Introduction: A secular trend in hip fracture incidence shows a decrease in Europe and North America. Although the results are age adjusted – is there a decreasing trend in all age groups or is just the fracture curve drifting?

Material and Methods: In this population based material includes all hip fractures 1993–2005, in men and women, age 50 years or older admitted to the University Hospital in Ume̊ in the northern part of Sweden. The total numbers of hip fractures in this age group over this 13 years period were 2919 (909(31%) in men and 2010 (69%) in women). The crude numbers of fractures and incidence were mean value over the time periods 1993–96 and 2001–05.

Results: The age adjusted incidence were showing a declining trend X 1993–96, Y 1997–2000, Z 2001–05(ska kolla med Hans hur jag räknar ut). The crude numbers of hip fractures per year are more or less constant 93–96 222,25 (150 women and 72,25 men), 01–05 221,1 (152,4 women and 68,8 men) When analyzing the age groups separately a 50% increase in crude fracture rate was noted in women age 90 or older (11,5 hip fractures/year(93–96) and 24,6 hip fractures/year(01–05). The latter fracture rate can be compared with the crude number of hip fractures in women age 75–79 (27,2 fractures/year 2001–05).

The incidence was also increasing during this period of time, from 270/10.000 to 390/10.000 among the women age 90 or older.

In men there were a declining trend in both incidence and crude numbers, only a small increase of crude fractures were noted among the oldest men age 90 or older.

Conclusion: Though the age adjusted incidence shows a declining secular trend the crude facture rate and incidence are increasing in the oldest women age 90 year or older. Changes in this particular age group does not affect the age adjusted incidence when analyzing the whole population age 50 years or older. (kolla upp) Women 90 years and older do now account for the same amount of hip fractures every year as the women 75–79 years. There seem to be a drifting in the hip fracture incidence pattern towards older. It is therefore important to analyze secular changes in incidence separately in different age groups so that changes in fracture pattern can be identified.