Declining hip fracture burden in Sweden 1998-2019 and consequences for projections through 2050

doi:10.21203/rs.3.rs-2592193/v1

Download PDF

Research Article

Declining hip fracture burden in Sweden 1998-2019 and consequences for projections through 2050

https://doi.org/10.21203/rs.3.rs-2592193/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Although it may seem trivial to use national diagnosis databases to identify the number of patients, analyze time trends, and make projections, here we use hip fracture to illustrate the complexity and difficulty of estimating the reliable number of injury events. Readmissions after the index hip fracture are common, and register identification of a real second fracture is crucial for data accuracy. Therefore, we aimed to estimate the absolute and age-standardized number of hip fractures in Sweden during the past two decades to produce time trends and future projections. We also illustrate how the naïve use of register data leads to invalid estimates.

We used nationwide Swedish register data from 1998 to 2019 and a validated algorithm to calculate the annual absolute and age-standardized number of incident hip fractures over time. According to our algorithm, the total hip fracture burden was 335,399 incident events over the 22-year period, with a change from 16,180 in 1998 to 13,929 in 2019, a 14% decrease. One decade after the index hip fracture event, 80% of the patients had died and 11% had a new hip fracture. After considering the steady growth of the older population, the decline in the age-standardized number of hip fractures was 30%. With a continued similar reduction in hip fracture incidence, we can predict that 14,800 hip fractures will occur in 2034 and 12,000 in 2050 despite a doubling of the oldest old (≥80 years). Without an algorithm, a naïve estimate of the total number of hip fractures over the study period was 539,947, with a second 10-year hip fracture risk of 35%.

Accordingly, the naïve use of register data can dramatically inflate estimates of the number of fractures. We note an ongoing decline in absolute and age-standardized actual number of hip fractures in Sweden, with consequences for future projections.

hip fracture

orthopedics

epidemiology

aging

incidence

National medical databases and disease registers are widely used for descriptive purposes, such as determining disease burden and future projections of disease occurrence (1-3). For such purposes, high completeness and coding accuracy of register data are crucial. Equally important is the correct identification of incident events, separating complications from the index event, identifying true relapsing cases, and avoiding counting readmissions or hospital transfers as occurrences of new disease. Without carefully applied algorithms, this is a common issue for several disease categories, including myocardial infarction (4), stroke (5), pneumonia (6), and injuries (e.g., fragility fractures) (7).

Hip fracture is the most devastating fragility fracture, almost always leading to hospitalization, and is thus captured by hospital registries. This fracture is associated with substantial morbidity, increased mortality, and high societal costs (8). Worldwide 1.6 million hip fractures are estimated to occur annually (8) at an average of 80 years of age. They account for half of all fracture-related costs in aging populations (9, 10). With the nearly complete hospital database capture of hip fractures, they can help monitor disease burden and healthcare response after implementation of clinical and organizational improvements in the care for the elderly (11, 12).

Several studies have shown a marked variation worldwide in hip fracture incidence, with the highest rates reported in Scandinavian countries, including Sweden (13). A recent analysis based on national databases (14) forecasted a steep increase in fragility fracture burden in Scandinavia and most other European countries, with 30% more fractures in 2034 than in 2019. Older projections estimated a 2-3 times higher number of hip fractures in 2050 than in 1990 (15) and 2002 (16). Such trends have clear implications for public health. However, the validity of such projections depends on the accuracy of their data inputs, particularly hip fracture incidence in the recent past, and relies on national databases that capture the experience of entire populations. But these databases are often not straightforward to use as they were originally designed simply to record healthcare encounters, not integrate separate entries to distinguish multiple encounters for a single injury from two or more events.

To clarify trends in hip fracture incidence in Sweden and illustrate the potential problems with national database incidence studies, we estimated the actual number of hip fractures in Sweden from the past two decades, calculated time trends over this period, estimated 10-year risks of a new hip fracture, and made annual projections through 2050. We also demonstrate how the naïve use of register data can generate seriously inaccurate estimates of hip fracture burden.

Hip fractures captured by a national register and development of our database

We estimated hip fracture rates and numbers of hip fractures from 1998 to 2019 using the Swedish National Patient Register (NPR), maintained by the Swedish National Board of Health and Welfare (1). The register has covered all inpatient care in Sweden since 1987. By Swedish law, all private and publicly funded physicians and hospitals are mandated to deliver complete data to the NPR. The estimated positive predictive value for a first diagnosis of hip fracture in the NPR verified by hospital records ranges from 95 to 99%, depending on the study, and the negative predictive value has been estimated to 100% (1, 17, 18). In addition to the unique personal identification number (PIN) that is given to all Swedish citizens (19), the register includes information on the primary diagnosis, secondary diagnoses, comorbidities, type of injury, surgical procedures, type of department, and type of admission (elective or urgent). There is no variable to indicate whether the admission is the first (“index”) admission for an injury or a transfer or readmission for continuing care. Since 1997, the diagnoses have been coded according to the International Classification of Diseases version 10 (ICD-10; ICD-9 was previously used).

We used a database extracted from the NPR covering all hospital admissions for injuries from 1998 to 2019. For these patients, we additionally identified all previous injury admissions in 1987-1997 using the PIN for linkage. We identified possible hip fracture patients from 1997 to 2019 using ICD-10 codes S720, S721, or S722 (7). For 1987-1996, hip fracture events were coded with ICD-9 code 820. In addition, national annual age- and sex-specific population numbers from the study period and the expected number of Swedish inhabitants through 2050 were retrieved from Statistics Sweden.

Using these data, we conducted two sets of analyses, differing only in the definition of incident hip fractures. Our first set of analyses naïvely accepted as an incident fracture any hospitalization with one of the hip fracture codes enumerated above. In a parallel analysis, incident hip fracture events were identified using a previously developed and validated algorithm identifying an incident injury, as opposed to a readmission for a previous, “index” injury (7). Our complete national database from the NPR, in combination with our algorithm, shows that only 1.6% of hip fractures in 1998-2019 occurred before the age of 50 and because fractures coded as hip fractures at younger ages have a different etiology than those at older ages (20, 21), we restricted our analyses to women and men >50 years of age.

Description of the algorithm and code used to identify incident hip fractures

Derivation and validation of the prediction model variables and their multiplicative interactions used logistic regression modeling to develop a prediction model that could separate incident hospital admissions from readmissions. The algorithm is based on age, sex, diagnoses, time intervals between admissions, type of admission, type of injury, and type of hospital department (7) — the derived SAS code appears in Appendix file 1. The model was developed and validated using data representative of the Swedish national population (7) after the identification of correctly classified incident hip fractures based on individual hospital records review (7). One key point is that cases in which a patient’s discharge was followed by a new admission dated within 1 day were classified as transfers between departments or hospitals and counted as one fracture episode. By applying the model on a separate dataset by data split (7), we could evaluate and quantify the model's performance, enabling sensitivity analysis to illustrate the effect of misclassification. The model, based on the results from the analysis of the separate validation data set, has 96% sensitivity for incident hip fracture in patients with a hospitalization due to injury and 96% specificity (7).

Statistical analysis

Our study was conducted through four interconnected statistical analyses. Our initial analysis compared the annual number of algorithm-estimated incident hip fracture admissions from 1998 to 2019 with estimates generated by the naïve approach (including hospital transfers, readmissions, and complications). In sensitivity analyses, we examined the use of combinations of primary diagnosis, secondary diagnoses, and hospitalization sequence in each chain (within ≤1 day) of hospitalizations to identify incident fractures.

In addition, our results were compared to the official annual number of hip fractures presented by the Swedish National Board of Health and Welfare (https://www.socialstyrelsen.se/en). Their SAS code for hip fracture ascertainment can be found in Appendix file 2. The underlying assumption for the official statistics is that an individual recorded with one or more care events including an ICD code for hip fracture in a given year is classified as having had one incident hip fracture in that year. Each care event that includes an ICD code for hip fracture is counted as an incident case, even though the patient could have been treated for a complication after a previous year’s hip fracture, or had a hospital transfer near the turn of the year. Even if the event contains an ICD complication code along with a code for a hip fracture, this event is counted as a new hip fracture. On the other hand, patients who sustained two incident hip fractures during the same year are classified with only one hip fracture, The authority officially presents the number of hip fractures annually using 3-year moving averages.

Second, we evaluated the cumulative incidence of a new hip fracture after an index hip fracture event for the whole study period and within two time windows: 1998-2008 and 2009-2019, the latter a period with stronger national secondary prevention initiatives (22). We performed this analysis using three approaches: 1) the algorithm identification of each new hip fracture; 2) the naïve approach for a new hip fracture case identification; and 3) the naïve approach but with exclusions of a new admission dated within 1 day after the incident hip fracture event. Because death rates are high after a hip fracture event, which affects the likelihood of a new hip fracture (23), we used individual linkage to the Swedish Cause of Death Register (24) to enable consideration of death as a competing event.

Third, we assessed age- and sex-standardized trends in numbers of hip fractures during the time period 1998-2019. For each calendar year, we computed age- and sex-specific hip fracture rates, using age strata 50-65, 65-69, 70-74, 75-79, 80-84, 85-89, 90‑94, ≥95 years, based on algorithm-estimated number of fractures and annual age-and sex-specific population data from Statistics Sweden. These incidence rates were used to calculate age-standardized annual numbers and relative changes in hip fractures using direct standardization with the 2019 population as a reference.

Fourth, projections of hip fracture numbers from 2020 to 2050 were made under three assumptions: 1) a continued annual decline in hip fracture rates from 2020 through 2050 corresponding to the average annual percent change in 1998-2019; 2) rates constant from 2019 and onwards; and 3) a combination of scenario 1 and 2 with a continued downward trend in hip fracture rates from 2020 through 2029 followed by no further change. The rates were calculated by 5-year age bands and sex, and applied to projected population numbers from Statistics Sweden. We displayed the population statistics in absolute numbers as supplementary information and relative changes by sex and age (50-74 and ≥75 years) from 1998 to 2050.

In the analysis that accepted any diagnosis of hip fracture, there were 539,947 separate hip fracture admissions from 1998 to 2019 in individuals >50 years, 366,787 in women and 173,160 in men. In contrast, the numbers from our algorithm were considerably lower: 335,399 hip fractures (234,475 for women and 100,924 for men). The naïve analysis overestimated the actual number of hip fractures by 56% in women and 72% in men. Official statistics from the Swedish National Board of Health and Welfare claim 367,697 hip fractures from the 22-year study period, 112,824 in men and 254,830 in women, i.e. 12 and 9% higher than our algorithm-derived numbers.

Our algorithm identified 16,180 hip fracture cases over age 50 in Sweden in 1998 (Figure 1), declining to 13,929 in 2019, a 14% decrease. This reduction was due entirely to trends in women, with a 21% reduction in hip fractures versus a slight increase in men. The relative decline in the actual number of hip fractures between 2019 and 1998 is comparable to those estimated using the naïve hip fracture identification approach (17% reduction) or official statistics from the Swedish National Board of Health and Welfare (13% reduction).

We additionally evaluated risks of a second hip fracture. Within the first decade after an index hip fracture event, about 80% of hip fracture patients had died and 11% had suffered a new hip fracture (Figure 2). This pattern was evident (Appendix Figure 1) from the first half (1998-2008) and the second half of the observational period (2009-2019). In contrast, using a naïve approach, the proportion of patients with a second hip fracture from the first decade after the index hip fracture increased to 35% (Appendix Figure 2), an apparent tripling compared to the results based on our algorithm. If we used the naïve approach but ruled out adjacent admissions as unique cases, the 10-year risk of a new hip fracture was estimated to be 18% (Appendix Figure 3).

To better interpret the alterations in the number of hip fractures, we need to consider changes in age-specific incidence using the population's age-, sex- and year-specific structure. Between 1998 and 2019, the life expectancy of a 65-year-old Swede increased by 2.0 years in women and 3.2 years in men. During the same period, we observed a continuously increasing mean age at hip fracture, reaching 83.5 years in women and 80.7 years in men (Appendix Figure 4). In our Swedish setting, approximately 80% of all hip fractures occurred after 75 years (82% in women and 76% in men). Compared to 1998, the number of individuals 75+ years in 2019 had increased by 24%, with a moderate 13% increase in women compared to 41% in men in this age category.

The aging Swedish population would increase the number of hip fractures. Still, the decreases in age-specific fracture rates during 1998-2019 (Figure 3 and Appendix Figure 5) more than compensated for this, as shown by the dramatic decline in the age-standardized number of hip fractures over time (Figure 4). With the population in 2019 as a reference, a 30% reduction in the age-standardized number of hip fractures from 1998 through 2019 can be observed (32% in women and 26% in men), corresponding to about 6000 fewer hip fractures in 2019 compared to 1998 if the age-sex population structure had been identical.

There is an expected continued demographic trend towards an older population in Sweden. From 2020 to 2050, the number of Swedish citizens >75 years is projected to increase by 45%. The relative increase will be more pronounced in men, although the absolute number of older women will continue to be higher (Appendix Figure 6). Assuming the average proportional decline in hip fracture incidence from 1998 to 2019 continues from 2020 to 2050 and using the population changes projected by Statistics Sweden, we estimate the absolute number of hip fractures in 2050 will fall to less than 12,000 per year (Figure 5, dashed lines). The absolute number of hip fractures in men will approach those in women. By contrast, if the decline in hip fracture rates does not continue, we can naturally expect an increase in the total number of hip fracture events because of the growing number of older individuals, bringing the total to about 22,000 in 2050 (Figure 5, dotted lines). In the third intermediate scenario, there would be approximately 18,500 hip fractures per year in 2050 (Figure 5, dash-dot lines). In a 15-year projection from 2019, the number of hip fractures expected in 2034 would be 14,380 under scenario 1, 19,200 under scenario 2, and 16,300 under scenario 3. These 2034 numbers sharply contrast with the 32,200 hip fractures in Sweden recently projected by Kanis and co-workers (5) .

Sensitivity analyses

In a sensitivity analysis (Appendix Figure 7), the algorithm estimates of the number of hip fractures were similar irrespective of whether we considered only the primary diagnosis at the first hospitalization in each unique chain of hospitalizations, the primary diagnosis at any hospitalization in each chain, or if we used primary or secondary diagnoses (but not both) in each chain of hospitalizations. This congruence in estimates was observed in women and men. Regardless of the method, there was a steady decline in incidence in both sexes from 1998 to 2019 (Appendix Figure 7).

This study demonstrates that declining incidence rates of hip fractures have more than offset the aging Swedish population, resulting in a marked decrease in the number of hip fractures in Sweden since 1998. With age- and sex-standardized calculations, the reduction is 30%. We also show that incident hip fractures and recurrent hip fracture events in Sweden can be substantially over-reported without careful use of population fracture data.

Our findings corroborate previous reports of a continuous decline in age-standardized actual hip fracture incidence in Swedish women and men (25). These authors used individually validated data from northern Sweden, though only from 1993 to 2005 (25). Norwegian investigators have similarly observed a decline in age-standardized hip fracture incidence from 1999 to 2013 using careful methodology (26). Other Scandinavian studies using less formal methods have also seen downward trends in recent years (27-30). We observed a downward trend in hip fracture events irrespective how fractures were ascertained, finding it with a naïve analysis, with official Swedish statistics as well as with our more accurate algorithm-based approach. The same downward trend has also been seen in other European settings (31-34), while trends have plateaued in North America (35). In contrast, hip fracture rates are increasing in Asia (36-39).

Multiple potential causes may explain the downward trend in Swedish hip fracture rates, including a more healthy lifestyle with the avoidance of sedentary behavior (40, 41), a reduction in cigarette smoking (down to a 6% prevalence in recent years (42)), healthier dietary habits, including higher intake of fruit and vegetables (43), reduced consumption of milk and higher consumption of fermented dairy products (44, 45), and lowered intakes of retinol/vitamin A (46, 47). Moreover, there have been increases in body weight (48, 49) and more common use of hip arthroplasties, precluding hip fracture injuries (50). Positive, healthy lifestyle trends are likely to continue in Nordic countries, with strong influences from different authorities advocating and instructing the inhabitants on how to live a healthy life ̶ the Nordic nutritional recommendations are just one example (51).

Other factors may have played a role in lowering hip fracture rates. Increases in cataract surgery in Sweden (52) could reduce the risk of falls and subsequent hip fractures (53). Because of a possible direct causal link between cardiovascular diseases and the risk of hip fracture (54), the 40% reduction in both ischemic heart disease (55) and stroke (56) incidence in Sweden during the past two decades may also have contributed to the reduced hip fracture burden.

Some may argue that preventing fractures by bone-specific drugs and Fracture Liaison Services (22, 57, 58) should have affected the number of hip fractures. However, from 2005 to 2020, among Swedish women older than 55 years, the prevalence of bisphosphonate users decreased from 4.85% to 3.6%, while in men the corresponding proportion has been approximately 1%. This pattern is unlikely to affect hip fracture rates at a population level (59). Moreover, we found a similar age-adjusted decrease in hip fracture in men and women, even though women have a higher frequency of bone-specific drug treatment. Over the past decade, denosumab has been introduced as an additional bone-specific drug. In 2019, approximately 0.7% of all women and 0.1% of all men >55 years were prescribed the drug. Our results show a decline in age-adjusted hip fracture burden before the initiation of the Fracture Liaison Services in 2012 and the proportion of patients with a new hip fracture remained stable during our study period.

A non-clinical factor contributing to the declining hip fracture rates might be the fact that in Sweden, the immigrant population has grown in recent decades [Statistics Sweden], and many have moved from settings with a lower risk of fractures than in Scandinavia. From 2000 through 2019, the proportion of foreign-born women and men >75 years in Sweden increased from 7 to 12%.

The burden of hip fracture in Sweden has likely been exaggerated in previous reports (13-15), with estimates 25 to 40% (or more) higher than with our validated algorithm. These inaccuracies in the number of hip fractures have correspondingly led to inflated projections of future risk (10, 16, 60, 61), suggesting a doubling of cases in Sweden (16) by 2050. We also estimate a higher number of hip fractures 30 years from now if hip fracture rates such as those in recent years remain constant. However, we estimate that the total number will only reach 20,000 annually, even with this conservative approach, not 30,000 per year as projected in older analyses (16). Of note, in a different scenario, we estimate that the number of hip fractures in 2050 would be lower despite a growing number of older adults if the incidence trends we observed in the past two decades continue into the mid-century. Previous projections have ignored current and future trends in hip fracture rates (10, 16, 60, 61).

Our results are important for clinicians.The erroneously high estimated number of hip fractures in Sweden (14) is used in the widely adopted fracture risk calculator, FRAX, with a consequential overestimate of the FRAX 10-year risk of hip fracture for individual patients. The FRAX risk estimate is used in decision trees to recommend treatment with bone-specific medications (62, 63). The ratio between Swedish hip fracture rates and those of other types of fragility fractures is used in the FRAX calculations for European and non-European countries without complete register information (14, 64, 65). Inaccuracies in the calculation of Swedish incident hip fracture numbers are therefore also a concern in many other countries.

Similarly, FRAX seriously overestimates the risks of second fractures. We computed the 10-year FRAX probability of a hip fracture for Swedish women aged 80-85 years of average height and weight and with a previous fracture of any type but without any other risk factors. The estimate is 20-22%, considerably higher than those we and other researchers estimated using population-based databases (66-68). In addition, according to the FRAX calculator, the 10-year risk of hip fracture is greater than 50% in that clinical setting if there is additionally a parental history of hip fracture, a marker of risk inferior to the patient’s own history of a previous hip fracture. The present study confirmed (66-68) that only about 10% of hip fracture patients have a second hip fracture within 10 years, and this moderate risk may be largely explained by high mortality rates following a hip fracture (23, 69).

The validity of our calculated number of hip fractures has been confirmed by manual case identification using individual patient hospital records in localized regions in Sweden and with external validity to national data (25, 70). However, whether the potential for substantial overestimation of fracture risk from national databases can be generalized to settings outside Sweden requires detailed validation of the case identification algorithms (if any) used. The naïve use of national hospital record data has been a common practice for some time (13). Nonetheless,country-specific FRAX fracture risk assessments based on register data have been developed for 77 countries (71). The assumptions and calculations underlying these predictions have not been published (72-74), limiting their relevance as a prediction tool (75, 76).

In addition to our algorithm for identifying incident injury cases (7), our study is strengthened by the national population-based design and individual-based register data linked using the Swedish PIN. The trend of lower hip fracture rates is gratifying, but our study design cannot decipher which individual lifestyle factors have most tangibly driven this development.

The methodological issues we faced in ascertaining hip fracture cases in a large population are relevant for other disease endpoints for which it is important to identify incident cases of disease, e.g., myocardial infarction (4, 77), stroke (5, 78), colorectal cancer (79), pneumonia (6), and different types of injuries (7) Methodology for accurate estimates requires an ascertainment algorithm validated against individual medical records. Using raw register data may result in serious bias. Many difficulties can be overcome in databases with unique PINs that enable reliable record linkage. This linkage allows for multiple medical encounters to be used for the identification of incident disease events, avoiding double counting. The issues we describe are even more important when using administrative databases such as those from Medicare in the U.S., databases without direct individual identifiers (80, 81).

Our findings show a substantial over-reporting of hip fractures in Sweden, and over the past two decades, there has been a continued decline in the overall and age-standardized number of hip fractures in Sweden. Accordingly, Sweden's future hip fracture burden will almost certainly be lower than previously calculated (14-16). This lower fracture incidence will affect health economic analyses and fracture prediction tools in Sweden and other countries.

Disclaimer:

Rolf Gedeborg is also employed by the Medical Products Agency (MPA) in Sweden. The MPA is a Swedish Government Agency. The views expressed in this article may not represent the views of the MPA.

DISCLOSURE PAGE

Disclosure of potential conflicts of interest.

The authors declare no conflict of interest.

Data Availability Statement. Data are available upon application to the National Board of Health and Welfare. Our code to identify incident hip fracture events can be accessed from the corresponding author.

Acknowledgments

Contributors: KM designed the study and drafted the manuscript. KM and RG analyzed the data. JAB, LB, MK, SL, and HM contributed to interpreting the data and revising the manuscript. KM had primary responsibility for the final content. All authors read and approved the final manuscript.

Funding: The study was supported by grants from the Swedish Research Council. The study funders had no role in the study design, data collection, data analysis, data interpretation, writing of the report, or the decision to submit the article for publication.

Ethical approval: The study was approved by the Regional Research Ethics Board at Uppsala University.

The SAS code for calculating incident injuries, including hip fractures, can be found in the Appendix file 1.

Ludvigsson JF, Andersson E, Ekbom A, et al. External review and validation of the Swedish national inpatient register. BMC Public Health. 2011;11:450. doi:10.1186/1471-2458-11-450
Calltorp J, Adami HO, Astrom H, et al. Country profile: Sweden. Lancet. 1996;347(9001):587-94.
Thygesen LC, Ersboll AK. When the entire population is the sample: strengths and limitations in register-based epidemiology. Eur J Epidemiol. 2014;29(8):551-8. doi:10.1007/s10654-013-9873-0
Modig K, Berglund A, Talback M, Ljung R, Ahlbom A. Estimating incidence and prevalence from population registers: example from myocardial infarction. Scandinavian journal of public health. 2017;45(17_suppl):5-13. doi:10.1177/1403494817702327
Ellekjaer H, Holmen J, Kruger O, Terent A. Identification of incident stroke in Norway: hospital discharge data compared with a population-based stroke register. Stroke. 1999;30(1):56-60. doi:10.1161/01.str.30.1.56
Naucler P, Henriques-Normark B, Hedlund J, Galanis I, Granath F, Ortqvist A. The changing epidemiology of community-acquired pneumonia: nationwide register-based study in Sweden. J Intern Med. 2019;286(6):689-701. doi:10.1111/joim.12956
Gedeborg R, Engquist H, Berglund L, Michaelsson K. Identification of incident injuries in hospital discharge registers. Epidemiology. 2008;19(6):860-7.
Riggs BL, Melton LJd. The prevention and treatment of osteoporosis N Engl J Med. 1992;327:620-7.
Kanis JA, Oden A, Johnell O, Jonsson B, de Laet C, Dawson A. The burden of osteoporotic fractures: a method for setting intervention thresholds. Osteoporos Int. 2001;12(5):417-27. doi:10.1007/s001980170112
Hernlund E, Svedbom A, Ivergard M, et al. Osteoporosis in the European Union: medical management, epidemiology and economic burden. A report prepared in collaboration with the International Osteoporosis Foundation (IOF) and the European Federation of Pharmaceutical Industry Associations (EFPIA). Archives of osteoporosis. 2013;8(1-2):136. doi:10.1007/s11657-013-0136-1
Qureshi A, Gwyn Seymour D. Growing knowledge about hip fracture in older people. Age Ageing. 2003;32(1):8-9. doi:10.1093/ageing/32.1.8
Nordstrom P, Gustafson Y, Michaelsson K, Nordstrom A. Length of hospital stay after hip fracture and short term risk of death after discharge: a total cohort study in Sweden. BMJ. 2015;350:h696. doi:10.1136/bmj.h696
Johnell O, Gullberg B, Allander E, Kanis JA. The apparent incidence of hip fracture in Europe: a study of national register sources. Osteoporos. Int. 1992;2:298-302.
Kanis JA, Norton N, Harvey NC, et al. SCOPE 2021: a new scorecard for osteoporosis in Europe. Archives of osteoporosis. 2021;16(1):82. doi:10.1007/s11657-020-00871-9
Gullberg B, Johnell O, Kanis JA. World-wide projections for hip fracture. Osteoporos Int. 1997;7(5):407-13.
Rosengren BE, Karlsson MK. The annual number of hip fractures in Sweden will double from year 2002 to 2050: projections based on local and nationwide data. Acta Orthop. 2014;85(3):234-7. doi:10.3109/17453674.2014.916491
Michaelsson K, Baron JA, Farahmand BY, et al. Hormone replacement therapy and risk of hip fracture: population based case-control study. The Swedish Hip Fracture Study Group. BMJ. 1998;316(7148):1858-63.
Michaëlsson K, Melhus H, Ferm H, Ahlbom A, Pedersen NL. Genetic liability to fractures in the elderly. Arch Intern Med. 2005;165(16):1825-30.
Ludvigsson JF, Otterblad-Olausson P, Pettersson BU, Ekbom A. The Swedish personal identity number: possibilities and pitfalls in healthcare and medical research. Eur J Epidemiol. 2009;24(11):659-67. doi:10.1007/s10654-009-9350-y
Robinson CM, Court-Brown CM, McQueen MM, Christie J. Hip fractures in adults younger than 50 years of age. Epidemiology and results. Clin Orthop Relat Res. 1995(312):238-46.
Verettas DA, Galanis B, Kazakos K, Hatziyiannakis A, Kotsios E. Fractures of the proximal part of the femur in patients under 50 years of age. Injury. 2002;33(1):41-5. doi:10.1016/s0020-1383(01)00070-5
Andreasen C, Solberg LB, Basso T, et al. Effect of a Fracture Liaison Service on the Rate of Subsequent Fracture Among Patients With a Fragility Fracture in the Norwegian Capture the Fracture Initiative (NoFRACT): A Trial Protocol. JAMA Netw Open. 2018;1(8):e185701. doi:10.1001/jamanetworkopen.2018.5701
Michaelsson K, Nordstrom P, Nordstrom A, et al. Impact of hip fracture on mortality: a cohort study in hip fracture discordant identical twins. J Bone Miner Res. 2014;29(2):424-31. doi:10.1002/jbmr.2029
Brooke HL, Talback M, Hornblad J, et al. The Swedish cause of death register. Eur J Epidemiol. 2017;32(9):765-73. doi:10.1007/s10654-017-0316-1
Bergstrom U, Jonsson H, Gustafson Y, Pettersson U, Stenlund H, Svensson O. The hip fracture incidence curve is shifting to the right. Acta Orthop. 2009;80(5):520-4. doi:10.3109/17453670903278282
Sogaard AJ, Holvik K, Meyer HE, et al. Continued decline in hip fracture incidence in Norway: a NOREPOS study. Osteoporos Int. 2016;27(7):2217-22. doi:10.1007/s00198-016-3516-8
Nordstrom P, Bergman J, Ballin M, Nordstrom A. Trends in Hip Fracture Incidence, Length of Hospital Stay, and 30-Day Mortality in Sweden from 1998-2017: A Nationwide Cohort Study. Calcif Tissue Int. 2022;111(1):21-8. doi:10.1007/s00223-022-00954-4
Rosengren BE, Bjork J, Cooper C, Abrahamsen B. Recent hip fracture trends in Sweden and Denmark with age-period-cohort effects. Osteoporos Int. 2017;28(1):139-49. doi:10.1007/s00198-016-3768-3
Meyer AC, Ek S, Drefahl S, Ahlbom A, Hedstrom M, Modig K. Trends in Hip Fracture Incidence, Recurrence, and Survival by Education and Comorbidity: A Swedish Register-based Study. Epidemiology. 2021;32(3):425-33. doi:10.1097/EDE.0000000000001321
Karampampa K, Ahlbom A, Michaelsson K, Andersson T, Drefahl S, Modig K. Declining incidence trends for hip fractures have not been accompanied by improvements in lifetime risk or post-fracture survival--A nationwide study of the Swedish population 60 years and older. Bone. 2015;78:55-61. doi:10.1016/j.bone.2015.04.032
Longo UG, Vigano M, de Girolamo L, Banfi G, Salvatore G, Denaro V. Epidemiology and Management of Proximal Femoral Fractures in Italy between 2001 and 2016 in Older Adults: Analysis of the National Discharge Registry. Int J Environ Res Public Health. 2022;19(24). doi:10.3390/ijerph192416985
Unim B, Minelli G, Da Cas R, et al. Trends in hip and distal femoral fracture rates in Italy from 2007 to 2017. Bone. 2021;142:115752. doi:10.1016/j.bone.2020.115752
Jedynasty K, Zieba M, Adamski J, et al. Seasonally Dependent Change of the Number of Fractures after 50 Years of Age in Poland-Analysis of Combined Health Care and Climate Datasets. Int J Environ Res Public Health. 2022;19(15). doi:10.3390/ijerph19159467
Lippuner K, Rimmer G, Stuck AK, Schwab P, Bock O. Hospitalizations for major osteoporotic fractures in Switzerland: a long-term trend analysis between 1998 and 2018. Osteoporos Int. 2022;33(11):2327-35. doi:10.1007/s00198-022-06481-0
Lewiecki EM, Chastek B, Sundquist K, et al. Osteoporotic fracture trends in a population of US managed care enrollees from 2007 to 2017. Osteoporos Int. 2020;31(7):1299-304. doi:10.1007/s00198-020-05334-y
Shoji A, Gao Z, Arai K, Yoshimura N. 30-year trends of hip and vertebral fracture incidence in Japan: a systematic review and meta-analysis. J Bone Miner Metab. 2022;40(2):327-36. doi:10.1007/s00774-021-01288-1
Kawaji H, Ishii M, Sasaki K, Takakubo Y, Ishikawa H, Takagi M. Decennial trends in hip fractures over 20 years in Yamagata Prefecture, Japan. J Orthop Sci. 2022. doi:10.1016/j.jos.2022.01.001
Dong Y, Zhang Y, Song K, Kang H, Ye D, Li F. What was the Epidemiology and Global Burden of Disease of Hip Fractures From 1990 to 2019? Results From and Additional Analysis of the Global Burden of Disease Study 2019. Clin Orthop Relat Res. 2022. doi:10.1097/CORR.0000000000002465
Hou XL, Liu JY, Fan XH, et al. Secular trends of incidence and hospitalization cost of hip fracture in Tangshan, China. Osteoporos Int. 2022;33(1):89-96. doi:10.1007/s00198-021-06050-x
Stattin K, Hallmarker U, Arnlov J, James S, Michaelsson K, Byberg L. Decreased Hip, Lower Leg, and Humeral Fractures but Increased Forearm Fractures in Highly Active Individuals. J Bone Miner Res. 2018;33(10):1842-50. doi:10.1002/jbmr.3476
Stattin K, Michaelsson K, Larsson SC, Wolk A, Byberg L. Leisure-Time Physical Activity and Risk of Fracture: A Cohort Study of 66,940 Men and Women. J Bone Miner Res. 2017;32(8):1599-606. doi:10.1002/jbmr.3161
Olofsson H, Byberg L, Mohsen R, Melhus H, Lithell H, Michaëlsson K. Smoking and the risk of fracture in older men. Journal of Bone and Mineral Research. 2005;20(7):1208-15.
Byberg L, Bellavia A, Orsini N, Wolk A, Michaelsson K. Fruit and vegetable intake and risk of hip fracture: a cohort study of Swedish men and women. J Bone Miner Res. 2015;30(6):976-84. doi:10.1002/jbmr.2384
Michaelsson K, Wolk A, Langenskiold S, et al. Milk intake and risk of mortality and fractures in women and men: cohort studies. BMJ. 2014;349:g6015. doi:10.1136/bmj.g6015
Michaelsson K, Wolk A, Lemming EW, Melhus H, Byberg L. Intake of Milk or Fermented Milk Combined With Fruit and Vegetable Consumption in Relation to Hip Fracture Rates: A Cohort Study of Swedish Women. J Bone Miner Res. 2018;33(3):449-57. doi:10.1002/jbmr.3324
Melhus H, Michaëlsson K, Kindmark A, et al. Excessive dietary intake of vitamin A is associated with reduced bone mineral density and increased risk for hip fracture. Ann Intern Med. 1998;129(10):770-8.
Michaelsson K, Lithell H, Vessby B, Melhus H. Serum retinol levels and the risk of fracture. N Engl J Med. 2003;348(4):287-94. doi:10.1056/NEJMoa021171
Farahmand BY, Michaëlsson K, Baron JA, Persson PG, Ljunghall S. Body size and hip fracture risk. Swedish Hip Fracture Study Group. Epidemiology. 2000;11(2):214-9.
Zheng R, Byberg L, Larsson SC, Hoijer J, Baron JA, Michaelsson K. Prior loss of body mass index, low body mass index, and central obesity independently contribute to higher rates of fractures in elderly women and men. J Bone Miner Res. 2021;36(7):1288-99. doi:10.1002/jbmr.4298
Kjeldgaard HK, Meyer HE, O'Flaherty M, et al. Impact of total hip replacements on the incidence of hip fractures in Norway during 1999-2019. A NOREPOS study. J Bone Miner Res. 2022. doi:10.1002/jbmr.4660
Christensen JJ, Arnesen EK, Andersen R, et al. The Nordic Nutrition Recommendations 2022 - principles and methodologies. Food Nutr Res. 2020;64. doi:10.29219/fnr.v64.4402
The Swedish National Cataract Register. https://rcsyd.se/kataraktreg/.
Gutierrez-Robledo LM, Villasis-Keever MA, Avila-Avila A, Medina-Campos RH, Castrejon-Perez RC, Garcia-Pena C. Effect of Cataract Surgery on Frequency of Falls among Older Persons: A Systematic Review and Meta-Analysis. J Ophthalmol. 2021;2021:2169571. doi:10.1155/2021/2169571
Sennerby U, Melhus H, Gedeborg R, et al. Cardiovascular diseases and risk of hip fracture. JAMA. 2009;302(15):1666-73. doi:10.1001/jama.2009.1463
Journath G, Hammar N, Elofsson S, et al. Time Trends in Incidence and Mortality of Acute Myocardial Infarction, and All-Cause Mortality following a Cardiovascular Prevention Program in Sweden. PLoS One. 2015;10(11):e0140201. doi:10.1371/journal.pone.0140201
Rosengren A, Giang KW, Lappas G, Jern C, Toren K, Bjorck L. Twenty-four-year trends in the incidence of ischemic stroke in Sweden from 1987 to 2010. Stroke. 2013;44(9):2388-93. doi:10.1161/STROKEAHA.113.001170
Geiger I, Kammerlander C, Hofer C, et al. Implementation of an integrated care programme to avoid fragility fractures of the hip in older adults in 18 Bavarian hospitals - study protocol for the cluster-randomised controlled fracture liaison service FLS-CARE. BMC geriatrics. 2021;21(1):43. doi:10.1186/s12877-020-01966-1
Wu CH, Tu ST, Chang YF, et al. Fracture liaison services improve outcomes of patients with osteoporosis-related fractures: A systematic literature review and meta-analysis. Bone. 2018;111:92-100. doi:10.1016/j.bone.2018.03.018
Abrahamsen B, Vestergaard P. Declining incidence of hip fractures and the extent of use of anti-osteoporotic therapy in Denmark 1997-2006. Osteoporos Int. 2010;21(3):373-80. doi:10.1007/s00198-009-0957-3
Svedbom A, Hernlund E, Ivergard M, et al. Osteoporosis in the European Union: a compendium of country-specific reports. Archives of osteoporosis. 2013;8(1-2):137. doi:10.1007/s11657-013-0137-0
Cooper C, Campion G, Melton LJd. Hip fractures in the elderly: a world-wide projection. Osteoporos Int. 1992;2:285-9.
Ford N, Norris SL, Hill SR. Clarifying WHO's position on the FRAX(R) tool for fracture prediction. Bull World Health Organ. 2016;94(12):862. doi:10.2471/BLT.16.188532
Curtis EM, Reginster JY, Al-Daghri N, et al. Management of patients at very high risk of osteoporotic fractures through sequential treatments. Aging clinical and experimental research. 2022;34(4):695-714. doi:10.1007/s40520-022-02100-4
Kebaetse M, Nkhwa S, Mogodi M, et al. A country-specific FRAX model for Botswana. Archives of osteoporosis. 2021;16(1):90. doi:10.1007/s11657-021-00965-y
Saleh YAL, Sulimani RA, Alomary S, et al. Incidence of hip fracture in Saudi Arabia and the development of a FRAX model. Archives of osteoporosis. 2022;17(1):56. doi:10.1007/s11657-022-01085-x
Melton LJ, 3rd, Kearns AE, Atkinson EJ, et al. Secular trends in hip fracture incidence and recurrence. Osteoporos Int. 2009;20(5):687-94. doi:10.1007/s00198-008-0742-8
Center JR, Bliuc D, Nguyen TV, Eisman JA. Risk of subsequent fracture after low-trauma fracture in men and women. JAMA. 2007;297(4):387-94. doi:10.1001/jama.297.4.387
Omsland TK, Emaus N, Tell GS, et al. Ten-year risk of second hip fracture. A NOREPOS study. Bone. 2013;52(1):493-7. doi:10.1016/j.bone.2012.09.009
Lyles KW, Colón-Emeric CS, Magaziner JS, et al. Zoledronic acid and clinical fractures and mortality after hip fracture. New England Journal of Medicine. 2007;357(18):1799-809.
Leavy B, Michaelsson K, Aberg AC, Melhus H, Byberg L. The Impact of Disease and Drugs on Hip Fracture Risk. Calcif Tissue Int. 2017;100(1):1-12. doi:10.1007/s00223-016-0194-7
Chotiyarnwong P, McCloskey EV, Harvey NC, et al. Is it time to consider population screening for fracture risk in postmenopausal women? A position paper from the International Osteoporosis Foundation Epidemiology/Quality of Life Working Group. Archives of osteoporosis. 2022;17(1):87. doi:10.1007/s11657-022-01117-6
Jarvinen TL, Jokihaara J, Guy P, et al. Conflicts at the heart of the FRAX tool. CMAJ : Canadian Medical Association journal = journal de l'Association medicale canadienne. 2014;186(3):165-7. doi:10.1503/cmaj.121874
Jarvinen TL, Michaelsson K, Aspenberg P, Sievanen H. Osteoporosis: the emperor has no clothes. J Intern Med. 2015;277(6):662-73. doi:10.1111/joim.12366
Jarvinen TL, Michaelsson K, Jokihaara J, et al. Overdiagnosis of bone fragility in the quest to prevent hip fracture. BMJ. 2015;350:h2088. doi:10.1136/bmj.h2088
Bonnett LJ, Snell KIE, Collins GS, Riley RD. Guide to presenting clinical prediction models for use in clinical settings. BMJ. 2019;365:l737. doi:10.1136/bmj.l737
Collins GS, Reitsma JB, Altman DG, Moons KG. Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): the TRIPOD statement. Ann Intern Med. 2015;162(1):55-63. doi:10.7326/M14-0697
Mandair D, Tiwari P, Simon S, Colborn KL, Rosenberg MA. Prediction of incident myocardial infarction using machine learning applied to harmonized electronic health record data. BMC Med Inform Decis Mak. 2020;20(1):252. doi:10.1186/s12911-020-01268-x
Vodencarevic A, Weingartner M, Caro JJ, et al. Prediction of Recurrent Ischemic Stroke Using Registry Data and Machine Learning Methods: The Erlangen Stroke Registry. Stroke. 2022;53(7):2299-306. doi:10.1161/STROKEAHA.121.036557
Alsadhan N, Almaiman A, Pujades-Rodriguez M, et al. A systematic review of methods to estimate colorectal cancer incidence using population-based cancer registries. BMC medical research methodology. 2022;22(1):144. doi:10.1186/s12874-022-01632-7
D'Arcy ME, Sturmer T, Sandler RS, et al. Healthcare system engagement and algorithm-identified cancer incidence following initiation of a new medication. Pharmacoepidemiol Drug Saf. 2022. doi:10.1002/pds.5556
Baron JA, Karagas M, Barrett J, et al. Basic epidemiology of fractures of the upper and lower limb among Americans over 65 years of age. Epidemiology. 1996;7(6):612-8. doi:10.1097/00001648-199611000-00008

Appendixfile1.docx
Appendix File 1. The SAS code for the identification of incident hip fractures. The algorithm is based on age, sex, diagnoses, time intervals between admissions, type of admission, type of injury, and type of hospital department (7).
Appendixfile2.txt
Appendix File 2. The SAS code used by the Swedish National Board of Health and Welfare ( href="https://www.socialstyrelsen.se/en">https://www.socialstyrelsen.se/en) to calculate the official annual number of hip fractures in Sweden. The SAS code also displays the authority’s approach when calculating all annual hip fracture events (the naïve approach) and only the annual number of first hip fractures.
Appendixfigure1.pdf
Appendix Figure 1. Cumulative incidence of a second hip fracture and death in 1998-2008 and 2009-2019.
Appendixfigure2.pdf
Appendix Figure 2. Cumulative incidence of a second hip fracture and death in 1998-2008 and 2009-2019 using the naïve approach to identify a new hip fracture.
Appendixfigure3.pdf
Appendix Figure 3. Cumulative incidence of a second hip fracture and death in 1998-2008 and 2009-2019 using the naïve approach but ruled out adjacent admissions as unique new cases.
Appendixfigure4.pdf
Appendix Figure 4. Mean age of predicted incident hip fractures from 1998 through 2019. Grey shaded area indicates a 95% confidence interval.
Appendixfigure5.pdf
Appendix Figure 5. Hip fracture incidence from 1998 through 2019: 1) estimated by a prediction algorithm for incident fracture events) and observed population structure (square); 2) estimated by a naïve incidence calculation of all admissions and observed population structure (circle); and 3) estimated using prediction algorithm rates with age-standardization to the 2019 (population of 2019) predicted incident admission (triangle).
Appendixfigure6.pptx
Appendix Figure 6. Swedish population changes in people ≥50 years from 1998 through 2050. The cross-hatched areas represent ages 50-74 years, with the light grey area corresponding to the number of women and the dark grey area to the number of men (primary vertical axis). The dashed lines correspond to the relative change in the population, expressed as percentages relative to 1998 (secondary vertical axis). The solid areas and lines represent women and men aged ≥75 years.
Appendixfigure7.pdf
Appendix Figure 7. Incidence of hip fracture from 1998 through 2019: 1) estimated using a prediction algorithm and primary diagnosis from the first hospitalization in a chain of hospitalizations (square); 2) estimated using a prediction algorithm and primary diagnosis at any hospitalization in a chain of hospitalizations (circle); and 3) estimated using a prediction algorithm and primary or secondary diagnoses at any hospitalization in a chain of hospitalizations (triangle)

Download PDF

Version 1

posted

You are reading this latest preprint version

Declining hip fracture burden in Sweden 1998-2019 and consequences for projections through 2050

Status:

Version 1

Abstract

Figures

Introduction

Methods

Results

Discussion

Declarations

References

Supplementary Files

Status:

Version 1