Epidemiology of secondary fractures






Bernard CORTET,MD, PhD
Department of Rheumatology and EA 4490
University-Hospital of Lille
Lille, FRANCE

Epidemiology of secondary fractures


by B. Cortet, France



The term fracture cascade has been coined to describe the fact that the occurrence of a first fracture dramatically increases the risk of subsequent fractures. Although the etiopathogenesis of this phenomenon is not fully understood, it seems that the mechanisms are not identical for the vertebral and nonvertebral fracture cascades. For the vertebral fracture cascade, the main risk factors are low bone mineral density and impaired bone quality (which is also true for the nonvertebral fracture cascade); however, several non-bone parameters also characterize the vertebral fracture cascade: changes in local properties, alterations in bone macroarchitecture, spinal curvature and spinal loading abnormalities, and impaired intervertebral disc integrity. A past history of at least one vertebral fracture leads to a 4-fold increased risk of vertebral fracture. Moreover, it has been shown that the higher the number of prevalent vertebral fractures, the higher the risk of new vertebral fractures will be. In addition, a prevalent vertebral fracture also increases the risk of subsequent nonvertebral fracture. The risk is highest in the first year following a vertebral fracture and greater in men than in women. Similarly, for the nonvertebral fracture cascade, the occurrence of a nonvertebral fracture also increases both the risk of subsequent vertebral fractures—to varying degrees, according to the site of the prevalent nonvertebral fracture—and the risk of subsequent nonvertebral fractures. All of these findings are important since some fractures (both prevalent and incident) are associated with an excess in mortality (particularly those of the hip and vertebrae).

Medicographia. 2014;36:143-149 (see French abstract on page 149)



Numerous risk factors have been associated with the occurrence of fragility fractures. Among them, a prevalent fragility fracture is a major risk factor for future fractures. Whatever the site of the prevalent fracture, the risk of subsequent fracture is increased. The subsequent fracture can occur either at the same site or another site; however, the level of risk varies depending on the site of the prevalent fracture and whether it is a vertebral or nonvertebral fracture. A simple example can illustrate the weight of a prevalent fracture when assessing the risk of fracture and the FRAX® tool (Fracture Risk Assessment tool)1 is a simple tool with a clear educational value in this context. The 10-year major osteoporotic fracture risk of a 74-year-old woman (weight, 60 kg; height, 160 cm) whose only risk factor for fracture is a smoking history (10 packs per year) is 12%. However, if this patient had a past history of fragility fracture, her 10-year risk for major osteoporotic fracture would almost double (22%).

The term fracture cascade has been coined to illustrate the fact that once a first fracture has occurred, the risk of a second fracture increases greatly, and that once a second fracture occurs the risk of a third fracture increases even further. In this article we will review the burden of the fracture cascade according to the site of prevalent fracture.

Prevalent vertebral fracture and risk of subsequent vertebral fracture

Vertebral fractures are among the major fragility fractures and occur in 16% of postmenopausal women.2 Vertebral fractures are emblematic of osteoporosis. However, their frequency is difficult to determine since only 1 in 4 comes to clinical attention for various reasons such as the low diagnostic performance of radiography or the absence of recognized classic symptoms of vertebral fractures.3

A prevalent vertebral fracture dramatically increases the risk of future fracture, particularly in the spine, and the concept of the fracture cascade seems to be particularly relevant for vertebral fractures.

Etiopathogenesis
Although the etiopathogenesis of the fracture cascade is complex and not fully understood, there are several hypotheses.4 First, let’s consider bone properties. In clinical practice, bone properties are assessed by measuring bone mineral density (BMD) by dual-energy x-ray absorptiometry (DXA). However, although BMD measured by DXA is lower for patients with vertebral fractures compared with controls, the mean difference in terms of BMD is modest.4 Comparing measurements of subregional BMD in patients with and without vertebral fractures would be more relevant; yet, this approach is not used in clinical practice.4 These data suggest that other factors may explain the vertebral fracture cascade, in particular bone quality, spine properties, and neurophysiological properties. As previously mentioned, measurement of bone quality is a relevant factor when evaluating bone strength.5 However, in studies performed in patients with vertebral fractures, bone quality—and particularly bone microarchitecture— is usually measured using transiliac bone biopsies. Therefore, the conclusions of these studies may not be relevant to the concept of the fracture cascade. That’s because the skeleton is quite heterogeneous and the measurement of bone microarchitecture at the ilium does not necessarily reflect the situation at the vertebral level. In contrast, spine properties seem to play a major role. The first reason for this is that the fracture cascade is more marked in the spine than in other bone sites. The vertebral fracture cascade may be the consequence of changes in local properties and several studies have indicated that women with vertebral fractures have smaller vertebrae than control patients.6 It has also been shown that women with vertebral fractures have smaller lever arms (the distance between the erector spinae and vertebral centroids). These abnormalities (reduced vertebral cross-sectional areas and shorter lever arms) lead to increased mechanical loading. Therefore, in order to maintain the moment of equilibrium, the muscle force must increase. This is not a problem in normal situations4; however, in case of bone fragility, the vertebral body is not able to sustain the resultant increase in compressive and shear load and this leads to vertebral failure. Changes in load distribution may also be involved. Indeed, these changes are influenced by intervertebral disc integrity, and owing to the age of the osteoporotic population, disc degeneration is very common. Moreover, it has been shown that intervertebral disc narrowing is associated with an increased risk of vertebral fracture.7 The characteristics of the fracture should also be considered, and particularly its location, type, and severity.8 The risk of incident vertebral facture is dramatically increased when the prevalent fracture occurs between the T5 and T7 vertebrae for thoracic fractures, and between the L1 and L3 vertebrae for lumbar fractures.8 The risk is also increased when the deformation occurs in the anterior and middle parts of a vertebra. Finally, the greater the prevalent deformation, the greater the risk of incident fracture. Due to increased compression and shear loads in the vicinity of the vertebral fracture, the risk of subsequent vertebral fracture is particularly high on the vertebrae directly adjacent to the fracture.4

Global spine properties may also be involved in the etiopatho-genesis of the vertebral fracture cascade. Although there are conflicting data on this issue, one can speculate that osteoporotic vertebral fractures are associated with increases in thoracic curvature. More than 10 years ago, Cortet et al found a correlation between thoracic curvature and the vertebral deformity index (r=0.6, P<0.001) using a curviscope (a tool composed of angular potentiometers placed on the skin over the thoracolumbar area).9 The longitudinal part of the study also demonstrated a relationship between the occurrence of vertebral fracture at year 1 and 3 and an increase in spinal curvature at the thoracic level. However, other authors have not been able to confirm this relationship.10 Nevertheless, thoracic kyphosis is an independent and significant predictor of vertebral fracture. In a large prospective cohort of postmenopausal women with at least one vertebral fracture at baseline, Roux et al measured the kyphosis index (KI), which is defined as the percentage ratio between the maximum depth of thoracic curvature and the height measured from the T4 to the T12 vertebrae.11 Patients with the highest KI experienced significantly more new vertebral fractures over the 3-year study period (incidence, 27.36%) than those in the medium tertile (fracture incidence, 19.07%; relative risk [RR], 1.5; 95% confidence interval [CI], 1.19-1.96; P<0.001) or those in the lowest tertile (incidence, 17.31%; RR, 1.70; 95% CI, 1.32-2.21; P<0.001).




Finally, neurophysiological properties may also be involved in the fracture cascade. Although studies are scarce on this issue, it seems that vertebral fractures have a negative effect on balance control, with an increased risk of falling and, therefore, an increased risk of subsequent fracture.4 Muscle abnormalities characterized by a reliance on trunk muscle co-contraction have been identified in patients with vertebral fractures. Co-contraction of the trunk musculature is defined by a simultaneous contraction of agonist and antagonist muscles and causes an increase in spinal loading. In addition, electromyographic abnormalities in the thoracic erector spinae in patients with vertebral fracture have also been observed.4

The different factors involved in the vertebral fracture cascade are summarized in Figure 1.

Relationship between prevalent and incident vertebral fractures
Several studies have focused on the relationship between a prior vertebral fracture and the occurrence of a new vertebral fracture. In a meta-analysis (the first published on this issue), Klotzbuecher et al demonstrated that a prevalent vertebral fracture dramatically increases the risk of new vertebral fracture with an odds ratio of 4.4 (95% CI, 3.6-5.4).12 In this meta-analysis (see below), all types of prevalent fractures increased the risk of a new fracture but the strongest association was observed between prevalent vertebral fractures and incident vertebral fractures.

In this meta-analysis, about 50% of the patients were included on the basis of systematic spine radiographs (morphometric diagnosis). The remaining 50% patients were included on the basis of prior clinical vertebral fracture. Most of the studies selected included postmenopausal women.

These two types of patient recruitment did not seem to influence the magnitude of the association. Overall, the risk of subsequent fractures was shown to dramatically increase with the number of prevalent vertebral fractures. For example, women with at least 5 prevalent vertebral fractures (an unusual situation, fortunately) had a 35-fold higher risk of subsequent vertebral fracture than women with no vertebral fracture at baseline.13 More recently, Siris et al studied 2651 postmenopausal women from the placebo groups of both the FIT (Fracture Intervention Trial) and MORE trials (Multiple Outcomes of Raloxifene Evaluation).14 They confirmed the findings from Klotzbuecher et al12 and found that women with a semi-quantitative score (SQ) of 1 at baseline had a 3-fold higher risk of subsequent vertebral fracture after 2 years of follow-up than women with a SQ score of 0. A SQ score of 2 led to an approximately 5-fold increased risk and a SQ score of 3 led to an approximately 10-fold increased risk. Kanis et al published another meta-analysis including a large cohort (44 902 women and 15 259 men) and studied the relationship between a previous fracture and the subsequent risk of fracture.15 Unfortunately, in this meta-analysis there was no separate analysis focusing on vertebral fractures.


Figure 1
Figure 1. Etiopathogenesis of the vertebral fracture cascade.

After reference 4: Briggs et al. Osteoporos Int. 2007;18:575-584. © 2007, International Osteoporosis Foundation and National Osteoporosis Foundation.



Johnell et al studied a large cohort of patients (both men and women) with different types of prevalent fractures (hip, forearm, spine, and shoulder).16 They confirmed the association between the presence of a vertebral fracture at baseline and the risk of subsequent vertebral fracture. However, whereas the overall risk of subsequent fracture seemed to decrease over time, it was not the case for vertebral fractures. By contrast, in a previous study, Johnell et al had showed a decreasing risk of fracture over time in a subpopulation of patients with vertebral fracture requiring hospitalization.17


Figure 2
Figure 2.
Progression of the vertebral fracture cascade over time.

A. Fracture of the T6 vertebra at baseline.
B. Fractures of the T6 and T9 vertebrae at the 12-month evaluation.
C. Fractures of the T6, T9, and T11 vertebrae at the 24-month evaluation.



Roux et al focused their study on women with mild vertebral fractures.18 This is a particularly interesting population since, as previously mentioned, mild vertebral fractures are often underdiagnosed. Data were extracted from the placebo group of two phase 3 studies of strontium ranelate. After 4 years of follow-up, they found a 1.8-fold (95% CI, 1.3-2.4) increased risk of subsequent vertebral fracture in women with mild vertebral fractures at baseline. Obviously, the association was stronger in women with at least one grade ≥2 fracture at baseline (odds ratio, 2.7; 95% CI, 2.3-3.3).

Lindsay et al studied the risk of new vertebral fracture in the year following a fracture.19 Data were extracted from the placebo groups of four large clinical trials that assessed the efficacy of risedronate. A total of 2725 postmenopausal women were included. First, the authors confirmed previous findings and found that women with at least one vertebral fracture at baseline had a 5-fold increased risk of subsequent vertebral fracture compared with women with no vertebral fracture. In total, 381 women developed an incident vertebral fracture. Among them, 19.2% (95%CI, 13.6%-24.8%) had a new vertebral fracture in the subsequent year. A few years later, Lindsay et al evaluated the probability of vertebral fracture in a population of postmenopausal women with osteoporosis but without vertebral fracture at baseline.20 The population studied was the same as the one described previously and a Markov model was used to show the distribution of fracture prevalence over time. Lindsay et al found that an osteoporotic woman with no existing vertebral fractures had a 7.7% probability of having a vertebral fracture within 1 year. After 5 years, the percentage was 33%, and after 10 years, 55%.

Finally, in the recent GLOW study (Global Longitudinal study of Osteoporosis in Women) Gehlbach et al evaluated the association between a past history of fracture and the incidence of subsequent fractures.21 The strengths of the GLOW study were the high number of women recruited (n=51 762) and the high number of sites assessed for both prevalent and subsequent fractures. Vertebral fractures were clinically diagnosed and the diagnosis was confirmed by radiography. The association between a past history of vertebral fracture and the occurrence of a subsequent vertebral fracture was found to be very strong, with an odds ratio of 7.34 (95% CI, 5.42-9.92). Figure 2 illustrates the progression of the vertebral fracture cascade over time.

Prevalent vertebral fracture and risk of subsequent nonvertebral fracture

A prevalent vertebral fracture also increases the risk of nonvertebral fracture. The etiopathogenesis of this phenomenon is not fully understood—as in the case of vertebral fractures— although bone properties seem to play a major role. Bone properties reflect both bone quantity and bone quality, but in clinical practice only bone quantity can be measured using DXA. Even if BMD is usually lower in women with vertebral fractures, BMD alone cannot explain the increased risk of non- vertebral fracture in women with a prevalent vertebral fracture. Indeed, after adjustment for BMD, the increased risk is attenuated (by about 20%) but remains significant.12 This is also true for women with nonvertebral fractures at baseline (see below). In their meta-analysis, Klotzbuecher et al showed that among peri-/postmenopausal women the association with subsequent hip fracture was stronger than with all subsequent fractures combined, with corresponding odds ratios of 2.3 (95% CI, 2-2.8) and 1.8 (95% CI, 1.7-1.9), respectively.12 However, the association seems to be weaker for wrist fractures than for other subsequent fractures, with an odds ratio of 1.4 (95% CI, 1.3-2.4). The association between prevalent vertebral fracture and subsequent fractures was later demonstrated in several studies.16,19,22,23 Moreover, Johnell et al showed that the risk of nonvertebral fracture is highest the first year and decreases over time, although it remains significant after 5 years of follow-up.16 Van Staa et al showed that the risk is lower in elderly women (>85 years).22 They also demonstrated that the association was stronger in men than in women, whatever their age. These findings were confirmed in the Dubbo Osteoporosis Epidemiology Study, with odds ratio of 6.18 (95% CI, 4.17-9.14) for fractures of any type in men with a past history of vertebral fracture, and 2.52 (95% CI, 1.99- 3.19) in women.23

Prevalent nonvertebral fracture and the risk of subsequent fracture

There is also convincing data regarding the risk of subsequent fracture (both in the spine and at other sites) in women (but also in men) with prevalent nonvertebral fracture. The mechanisms underlying this association are probably similar to those previously described for the association between prevalent vertebral fracture and the occurrence of a subsequent nonvertebral fracture.

Prevalent nonvertebral fracture and risk of subsequent vertebral fracture
In the meta-analysis of Klotzbuecher et al, the odds ratio for a vertebral fracture in those with a prevalent wrist fracture was 1.7 (95% CI, 1.4-2.1).12 For those with a prevalent hip fracture the odds ratio was 2.5 (95% CI, 1.8-3.5). Johnell et al studied 1918 patients with fractures at various sites who were diagnosed and followed for 5 years in the department of radiology in Malmö University Hospital (Sweden).16 They found that the risk of vertebral fracture was higher in both men and women with a past history of hip fracture than in the general population. However, the risk was significant in younger people only. Interestingly, the findings were similar for those with a past history of shoulder fracture. Using the GPRD database (General Practice Research Database), van Staa et al had the opportunity to study 222 369 subjects (119 317 women and 103 052 men) who had sustained at least one fracture during follow-up.22 Whatever the site of the prevalent fracture (tibia/fibula/ankle, femur/hip, radius/ulna, ribs, and humerus), the odds ratios for a subsequent vertebral fracture were always increased. They ranged from 1.5 (95% CI, 1.3-1.8) for radius/ ulna to 4.3 (95% CI, 3.7-5.2) for ribs. The association was significant in both women and men but was stronger in men than in women. More recently, Gehlbach et al also studied the association between a past history of nonvertebral fracture and the occurrence of subsequent vertebral fractures in the GLOW cohort (n=51 762).21 As previously mentioned, one of the strengths of the GLOW study was the high number of fracture sites evaluated for both prevalent and incident fractures. The association was significant for prevalent nonvertebral fracture located at the rib and wrist, with respective odds ratios of 2.28 (95% CI, 1.64-3.17) and 1.37 (95% CI, 1.01-1.85). For the other sites of prevalent nonvertebral fracture (ie, hip, upper arm, ankle, lower leg, upper leg, clavicle, and pelvis) the associations were not significant.

Prevalent nonvertebral fracture and risk of subsequent nonvertebral fracture
Numerous studies have found an association between prevalent nonvertebral fracture and the risk of subsequent nonvertebral fracture. In the meta-analysis of Klotzbuecher et al, the authors found that a past history of wrist fracture increased not only the risk of subsequent wrist fracture, but also the risk of hip fracture.12 The odds ratios were 3.3 (95% CI, 2.0-5.3) and 1.9 (95% CI, 1.6-2.2), respectively. Similarly, a past history of hip fracture was shown to increase the risk of subsequent hip fracture, with an odds ratio of 2.3 (95% CI, 1.5-3.7) in peri- or postmenopausal women. In this meta-analysis, a prevalent fracture, regardless of its site, was shown to increase the risk of subsequent fracture (pooled vertebral and nonvertebral fractures) with an odds ratio of 2 (95% CI, 1.8- 2.1) in peri- and postmenopausal women. The conclusion of the meta-analysis of Kanis et al was quite similar, with an odds ratio of 1.86 (95% CI, 1.75-1.98).15 Moreover, adjustment for BMD did not modify the association, and the strength of the association was similar regardless of age. Johnell et al found that the risk of hip fracture was higher in women and men with a past history of both shoulder and hip fracture than in the general population.16 However, for those with a previous history of shoulder fracture, the risk was significant for younger people only. In addition, a past history of both shoulder and hip fractures increased the risk of wrist fracture. However, the association between a prevalent hip fracture and a subsequent forearm fracture was significant in men only. In the GPRD cohort, Van Staa et al confirmed these findings.22 They demonstrated that a previous fracture of the tibia/fibula/ankle, femur/hip, radius/ulna, ribs, and humerus increased the risk of subsequent fracture at these sites, with corresponding odds ratios ranging from 2 to 3. The strongest association was observed for a past history of radius/ulna fracture and the risk of humerus fracture, with an odds ratio of 5.8 (95% CI, 5.5-6.1). In addition, the association was stronger in men than in women, whereas, findings were quite similar overall whatever the age of the selected population. In the GLOW study, Gehlbach et al demonstrated that a prevalent fracture (whatever its site) multiplied the risk of subsequent fracture (of any bone) by 2.19 (95% CI, 2.03-2.35).21 The association was quite similar for the risk of incident hip fracture, where the odds ratio was 2.02 (95% CI, 1.55-2.63). Similarly, Gehlbach et al demonstrated that a past history of fracture increased the risk of fracture in other weight-bearing bones (ie, pelvis, upper leg, lower leg, ankle, foot, and knee, but not spine or hip) but also non–weight-bearing bones (ie, rib, wrist, upper arm, clavicle, hand, elbow, and shoulder).21 The corresponding odds ratios were 2.21 (95% CI, 1.96-2.49) and 2.15 (95% CI, 1.93-2.39), respectively. However, the association with a prevalent nonvertebral fracture (ie, either rib, hip, wrist, upper arm, ankle, lower leg, upper leg, clavicle, or pelvis) and the risk of subsequent hip fracture was significant for hip, upper leg, and pelvis only. The corresponding odds ratios were 3.50 (95% CI, 2.30-5.32), 2.15 (95% CI, 1.12-4.14), and 2.62 (95% CI, 1.44-4.77), respectively. The association between a previous nonvertebral fracture and the risk of subsequent other weight-bearing bone fracture was significant for nearly all the sites evaluated, except for the hip and clavicle. By contrast, the association between a prevalent nonvertebral fracture and the risk of subsequent non–weight-bearing bone was significant for rib, hip, wrist, and upper arm. Recently, Omsland et al studied the association between a prevalent hip fracture and the risk of a second hip fracture over a 10-year period.24 They collected data on prevalent hip fractures in Norwegian hospitals between 1999 and 2008. Among the 81 867 people who suffered from a first hip fracture, 6161 women (15%) and 1782 men (11%) suffered from a second hip fracture. The overall age-adjusted hazard ratio did not differ according to sex. However, by taking into account the higher risk of death in men than in women after a hip fracture, the corresponding age-adjusted hazard ratio was higher in men than in women: 1.40 (95% CI, 1.33-1.47).

Conclusion

In conclusion, a previous fracture dramatically increases the risk of subsequent fracture, whatever the site of the prevalent fracture (vertebral or nonvertebral), and this is the case for both vertebral and nonvertebral fractures. The concept of fracture cascade is particularly relevant since some of these fractures—which Bliuc et al25 call “major fractures” (vertebral, hip, pelvis, distal femur, proximal tibia, 3 or more simultaneous ribs, and proximal humerus)—are associated with an increased risk of death. This association was demonstrated for both prevalent major fractures and incident major fractures.

Finally, let’s consider the cost related to the fractures occurring as a result of the fracture cascade. In France, Maravic et al showed that in 2008, a total of 67 807 hospitalizations were related to osteoporotic hip fractures in women.26 The cost associated with these hospitalizations was €415.4 million and the cost of rehabilitation was €331.8 million. In Canada, Leslie et al analyzed the direct health care costs for 5 years post-fracture.27 They showed that the incremental median costs were highest the first year after a hip fracture. They were Can$25 306 in women, and Can$21 396 in men (in 2009 constant Canadian dollars). In addition, they also showed that —unsurprisingly—the costs decreased over time. However, in women and men who survived a hip fracture, the costs at 5 years remained above the prefracture costs. The figures were similar for vertebral fractures; however the costs at 1 year (and thereafter) were lower. For vertebral fractures, the costs were Can$15 392 in women and Can$11 309 in men at 1 year.

Fractures caused by osteoporosis represent a major health concern comparable to that posed by other severe diseases. Piscitelli et al studied the costs of hip fractures, acute myocardial infarctions, and strokes (both hemorrhagic and ischemic) between 2001 and 2005 in Italy.28 Overall, the number of hospitalizations related to these diseases increased between 2001 and 2005. In 2005, the costs related to hip fractures were similar to those associated with strokes (both hemorrhagic and ischemic) and were higher than those associated with acute myocardial infarctions. Rehabilitations costs associated with hip fractures and acute myocardial infarctions were comparable but lower than those associated with strokes. ■

References
1. Kanis JA; World Health Organization Scientific Group. Assessment of osteoporosis at the primary health care level. Sheffield, UK: WHO Collaborating Centre, University of Sheffield. 2008. Accessed March 21, 2011. Available at: http://www. shef.ac.uk/FRAX/.
2. Melton LJ 3rd, Chrischilles EA, Cooper C, Lane AW, Riggs BL. Perspective. How many women have osteoporosis? J Bone Miner Res. 1992;7:1005-1010.
3. Delmas PD, van de Langerijt L, Watts NB, et al. Underdiagnosis of vertebral fracture is a worldwide problem: the IMPACT study. J Bone Miner Res. 2005;20: 557-563.
4. Briggs AM, Greig AM, Wark JD. The vertebral fracture cascade in osteoporosis: a review of aetiopathogenesis. Osteoporos Int. 2007;18:575-584.
5. Aaron JE, Shore PA, Shore RC, Beneton M, Kanis JA. Trabecular architecture in women and men of similar bone mass with and without vertebral fracture: II. Three-dimensional histology. Bone. 2000;27:277-282.
6. Gilsanz V, Loro ML, Roe TF, Sayre J, Gilsanz R, Schulz EE. Vertebral size in elderly women with osteoporosis. Mechanical implications and relationship to fractures. J Clin Invest. 1995;95:2332-2337.
7. Sornay-Rendu E, Munoz F, Duboeuf F, Delmas PD; OFELY Study. Disc space narrowing is associated with an increased vertebral fracture risk in postmenopausal women: the OFELY Study. J Bone Miner Res. 2004;19:1994-1999.
8. Lunt M, O’Neill TW, Felsenberg D, et al; European Prospective Osteoporosis Study Group. Characteristics of a prevalent vertebral deformity predict subsequent vertebral fracture: results from the European Prospective Osteoporosis Study (EPOS). Bone. 2003;33:505-513.
9. Cortet B, Roches E, Logier R, et al. Evaluation of spinal curvatures after a recent osteoporotic vertebral fracture. Joint Bone Spine. 2002;69:201-208.
10. Schneider DL, von Mühlen D, Barrett-Connor E, Sartoris DJ. Kyphosis does not equal vertebral fractures: the Rancho Bernardo study. J Rheumatol. 2004;31: 747-752.
11. Roux C, Fechtenbaum J, Kolta S, Said-Nahal R, Briot K, Benhamou CL. Prospective assessment of thoracic kyphosis in postmenopausal women with osteoporosis. J Bone Miner Res. 2010;25:362-368.
12. Klotzbuecher CM, Ross PD, Landsman PB, Abbott TA 3rd, Berger M. Patients with prior fractures have an increased risk of future fractures: a summary of the literature and statistical synthesis. J Bone Miner Res. 2000;15:721-739.
13. Ross PD, Genant HK, Davis JW, Miller PD, Wasnich RD. Predicting vertebral fracture incidence from prevalent fractures and bone density among nonblack, osteoporotic women. Osteoporos Int. 1993;3:120-126.
14. Siris ES, Genant HK, Laster AJ, Chen P, Misurski DA, Krege JH. Enhanced prediction of fracture risk combining vertebral fracture status and BMD. Osteoporos Int. 2007;18:761-770.
15. Kanis JA, Johnell O, De Laet C, et al. A meta-analysis of previous fracture and subsequent fracture risk. Bone. 2004;35:375-382.
16. Johnell O, Kanis JA, Odén A, et al. Fracture risk following an osteoporotic fracture. Osteoporos Int. 2004;15:175-179.
17. Johnell O, Oden A, Caulin F, Kanis JA. Acute and long-term increase in fracture risk after hospitalization for vertebral fracture. Osteoporos Int. 2001;12:207- 214.
18. Roux C, Fechtenbaum J, Kolta S, Briot K, Girard M. Mild prevalent and incident vertebral fractures are risk factors for new fractures. Osteoporos Int. 2007;18: 1617-1624.
19. Lindsay R, Silverman SL, Cooper C. Risk of new vertebral fracture in the year following a fracture. JAMA. 2001;285:320-323.
20. Lindsay R, Pack S, Li Z. Longitudinal progression of fracture prevalence through a population of postmenopausal women with osteoporosis. Osteoporos Int. 2005;16:306-312.
21. Gehlbach S, Saag KG, Adachi JD, et al. Previous fractures at multiple sites increase the risk for subsequent fractures: the Global Longitudinal Study of Osteoporosis in Women. J Bone Miner Res. 2012;27:645-653.
22. van Staa TP, Leufkens HG, Cooper C. Does a fracture at one site predict later fractures at other sites? A British cohort study. Osteoporos Int. 2002;13:624- 629.
23. Center JR, Bliuc D, Nguyen TV, Eisman JA. Risk of subsequent fracture after lowtrauma fracture in men and women. JAMA. 2007;297:387-394.
24. Omsland TK, Emaus N, Tell GS, et al. Ten-year risk of second hip fracture. A NOREPOS study. Bone. 2013;52:493-497.
25. Bliuc D, Nguyen ND, Milch VE, Nguyen TV, Eisman JA, Center JR. Mortality risk associated with low-trauma osteoporotic fracture and subsequent fracture in men and women. JAMA. 2009;301:513-521.
26. Maravic M, Jouaneton B, Vainchtock A, Tochon V. Economic burden of osteoporosis in women: data from the 2008 French hospital database (PMSI). Clin Exp Rheumatol. 2012;30:222-227.
27. Leslie WD, Lix LM, Finlayson GS, Metge CJ, Morin SN, Majumdar SR. Direct healthcare costs for 5 years post-fracture in Canada: a long-term populationbased assessment. Osteoporos Int. 2013;24:1697-1705.
28. Piscitelli P, Iolascon G, Argentiero A. Incidence and costs of hip fractures vs strokes and acute myocardial infarction in Italy: comparative analysis based on national hospitalization records. Clin Interv Aging. 2012;7:575-583.



Keywords: etiopathogenesis; fracture cascade; nonvertebral fracture; osteoporosis; risk factor; vertebral fracture