n = 3,043
n = 1,249
n = 838
n = 5,310
n = 2,983
n = 1,555
n = 1,187
n = 5,725
Family history of hip fracture
Personal history of previous osteoporotic fracture
Body mass index <20
Regular alcohol intake (>17 U/week for women, >28 U/week for men)
Current smoker (any quantity of tobacco)
Falls (≥1 in the last year)
Calcium intake <500 mg
Glucocorticoid use (prednisolone equivalent >5 mg/day at least 3 months in previous year)
Menopause before 41 years old
Other osteopenic diseases
Baseline characteristics of the ESOSVAL cohort: risk probability of osteoporotic fracture (FRAX tool)
10-year risk of major fracture
10-year risk of hip fracture
Baseline characteristics of the ESOSVAL cohort: densitometric testing, bone mineral density, and antiosteoporotic treatments
Densitometric test in ±24 months from the recruitment date
Bone mineral density (T-score)b
Less than or equal to −2.5
−1 to −2.5
Calcium and/or vitamin D supplements
Antiosteoporotic treatment (any drug)
In contrast with the considerable efforts made to identify risk factors for osteoporosis and osteoporotic fracture, there is a lack of studies assessing their population prevalence, a key issue to assessing their true impact on the incidence of fractures. This study uses a large cohort of women and men to describe the population prevalence of numerous risk factors for osteoporotic fracture stratified by age and sex. Our study shows a high prevalence of falls in the previous year, a sedentary life, a personal and family history of fracture, a low calcium intake, being a current smoker, and osteopenic diseases. Additionally, it shows that the prevalence of these factors varies according to age and sex strata. Despite some discrepancies due to differences in population age profile and the different ways of defining some risk factors, the totals in our study are consistent with previous studies for women in Spain [17, 30, 31], but we have not found any relevant Spanish studies for men or any with information stratified by age groups. Other Spanish studies have described the prevalence of some risk factors, but comparisons are limited because those studies use specific samples (recruited during the performance of densitometries, in rheumatology services after hip fracture, or excluding nontreated people) [32–36].
From an epidemiological point of view, the importance of the following three risk factors should be emphasized: a lack of physical activity, falls, and a low intake of calcium. Physical activity is one of the most important factors conditioning bone mass. Several studies have confirmed that association in different populations and moreover physical activity has been identified as an independent predictor of the incidence of osteoporotic fracture [14, 37]. Sedentary lifestyles are especially high among women, although their prevalence depends heavily on the definition used. In this study, it was about 25 %, but in other Spanish studies, it reached more than 50 % in older women . Falls are also an important independent predictor of the incidence of osteoporotic fracture  and have recently been included in some tools such as the Garvan Institute Assessment Tool to estimate the risk of osteoporotic fracture . As in other studies [17, 31, 38, 39], we found a high prevalence of a patient's self-reported falls, especially in the oldest women (33 %). The clinically significant threshold of a low calcium intake is not well established, but several studies have found a consistent association between a low calcium intake and osteoporosis and osteoporotic fracture , and recently it has been suggested that intakes of less than 500 mg might increase the circulating parathyroid hormone and worsen bone health . We did not find a lower prevalence of low calcium intake among older people as in other studies [37, 39], but a low calcium intake prevalence in men which is almost double that of women. From a clinical point of view, although the strength of association between these factors and the risk of fracture may be less than that shown by other risk factors such as BMD, its importance is not negligible (e.g., low physical activity increases by between 1.64 and 2.08 times the risk of osteoporotic fracture [14, 37]). All these three factors are highly prevalent in the Spanish population and should be taken into account for clinical decisions, despite the fact that they are not included in the most widely used tools for determining the risk of osteoporotic fractures as in FRAX® .
Our results describe the 10-year risk of a major fracture and hip fracture for women and men using the FRAX® tool for the Spanish population. The risk of a major fracture among women was lower than that found in the population-based Canadian Multicentre Osteoporosis Study  or in the Manitoba cohort [41, 42], but the risk of hip fracture was analogous. Another large population study in France found results that were close to our study for hip and major fracture risks . The ECOSAP study in Spain found higher FRAX® risk scores in women (8.8 for major fracture and 3.7 for hip fracture) , but their cohort was older than the ESOSVAL cohort (ECOSAP mean age at baseline was 72.3 years vs. 64.3 years in the ESOSVAL cohort).
Our study identifies 28.8 % of women with a 10-year risk of hip fracture ≥3 %, but only 6.4 % had a 10-year risk of a major fracture >20 %. While some guidelines use one, the other, or both of these values as a reference for defining high-risk patients or as thresholds for treatment [9–11, 28], our results show no equivalence and strongly suggest that a threshold of 20 % for major fracture excludes most of the population at high risk. This inconsistency between reference risk thresholds for hip and major fracture has also been found in other studies [41, 44]. In Spain, FRAX® reference values for defining high-risk populations are not yet available and some studies use recommended thresholds from other countries. The results of our study suggest that a major fracture risk threshold of 20 % is inadequate in the Spanish setting and should not be used without further study.
One in 4 of the women and 1 in 20 of the men had received one densitometric test in the 30 months around the date of recruitment; 3 in 10 of the women were treated with antiosteoporotic drugs and a similar number received calcium and/or vitamin D supplements alone or in combination with antiosteoporotic treatment. Strikingly, one in four of the women in the 50–65 years group had received a densitometric test within a maximum of a 30-month period and one in five had been treated with antiosteoporotics despite the low risk of fracture in this age stratum (less than 1 % of the women exceed the 10 % 10-year risk of major fracture or the 3 % 10-year risk of hip fracture). The quantification of overuse and underuse would require the individual assessment of each patient with predefined criteria; however, the contrast between the high levels of treatment and the low prevalence of risk factors in young adults suggests a population overuse of densitometry testing and antiosteoporotic treatment in young low-risk women, which would have a notable impact on health spending because this age group represents more than half of women over 50 years old. Furthermore, higher percentages of treated low-risk women have also been observed in other studies [32, 45]. The results of this study also suggest an underuse of treatment in women over 75 (85 % have a FRAX hip fracture risk higher than 3 %, while only 34 % are treated).
Several potential limitations of this study should be addressed. First, the patients in the ESOSVAL cohort were recruited opportunistically from among patients who attended the primary health-care centers for any reason. The population in these samples is usually different from the general population in two opposing directions: people who attend medical centers tend to be sicker than the average population and, on the other hand, people with disabilities or who are very sick (such as very old persons) who cannot attend health-care centers are not represented. In practice and regarding the general population, this limitation will probably have resulted in some overestimation of risk in younger patients and some underestimation of risk in the elderly. Second, while the FRAX tool is not designed to evaluate fracture risk in treated people, we used it in treated and untreated patients. Although initially a certain underestimation of risk in treated patients might be expected, the high rates of treatment in very low-risk patients (in which no significant changes are expected with treatment) suggest a limited importance of this bias. Third, as expected in a study with the characteristics of ESOSVAL, developed under routine clinical practice conditions, several variables have a large percentage of missing data. The effect of missing data is difficult to estimate and may vary depending on the risk factors, but in most cases, we can expect it to correspond to normal values (e.g., a person without a record of glucocorticoid therapy is not likely to take it). This bias would result in an overestimation of the prevalence of the risk factors with a significant percentage of missing data. Fourth, the data sources used do not permit the identification of the over-the-counter purchase calcium and vitamin D supplements, treatments dispensed in pharmacies located outside the region, or test and treatments prescribed by private doctors, implying some underestimation of the prevalence of these variables. We have no information about the volume of densitometries, antiosteoporotic treatments, or supplements dispensed outside public coverage (or outside the region) for people who regularly attend their primary care doctor in the public system, but we do not expect them to be very important, except perhaps in the case of calcium and vitamin D supplements. Finally, the ESOSVAL cohort represents the population of a Spanish region (with 11 % of the overall Spanish population) and its results cannot necessarily be extrapolated to other regions or to the entire Spanish population, especially because some studies have shown a considerable variability in hip fracture rates and antiresorptive treatment between different territories in the same country.
In conclusion, this study provides information on the prevalence of the most important risk factors for osteoporosis and osteoporotic fracture in a large sample of women and men in the Valencia region and also provides the FRAX 10-year major and hip fracture risks for this population, as well as data about the use of diagnostic tests and antiosteoporotic treatments. Our study suggests that some risk thresholds used in other countries or recommended by international guidelines could be inadequate in the Spanish setting. Additionally, beyond their contribution to enhancing the knowledge of osteoporotic fracture risk factors in a specific Mediterranean region, the results of our study suggest the quality of care problems in the management of these risks by an overuse of tests and treatments in low-risk groups.
The ESOSVAL research program is funded by the Instituto de Salud Carlos III (grants PS09/02500 and PI11/00238) as part of the Spanish Ministry of Health and Social Policies; the 2009, 2010, and 2011 collaboration agreements between MSD Spain, the Fundación para el Fomento de la Investigación Sanitaria y Biomédica (FISABIO), and the Valencia Department of Health to conduct training and research into musculoskeletal disorders and osteoporosis; and the 2010 collaboration agreement between AMGEN and the Centro Superior de Investigación en Salud Pública (CSISP) of the Valencia Ministry of Health. In 2011, JSG was partially funded by the Instituto de Investigación Sanitaria INCLIVA through grant INT10/122 from the Spanish National Health System program for stabilization of research and intensification of research activity. The organizations funding the research project do not necessarily share the contents of this manuscript.
We are grateful to all the doctors and nurses in the Valencia Health Agency primary health-care centers participating in the ESOSVAL study for their collaboration and to the Valencia Department of Health for its enthusiastic and continuous support of the ESOSVAL research projects.
Conflicts of interest
None of the sponsors played any role in the design of the ESOSVAL studies; the collection, analysis or interpretation of data; the writing of the manuscripts; or in the decision to submit it for publication. JSG and SP have received research grants from various pharmaceutical companies and from fees for participating in scientific meetings sponsored by pharmaceutical companies.