Elsevier

Bone

Volume 114, September 2018, Pages 116-124
Bone

Full Length Article
Fracture incidence and secular trends between 1989 and 2013 in a population based cohort: The Rotterdam Study

https://doi.org/10.1016/j.bone.2018.06.004Get rights and content

Highlights

  • Age is the most important determinant of fracture with those of the hip, wrist and proximal humerus having the highest rates

  • Bone mineral density is still predictive of fractures for up to 20 years

  • No secular trend differences in fracture incidence rates were observed between the 1989-2001 and 2001-2013 periods

Abstract

Fracture incidence needs to be evaluated over time to assess the impact of the enlarging population burden of fractures (due to increase in lifespan) and the efficacy of fracture prevention strategies. Therefore, we aimed to evaluate the association of femoral neck bone mineral density (FN-BMD) measured using dual-energy X-ray absorptiometry (DXA) at baseline with fracture risk over a long follow-up time period. Incident non-vertebral fractures were assessed in 14,613 individuals participating in the Rotterdam Study with up to 20 years of follow-up. During a mean follow-up of 10.7 ± 6.2 years, 2971 (20.3%) participants had at least one incident non-vertebral fracture. The risk for any non-vertebral fracture was 1.37 (95% Confidence Interval (CI): 1.25–1.49) and 1.42 (95%CI: 1.35–1.50) for men and women, respectively. The majority (79% in men and 75% in women) of all fractures occurred among participants a normal or osteopenic T-score. The incidence rates per 1000 person-years for the most common fractures were 5.3 [95%CI: 5.0–5.7] for hip, 4.9 [95%CI: 4.6–5.3] for wrist and 2.3 [95%CI: 2.0–2.5] for humerus. To examine the predictive ability of BMD through follow-up time we determined fracture hazard ratios (HR) per standard deviation decrease in femoral neck BMD across five year bins. No differences were observed, with a HR of 2.5 (95%CI: 2.0–3.1) after the first 5 years, and of 1.9 (95%CI: 1.1–3.3) after 20 years. To assess secular trends in fracture incidence at all skeletal sites we compared participants at an age of 70–80 years across two time periods: 1989–2001 (n = 2481, 60% women) and 2001–2013 (n = 2936, 58% women) and found no statistically significant difference (p < 0.05) between fracture incidence rates (i.e., incidence of non-vertebral fractures of 26.4 per 1000 PY [95%CI: 24.4–28.5]) between 1989 and 2001, and of 25.4 per 1000 PY [95%CI: 23.0–28.0] between 2001 and 2013. In conclusion, BMD is still predictive of future fracture over a long period of time. While no secular changes in fractures rates seem to be observed after a decade, the majority of fractures still occur above the osteoporosis threshold, emphasizing the need to improve the screening of osteopenic patients.

Introduction

Fractures of the hip, wrist and pelvis are among the most common osteoporotic fractures [1] affecting hundreds of millions of people worldwide. In the European Union, the annual costs of all osteoporotic fractures has been estimated at €37 billion [2] in 2010, of which 54% of the costs are attributed to hip fractures [2]. Due to the aging of industrialized societies, the incidence of osteoporosis and fragility fractures is expected to increase in the years to come [3, 4].

However, not all studies show a clear increase of fractures rates. For instance, previous studies have yielded conflicting perspectives indicating that the incidence of hip fracture has either increased, plateaued, or even decreased in the last decades [[5], [6], [7], [8], [9], [10]]. Such discrepancies may be explained by multiple factors, including: secular periods in which the fractures occur, changes in clinical practice and drug prescription or compliance, distribution of age and demographics within age and sex strata, migrations and/or geographical origin of the report [7]. For example, studies performed in the USA have reported a decline in the incidence of hip fractures between 1980 and 2000 [8, 9]. In contrast, in the Netherlands, an initial linear increase in hip fracture incidence during the period between 1972 and 1987 [10] is now proposed to have plateaued during the following decades [11]. Moreover, recent information on incidence rates and trends for other sites of non-vertebral fractures is currently lacking despite that non-hip fractures are also associated with higher disability [12] and mortality [13] rates. Depiction of robustly-assessed overall and site-specific fracture trends is important to estimate the burden of osteoporosis and to establish proper and cost-effective prevention strategies.

Part of delineating cost-effective prevention strategies involves determining the predictive ability of fracture risk factors in time. Although fractures have a multifactorial background, apart from age, low BMI and falls, a low bone mineral density (BMD) remains one of the most important risk factors. Studies have shown that the reduction of femoral neck (FN-) BMD is essentially linear over time, comprising ~1–2% of baseline BMD per year [14]. It is well established that BMD can predict fracture risk over a period of 5–10 years [15, 16], but little is known [17, 18] about the predictive value of BMD over longer periods.

Altogether, our aims were to 1) evaluate the association between BMD at baseline (both continuously and using clinical cutoffs) and fracture risk over a long time period, 2) estimate the incidence of non-vertebral fractures (overall and site-specific) during 23 years of follow-up and 3) evaluate whether incidence rates have changed during two subsequent secular follow-up periods. All analyses were performed for the total group and stratified by sex, considering the well-established sexual differences in fracture risk.

Section snippets

Study population

Our study included participants from the Rotterdam Study, an ongoing population-based prospective cohort comprising 14,926 Dutch individuals aged 45 years and older examined across three population sets [19]. At its start in 1990, a total of 7983 participants aged 55 years and older were included in the initial study wave (RSI). The cohort was expanded in 2000 with 3011 participants (RSII) aged 55 years and older; and in 2006 with 3932 participants (RSIII) aged 45 years and older, or who had

Statistical analyses

All analyses were performed for the total group of participants and stratified by sex, considering the well-established sex differences in fracture risk [[22], [23], [24]]. First, Cox proportional hazard models adjusted for age at baseline and cohort were used to estimate the hazard ratio (HR) of first fracture associated with 1 SD decrease in FN-BMD across 1) all non-vertebral fractures, 2) for specific types of fractures and 3) all non-vertebral fractures in groups of subjects classified by

Characteristics of the study population

Fig. 1 is a flow diagram describing the selection of study participants. At baseline, the mean (±SD) age of the participants was 64.7 (±9.4) years for men and 66.5 (±10.9) years for women. The prevalence of osteoporosis (BMD T-score <−2.5) was 10.8% in women and 6.6% in men. The prevalence increased exponentially with age for all types of fractures but this relation is less prominent for wrist fractures (Fig. 2). Above the age of 85 years, 40.9% of the women and 31.7% of the men had

Discussion

In this population-based prospective cohort, a single FN-BMD measurement at baseline remains a strong predictor of incident non-vertebral fragility fractures over a period of 20 years. The overall incidence rate of suffering non-vertebral fractures was 21.1 per 1000 PY [95%CI: 20.3–21.9] with a higher incidence rate in women than in men. The most frequent non-vertebral fractures in elderly men and women continues to be fractures of the hip, wrist, proximal humerus and hand, with most events

Conclusion

In conclusion, BMD remains a strong predictor of hip and non-vertebral fractures over 20 years in both men and women. The majority of fractures continue to occur above the osteoporosis threshold emphasizing the need to improve the screening of osteopenic patients. Most importantly, we established there are no difference in fracture trends between the periods of 1989–2001 and 2001–2013, welcoming active actions seeking to improve the diagnoses, treatment and prevention strategies to this costly

Acknowledgements and statement of authors' contributions to manuscript

We gratefully acknowledge the contribution of the participants of the Rotterdam Study, research assistants, the general practitioners, hospitals and pharmacies in Rotterdam.

Funding sources

The Rotterdam Study is funded by Erasmus Medical Center and Erasmus University Rotterdam, Netherlands Organization for the Health Research and Development (ZonMw), the Research Institute for Diseases in the Elderly (014-93-015; RIDE2, RIDE), the Ministry of Education, Culture and Science, the Ministry for Health, Welfare and Sports, the European Union Seventh Framework Programme (FP7/2007–2013) under grant agreement No. 601055, VPH-DARE@IT, and the Municipality of Rotterdam.

References (46)

  • A.J. Søgaard et al.

    Continued decline in hip fracture incidence in Norway: a NOREPOS study

    Osteoporos. Int.

    (2016)
  • R. Hiebert et al.

    Temporal and geographic variation in hip fracture rates for people aged 65 or older, New York State, 1985–1996

    Am. J. Orthop. (Belle Mead NJ)

    (2005)
  • D. Zingmond et al.

    Increasing hip fracture incidence in California Hispanics, 1983 to 2000

    Osteoporos. Int.

    (2004)
  • F.T. Boereboom et al.

    The incidence of hip fractures in The Netherlands

    Neth. J. Med.

    (1991)
  • W.G. Goettsch et al.

    Developments of the incidence of osteoporosis in The Netherlands: a PHARMO study

    Pharmacoepidemiol. Drug Saf.

    (2007)
  • H.A. Fink et al.

    Disability after clinical fracture in postmenopausal women with low bone density: the fracture intervention trial (FIT)

    Osteoporos. Int.

    (2003)
  • T. Tran et al.

    Population-wide impact of non-hip non-vertebral fractures on mortality

    J. Bone Miner. Res.

    (2017)
  • L.J. Melton et al.

    Determinants of bone loss from the femoral neck in women of different ages

    J. Bone Miner. Res.

    (2000)
  • L.J. Melton et al.

    Relative contributions of bone density, bone turnover, and clinical risk factors to long-term fracture prediction

    J. Bone Miner. Res.

    (2003)
  • D.M. Black et al.

    The ability of a single BMD and fracture history assessment to predict fracture over 25 years in postmenopausal women: the study of osteoporotic fractures

    J. Bone Miner. Res.

    (2017)
  • E. Sornay-Rendu et al.

    How to predict fragility fracture beyond 10 years? The OFELY study

    J. Clin. Endocrinol. Metab.

    (2014)
  • M.A. Ikram et al.

    The Rotterdam Study: 2018 update on objectives, design and main results

    Eur. J. Epidemiol.

    (2017)
  • A.C. Looker et al.

    Updated data on proximal femur bone mineral levels of US adults

    Osteoporos. Int.

    (1998)
  • Cited by (69)

    View all citing articles on Scopus
    View full text